Updated textproc/p5-Text-SimpleTable to 2.05

(wen)

Update to 2.05

Upstream changes:
2.05  2010-06-06 16:38:00
        - Add CJK support (usausat)
        - Add unicode box support (pjsg)

(wen)

Updated textproc/p5-YAML-LibYAML to 0.70

(wen)

Update to 0.70

Upstream changes:
0.70 Sat Jun  9 21:53:05 CEST 2018
- Fix format specifier/argument mismatch PR#79 @a3f++
- Travis CI: Test on Perl 5.26 PR#80 @a3f++
- Fix a C90-compatibility issue PR#81 @stphnlyd++

(wen)

Updated time/p5-DateTime-Locale to 1.22

(wen)

Update to 1.22

Upstream changes:
1.22    2018-06-08

- Update the glibc source data to the data from glibc 2.27.

- Added documentation for this glibc data.

- Reformatted the DateTime::Locale::FromData docs.

1.21    2018-06-08

- Added a new DateTime::Local->register_from_data method. This replaces the
  old custom locale registration system with something much simpler. Instead
  of providing a new class you simply provide a hash of data for the locale
  and it becomes available for use via DateTime::Local->load. Implemented by
  Ryley Breiddal. PR #21. Addresse GH #19.

(wen)

Updated time/p5-Time-Local to 1.27

(wen)

Update to 1.27

Upstream changes:
1.27    2018-06-09

- This release is identical to 1.26 except that it is a stable release, not a
  trial.

- Fixed the tests so they will not break in the year 2020. The tests were
  passing in 2-digit years on the assumption that a year like "70" would be
  treated as 1970, but starting in 2020 this would be treated as 2070
  instead. Reported by Bernhard M. Wiedemann. RT #124787.

- Added timelocal_modern and timegm_modern variants as optional exports. These
  versions do not munge years at all. They treat the year value that they are
  given exactly as-is.

1.26    2018-05-20 (TRIAL RELEASE)

- Fixed the tests so they will not break in the year 2020. The tests were
  passing in 2-digit years on the assumption that a year like "70" would be
  treated as 1970, but starting in 2020 this would be treated as 2070
  instead. Reported by Bernhard M. Wiedemann. RT #124787.

- Added timelocal_modern and timegm_modern variants as optional exports. These
  versions do not munge years at all. They treat the year value that they are
  given exactly as-is.

(wen)

Updated math/py-numpy to 1.14.4

(wen)

Update to 1.14.4

Upstream changes:
NumPy 1.14.4 Release Notes

This is a bugfix release for bugs reported following the 1.14.3 release. The most significant fixes are:

    fixes for compiler instruction reordering that resulted in NaN's not being properly propagated in np.max and np.min,
    fixes for bus faults on SPARC and older ARM due to incorrect alignment checks.

(wen)

Updated devel/p5-Devel-NYTProf to 6.06

(wen)

Update to 6.06

Upstream changes:
Changes in Devel::NYTProf 6.06 - 4th June 2018

Fix sorting of numbers ending ...5s as microsec
  thanks to pichi. #120

Fix tests for Strawberry Perl portable
  thanks to shawnlaffan. #123

Fixed broken link in the pod to YAPC::NA 2014 talk video
  thanks to manwar. #116

Add "NYTProf" to buffer overflow error message for easier triage
  thanks to atomicstack. #119.

Add appveyor config file for CI on Windows
  thanks to shawnlaffan. #117

(wen)

Updated devel/p5-Devel-CheckLib to 1.13

(wen)

Update to 1.13

Upstream changes:
1.13    2018-06-23      Improve Makefile.PL
                        update compatibility code from App-EUMM-Upgrade-1.0

(wen)

Updated devel/p5-Date-Manip to 6.72

(wen)

Update to 6.72

Upstream changes:
6.72  2018-06-06
  -  Fixed recurrence bug
        The fix in the previous version for a rare recurrence problem broke
        another recurrence form. It is now corrected. Michael Schout (GitHub
        #20)

  -  Fixed version problem
        The wrong version was included in two files for some reason. This is
        fixed.

  -  Documentation fixes
        Fixed a grammatical error reported by Xavier Guimard (GitHub #19).

6.71  2018-06-01
  -  Fixed an extremely rare problem with recurrences
        It is possible to specify a recurrence that never produces a valid
        date. In these cases, looking for dates went into an infinite loop.
        The MaxRecurAttempts config variable was added which will stop that
        from happening. If no occurrence was found, an error condition will
        be set. Dean Hamstead (RT 123708)

  -  Changes file supported
        It was requested that I include a valid Changes file. I wrote a
        simple script to convert the Change6.pod file into a valid Change
        file. The Changes6.pod file is still the canonical source of this
        information! Requested by H. Merijn Brand

  -  Fix for timezone determination
        On MacOS X High Sierra, some of the timezone files were symlinks,
        but not properly followed. This was fixed by Stu Tomlinson (GitHub
        #15).

  -  Time zone fixes
        Newest zoneinfo data (tzdata 2018e)

  -  Documentation fixes
        Minor fix provided by Mohammad S Anwar (GitHub #17)

(wen)

Updated devel/p5-App-perlbrew to 0.83

(wen)

Update to 0.83

Upstream changes:
0.83
        - Released at 2018-05-30T08:04:43+0900
        - Thanks to our contributors: Alceu Rodrigues de Freitas Junior, Dan Book, Luca Ferrari, Mar鱈a In辿s Parnisari, Mohammad S Anwar, Shoichi Kaji, sylints
        - my respect to search.cpan.org
        - MAKE env var can now be used to specify the "make" program.
        - New command: install-cpm
        - "install" command now has a "--builddir" option for specifying a work directory
        - "list" command now takes "--verbose" option.
        - "clone-modules" command is now better documented.

(wen)

Updated devel/p5-Alien-Base-ModuleBuild to 1.04

(wen)

Update to 1.04

Upstream changes:
1.04 2018-06-06 09:00:21 -0400
  - ALIEN_INSTALL_TYPE is now fully compatible with Alien::Build
    Previously there were some edge cases with setting it to
    'default'

1.03 2018-05-09 08:15:56 -0400
  - Fix warning

(wen)

Updated devel/p5-Alien-Build to 1.43

(wen)

Update to 1.43
Add missing BUILD_DEPENDS

Upstream changes:
1.43      2018-06-03 06:41:28 -0400
  - Documentation fixes
  - Use a custom site.config to ensure libraries are installed in lib
    and not lib64 or lib/64 for share installs

1.42      2018-05-09 08:19:18 -0400
  - Corrected typo for install_prop->{old}->{prefix} (had been preifx)

(wen)

Updated misc/p5-Locale-Codes to 3.57

(wen)

Update to 3.57

Upstream changes:
3.57  2018-06-01  sbeck
  -  NEW CODE(s)

(wen)

Updated security/p5-Net-DNS-SEC to 1.09

(wen)

Update to 1.09

Upstream changes:
**** 1.09 Jun 4, 2018

        Avoid use of EC_GROUP_new, EC_GROUP_set_curve_GFp, and
        EC_GFp_mont_method which are expected to disappear.

        Fix filename conflict when tests run in parallel using make -j

**** 1.08 May 11, 2018

        Internal reorganisation to use OpenSSL EVP interface

(wen)

Updated textproc/p5-Lingua-EN-Tagger to 0.29

(wen)

Update to 0.29

Upstream changes:
0.29            Aaron Coburn            5/18/18
        Add an add_tags_incrementally method for a more
        stream-based parsing mechanism. Updated email address.

(wen)

Updated textproc/p5-Search-Indexer to 0.79

(wen)

Update to 0.79

Upstream changes:
0.79 30.05.2018
  - fixed bug on full text search

(wen)

Updated textproc/p5-Text-Format to 0.61

(wen)

Update to 0.61

Upstream changes:
0.61    2018-05-21
    - Convert to Dist-Zilla

(wen)

Updated textproc/p5-YAML to 1.26

(wen)

Update to 1.26

Upstream changes:
1.26 Fri May 18 21:57:20 CEST 2018
- Fix bug introduced in 1.25 - loading of quoted string with colon as
  sequence element (tinita, fixes issue#208)
- Support zero indented block sequences (PR# 207 tinita)

(wen)

Updated time/p5-DateTime-TimeZone to 2.19

(wen)

Update to 2.19

Upstream changes:
2.19    2018-05-13

- This release is based on version 2018e of the Olson database. This release
  includes contemporary changes for North Korea.

(wen)

Updated www/p5-HTTP-OAI to 4.07

(wen)

Update to 4.07

Upstream changes:
4.07  2018-05-28 15:14:54 CEST
  - Fixing counting number of tests
  - Fixed adding XSLT stylesheet process instructions #4

(wen)

Updated www/p5-CGI-Lite to 3.02

(wen)

Update to 3.02

Upstream changes:
* v3.02 - 19th May 2018

No changes from v3.01_02

* v3.01_02 - 8th May 2018

BUG FIX: Uploads on MSWin32 were not correctly converted if the first
    character of a buffer was LF. Changed regex to negative lookarounds.

* v3.01_01 - 4th May 2018

BUG FIX: Uploads could fail if n * buffer_size fell inside the boundary
    marker. Thanks pru-mike (PR 1). (possibly fixes issue 116670)

BUG FIX: Changing the boundary marker for second and subsequent uploads
    causes errors. Thanks pru-mike (PR 1). (issue 125130)

Test suite enhanced: now with Test::Vars (0.12) and fix of tmp filename
    race condition.

(wen)

Updated www/p5-HTTP-Message to 6.18

(wen)

Update to 6.18

Upstream changes:
6.18      2018-06-05 16:29:15Z
    - Revert status_message to original code (GH#111) (Theo van Hoesel)

6.17      2018-06-05 01:55:34Z
    - Documented status code 451 in the list of constants (GH #104) (Robert Rothenberg)
    - Status code 451 is cachable by default, as per RFC 7725 (GH #104) (Robert Rothenberg)
    - Add default status_message for unknown status codes (GH#105) (Robert Rothenberg)
    - Corrections to the documentation (GH#102) (Jonas B. Nielsen)

(wen)

Updated www/p5-libwww to 6.34

(wen)

Update to 6.34

Upstream changes:
6.34      2018-06-05 18:48:59Z
    - timeout() documentation has more details about the response (thanks,
      Andrew Grangaard!)
    - HTTP::Status is now upgraded to 6.18 if 6.17 is installed (which had a
      bug).

(wen)

Updated security/p5-String-Random to 0.30

(wen)

Update to 0.30

Upstream changes:
0.30    2018-05-22
    - Convert to Dist-Zilla.

(wen)

Updated math/R-VGAM to 1.0.5

(wen)

Update to 1.0.5

Upstream changes:
CHANGES IN VGAM VERSION 1.0-5

NEW FEATURES

    o  extlogit() now handles 'deriv = 3'.
    o  Generic function hdeff() implements an analytic solution
        for the following families: borel.tanner(), felix(), lindley().
        For almost all other families, finite-difference approximations
        to derivatives means that first and second derivatives can be
        computed, even with models with 'xij' terms.
    o  Generic function wald.stat() implements Wald tests with SEs evaluated
        at the null values, not at the original MLE, so do not suffer from the
        Hauck-Donner effect.
    o  "vglm" objects have a new "charfun" slot, for the
        characteristic function.
    o  "summary.vglm" and "summary.vlm" objects have new
        "coef4lrt0", "coef4score0", "coef4wald0" slots, for
        storing the 'Wald table' equivalent of LRTs, score tests and
        modified Wald tests.
        The latter has its SEs computed at the null values with the
        other coefficients obtained by further IRLS iterations, etc.
        Function summaryvglm() has arguments 'lrt0.arg', 'score0.arg',
        'wald0.arg'.
    o  TIC() is new, for the Takeuchi's Information Criterion.
        Thanks to Khedhaouiria Dikra for suggesting this.
    o  mills.ratio() and mills.ratio2() are exported.
    o  New functions: lrt.stat(), score.stat(), wald.stat(),
        which.etas(), which.xij().
    o  cauchy1() and cauchy() handle multiple responses and have
        been modernized a bit.
    o  Tested okay on R 3.4.3.

BUG FIXES and CHANGES

    o  Setting 'deriv.arg' a positive value in plotvgam() when there
        are no s() terms results in a warning.
        Thanks to Barry Goodwin for detecting this.
    o  cens.poisson() can better handle large lambda values, at least for
        left and right censored data (but not for interval-censored data yet).
        Thanks to Eugenie Hunsicker for picking up deficiencies in the code.
    o  In multinomial.Rd, it was stated that setting parallel = TRUE
        did not make the intercepts the same. It does make them the same.
        Thanks to Stuart Coles for picking this up.
    o  binomialff(multiple.responses = TRUE) returned an incorrect deviance.
    o  bilogistic() uses SFS rather than BFGS as its algorithm.
    o  Deprecated: lrp(), normal1() [use uninormal() instead].

(wen)

Updated devel/p5-Devel-CheckLib to 1.12

(wen)

Update to 1.12

Upstream changes:
1.12    2018-05-23      Fix order of flags for gcc
                        Fix problems with executable binaries

(wen)

Updated devel/p5-Test-POE-Client-TCP to 1.14

(wen)

Update to 1.14

Upstream changes:
ersion 1.14 at 2018-05-29 08:22:21 +0000
-----------------------------------------

  Change: 7ba57fa9ee6d411fc7cd1336a985e57565d082c9
  Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
  Date : 2018-05-29 09:22:21 +0000

    Added accessor for 'context' and update documentation

(wen)

Updated math/R-akima to 0.6.2

(wen)

Update to 0.6.2

(wen)

Updated math/R-sp to 1.2.7

(wen)

Update to 1.2.7

No upstream changelog found.

(wen)

Updated devel/p5-Test-Warn to 0.34

(wen)

Update to 0.34

Upstream changes:
0.34  2018-05-31
        - Added a note that XS warnings might not be caught
          https://rt.cpan.org/Ticket/Display.html?id=42070&results=3c71d1b101a730e185691657f3b02f21
          https://github.com/hanfried/test-warn/issues/1
        - Removed TODO section
        - Updated Copyright section

0.33  2018-05-29
        - Patch from Era Eriksson to clean up some manpage language
          Bug-Debian: #322351
          https://rt.cpan.org/Ticket/Display.html?id=49519

(wen)

Updated devel/R-rlang to 0.2.1

(wen)

Update to 0.2.1

Upstream changes:
rlang 0.2.1
This is a maintenance release that fixes several tidy evaluation issues.

Functions with tidy dots support now allow splicing atomic vectors.

Quosures no longer capture the current srcref.

Formulas are now evaluated in the correct environment by eval_tidy(). This fixes issues in dplyr and other tidy-evaluation interfaces.

(wen)

Updated devel/R-inline to 0.3.15

(wen)

Update to 0.3.15

Upstream changes:
Changes in inline version 0.3.15 (2018-05-18)
Correct requireNamespace() call thanks (Alexander Grueneberg in #5).

Small simplification to .travis.yml; also switch to https.

Use seq_along instead of seq(along=...) (Watal M. Iwasaki) in #6).

Update package manual page using references to DESCRIPTION file [CRAN request].

Minor packaging updates.

(wen)

Updated math/R to 3.5.0

(wen)

Update to 3.5.0

Upstream changes:
CHANGES IN R 3.5.0
SIGNIFICANT USER-VISIBLE CHANGES
All packages are by default byte-compiled on installation. This makes the installed packages larger (usually marginally so) and may affect the format of messages and tracebacks (which often exclude .Call and similar).

NEW FEATURES
factor() now uses order() to sort its levels, rather than sort.list(). This allows factor() to support custom vector-like objects if methods for the appropriate generics are defined. It has the side effect of making factor() succeed on empty or length-one non-atomic vector(-like) types (e.g., "list"), where it failed before.

diag() gets an optional names argument: this may require updates to packages defining S4 methods for it.

chooseCRANmirror() and chooseBioCmirror() no longer have a useHTTPS argument, not needed now all R builds support https:// downloads.

New summary() method for warnings() with a (somewhat experimental) print() method.

(methods package.) .self is now automatically registered as a global variable when registering a reference class method.

tempdir(check = TRUE) recreates the tempdir() directory if it is no longer valid (e.g. because some other process has cleaned up the ‘/tmp’ directory).

New askYesNo() function and "askYesNo" option to ask the user binary response questions in a customizable but consistent way. (Suggestion of PR#17242.)

New low level utilities ...elt(n) and ...length() for working with ... parts inside a function.

isTRUE() is more tolerant and now true in

  x <- rlnorm(99)
  isTRUE(median(x) == quantile(x)["50%"])
New function isFALSE() defined analogously to isTRUE().

The default symbol table size has been increased from 4119 to 49157; this may improve the performance of symbol resolution when many packages are loaded. (Suggested by Jim Hester.)

line() gets a new option iter = 1.

Reading from connections in text mode is buffered, significantly improving the performance of readLines(), as well as scan() and read.table(), at least when specifying colClasses.

order() is smarter about picking a default sort method when its arguments are objects.

available.packages() has two new arguments which control if the values from the per-session repository cache are used (default true, as before) and if so how old cached values can be to be used (default one hour).

These arguments can be passed from install.packages(), update.packages() and functions calling that: to enable this available.packages(), packageStatus() and download.file() gain a ... argument.

packageStatus()'s upgrade() method no longer ignores its ... argument but passes it to install.packages().

installed.packages() gains a ... argument to allow arguments (including noCache) to be passed from new.packages(), old.packages(), update.packages() and packageStatus().

factor(x, levels, labels) now allows duplicated labels (not duplicated levels!). Hence you can map different values of x to the same level directly.

Attempting to use names<-() on an S4 derivative of a basic type no longer emits a warning.

The list method of within() gains an option keepAttrs = FALSE for some speed-up.

system() and system2() now allow the specification of a maximum elapsed time (‘timeout’).

debug() supports debugging of methods on any object of S4 class "genericFunction", including group generics.

Attempting to increase the length of a variable containing NULL using length()<- still has no effect on the target variable, but now triggers a warning.

type.convert() becomes a generic function, with additional methods that operate recursively over list and data.frame objects. Courtesy of Arni Magnusson (PR#17269).

lower.tri(x) and upper.tri(x) only needing dim(x) now work via new functions .row() and .col(), so no longer call as.matrix() by default in order to work efficiently for all kind of matrix-like objects.

print() methods for "xgettext" and "xngettext" now use encodeString() which keeps, e.g. "\n", visible. (Wish of PR#17298.)

package.skeleton() gains an optional encoding argument.

approx(), spline(), splinefun() and approxfun() also work for long vectors.

deparse() and dump() are more useful for S4 objects, dput() now using the same internal C code instead of its previous imperfect workaround R code. S4 objects now typically deparse perfectly, i.e., can be recreated identically from deparsed code.

dput(), deparse() and dump() now print the names() information only once, using the more readable (tag = value) syntax, notably for list()s, i.e., including data frames.

These functions gain a new control option "niceNames" (see .deparseOpts()), which when set (as by default) also uses the (tag = value) syntax for atomic vectors. On the other hand, without deparse options "showAttributes" and "niceNames", names are no longer shown also for lists. as.character(list( c (one = 1))) now includes the name, as as.character(list(list(one = 1))) has always done.

m:n now also deparses nicely when m > n.

The "quoteExpressions" option, also part of "all", no longer quote()s formulas as that may not re-parse identically. (PR#17378)

If the option setWidthOnResize is set and TRUE, R run in a terminal using a recent readline library will set the width option when the terminal is resized. Suggested by Ralf Goertz.

If multiple on.exit() expressions are set using add = TRUE then all expressions will now be run even if one signals an error.

mclapply() gets an option affinity.list which allows more efficient execution with heterogeneous processors, thanks to Helena Kotthaus.

The character methods for as.Date() and as.POSIXlt() are more flexible via new arguments tryFormats and optional: see their help pages.

on.exit() gains an optional argument after with default TRUE. Using after = FALSE with add = TRUE adds an exit expression before any existing ones. This way the expressions are run in a first-in last-out fashion. (From Lionel Henry.)

On Windows, file.rename() internally retries the operation in case of error to attempt to recover from possible anti-virus interference.

Command line completion on :: now also includes lazy-loaded data.

If the TZ environment variable is set when date-time functions are first used, it is recorded as the session default and so will be used rather than the default deduced from the OS if TZ is subsequently unset.

There is now a [ method for class "DLLInfoList".

glm() and glm.fit get the same singular.ok = TRUE argument that lm() has had forever. As a consequence, in glm(*, method = <your_own>), user specified methods need to accept a singular.ok argument as well.

aspell() gains a filter for Markdown (‘.md’ and ‘.Rmd’) files.

intToUtf8(multiple = FALSE) gains an argument to allow surrogate pairs to be interpreted.

The maximum number of DLLs that can be loaded into R e.g. via dyn.load() has been increased up to 614 when the OS limit on the number of open files allows.

Sys.timezone() on a Unix-alike caches the value at first use in a session: inter alia this means that setting TZ later in the session affects only the current time zone and not the system one.

Sys.timezone() is now used to find the system timezone to pass to the code used when R is configured with --with-internal-tzcode.

When tar() is used with an external command which is detected to be GNU tar or libarchive tar (aka bsdtar), a different command-line is generated to circumvent line-length limits in the shell.

system(*, intern = FALSE), system2() (when not capturing output), file.edit() and file.show() now issue a warning when the external command cannot be executed.

The “default” ("lm" etc) methods of vcov() have gained new optional argument complete = TRUE which makes the vcov() methods more consistent with the coef() methods in the case of singular designs. The former (back-compatible) behavior is given by vcov(*, complete = FALSE).

coef() methods (for lm etc) also gain a complete = TRUE optional argument for consistency with vcov().
For "aov", both coef() and vcov() methods remain back-compatibly consistent, using the other default, complete = FALSE.

attach(*, pos = 1) is now an error instead of a warning.

New function getDefaultCluster() in package parallel to get the default cluster set via setDefaultCluster().

str(x) for atomic objects x now treats both cases of is.vector(x) similarly, and hence much less often prints "atomic". This is a slight non-back-compatible change producing typically both more informative and shorter output.

write.dcf() gets optional argument useBytes.

New, partly experimental packageDate() which tries to get a valid "Date" object from a package ‘DESCRIPTION’ file, thanks to suggestions in PR#17324.

tools::resaveRdaFiles() gains a version argument, for use when packages should remain compatible with earlier versions of R.

ar.yw(x) and hence by default ar(x) now work when x has NAs, mostly thanks to a patch by Pavel Krivitsky in PR#17366. The ar.yw.default()'s AIC computations have become more efficient by using determinant().

New warnErrList() utility (from package nlme, improved).

By default the (arbitrary) signs of the loadings from princomp() are chosen so the first element is non-negative.

If --default-packages is not used, then Rscript now checks the environment variable R_SCRIPT_DEFAULT_PACKAGES. If this is set, then it takes precedence over R_DEFAULT_PACKAGES. If default packages are not specified on the command line or by one of these environment variables, then Rscript now uses the same default packages as R. For now, the previous behavior of not including methods can be restored by setting the environment variable R_SCRIPT_LEGACY to yes.

When a package is found more than once, the warning from find.package(*, verbose=TRUE) lists all library locations.

POSIXt objects can now also be rounded or truncated to month or year.

stopifnot() can be used alternatively via new argument exprs which is nicer and useful when testing several expressions in one call.

The environment variable R_MAX_VSIZE can now be used to specify the maximal vector heap size. On macOS, unless specified by this environment variable, the maximal vector heap size is set to the maximum of 16GB and the available physical memory. This is to avoid having the R process killed when macOS over-commits memory.

sum(x) and sum(x1,x2,..,x<N>) with many or long logical or integer vectors no longer overflows (and returns NA with a warning), but returns double numbers in such cases.

Single components of "POSIXlt" objects can now be extracted and replaced via [ indexing with 2 indices.

S3 method lookup now searches the namespace registry after the top level environment of the calling environment.

Arithmetic sequences created by 1:n, seq_along, and the like now use compact internal representations via the ALTREP framework. Coercing integer and numeric vectors to character also now uses the ALTREP framework to defer the actual conversion until first use.

Finalizers are now run with interrupts suspended.

merge() gains new option no.dups and by default suffixes the second of two duplicated column names, thanks to a proposal by Scott Ritchie (and Gabe Becker).

scale.default(x, center, scale) now also allows center or scale to be “numeric-alike”, i.e., such that as.numeric(.) coerces them correctly. This also eliminates a wrong error message in such cases.

par*apply and par*applyLB gain an optional argument chunk.size which allows to specify the granularity of scheduling.

Some as.data.frame() methods, notably the matrix one, are now more careful in not accepting duplicated or NA row names, and by default produce unique non-NA row names. This is based on new function .rowNamesDF(x, make.names = *) <- rNms where the logical argument make.names allows to specify how invalid row names rNms are handled. .rowNamesDF() is a “workaround” compatible default.

R has new serialization format (version 3) which supports custom serialization of ALTREP framework objects. These objects can still be serialized in format 2, but less efficiently. Serialization format 3 also records the current native encoding of unflagged strings and converts them when de-serialized in R running under different native encoding. Format 3 comes with new serialization magic numbers (RDA3, RDB3, RDX3). Format 3 can be selected by version = 3 in save(), serialize() and saveRDS(), but format 2 remains the default for all serialization and saving of the workspace. Serialized data in format 3 cannot be read by versions of R prior to version 3.5.0.

The "Date" and “date-time” classes "POSIXlt" and "POSIXct" now have a working `length<-` method, as wished in PR#17387.

optim(*, control = list(warn.1d.NelderMead = FALSE)) allows to turn off the warning when applying the default "Nelder-Mead" method to 1-dimensional problems.

matplot(.., panel.first = .) etc now work, as log becomes explicit argument and ... is passed to plot() unevaluated, as suggested by Sebastian Meyer in PR#17386.

Interrupts can be suspended while evaluating an expression using suspendInterrupts. Subexpression can be evaluated with interrupts enabled using allowInterrupts. These functions can be used to make sure cleanup handlers cannot be interrupted.

R 3.5.0 includes a framework that allows packages to provide alternate representations of basic R objects (ALTREP). The framework is still experimental and may undergo changes in future R releases as more experience is gained. For now, documentation is provided in https://svn.r-project.org/R/branches/ALTREP/ALTREP.html.

UTILITIES
install.packages() for source packages now has the possibility to set a ‘timeout’ (elapsed-time limit). For serial installs this uses the timeout argument of system2(): for parallel installs it requires the timeout utility command from GNU coreutils.

It is now possible to set ‘timeouts’ (elapsed-time limits) for most parts of R CMD check via environment variables documented in the ‘R Internals’ manual.

The ‘BioC extra’ repository which was dropped from Bioconductor 3.6 and later has been removed from setRepositories(). This changes the mapping for 6–8 used by setRepositories(ind=).

R CMD check now also applies the settings of environment variables _R_CHECK_SUGGESTS_ONLY_ and _R_CHECK_DEPENDS_ONLY_ to the re-building of vignettes.

R CMD check with environment variable _R_CHECK_DEPENDS_ONLY_ set to a true value makes test-suite-management packages available and (for the time being) works around a common omission of rmarkdown from the VignetteBuilder field.

INSTALLATION on a UNIX-ALIKE
Support for a system Java on macOS has been removed — install a fairly recent Oracle Java (see ‘R Installation and Administration’ §C.3.2).

configure works harder to set additional flags in SAFE_FFLAGS only where necessary, and to use flags which have little or no effect on performance.

In rare circumstances it may be necessary to override the setting of SAFE_FFLAGS.

C99 functions expm1, hypot, log1p and nearbyint are now required.

configure sets a -std flag for the C++ compiler for all supported C++ standards (e.g., -std=gnu++11 for the C++11 compiler). Previously this was not done in a few cases where the default standard passed the tests made (e.g. clang 6.0.0 for C++11).

C-LEVEL FACILITIES
‘Writing R Extensions’ documents macros MAYBE_REFERENCED, MAYBE_SHARED and MARK_NOT_MUTABLE that should be used by package C code instead NAMED or SET_NAMED.

The object header layout has been changed to support merging the ALTREP branch. This requires re-installing packages that use compiled code.

‘Writing R Extensions’ now documents the R_tryCatch, R_tryCatchError, and R_UnwindProtect functions.

NAMEDMAX has been raised to 3 to allow protection of intermediate results from (usually ill-advised) assignments in arguments to BUILTIN functions. Package C code using SET_NAMED may need to be revised.

DEPRECATED AND DEFUNCT
Sys.timezone(location = FALSE) is defunct, and is ignored (with a warning).

methods:::bind_activation() is defunct now; it typically has been unneeded for years.

The undocumented ‘hidden’ objects .__H__.cbind and .__H__.rbind in package base are deprecated (in favour of cbind and rbind).

The declaration of pythag() in ‘Rmath.h’ has been removed — the entry point has not been provided since R 2.14.0.

BUG FIXES
printCoefmat() now also works without column names.

The S4 methods on Ops() for the "structure" class no longer cause infinite recursion when the structure is not an S4 object.

nlm(f, ..) for the case where f() has a "hessian" attribute now computes LL' = H + µI correctly. (PR#17249).

An S4 method that “rematches” to its generic and overrides the default value of a generic formal argument to NULL no longer drops the argument from its formals.

Rscript can now accept more than one argument given on the #! line of a script. Previously, one could only pass a single argument on the #! line in Linux.

Connections are now written correctly with encoding "UTF-16LE". (PR#16737).

Evaluation of ..0 now signals an error. When ..1 is used and ... is empty, the error message is more appropriate.

(Windows mainly.) Unicode code points which require surrogate pairs in UTF-16 are now handled. All systems should properly handle surrogate pairs, even those systems that do not need to make use of them. (PR#16098)

stopifnot(e, e2, ...) now evaluates the expressions sequentially and in case of an error or warning shows the relevant expression instead of the full stopifnot(..) call.

path.expand() on Windows now accepts paths specified as UTF-8-encoded character strings even if not representable in the current locale. (PR#17120)

line(x, y) now correctly computes the medians of the left and right group's x-values and in all cases reproduces straight lines.

Extending S4 classes with slots corresponding to special attributes like dim and dimnames now works.

Fix for legend() when fill has multiple values the first of which is NA (all colours used to default to par(fg)). (PR#17288)

installed.packages() did not remove the cached value for a library tree that had been emptied (but would not use the old value, just waste time checking it).

The documentation for installed.packages(noCache = TRUE) incorrectly claimed it would refresh the cache.

aggregate(<data.frame>) no longer uses spurious names in some cases. (PR#17283)

object.size() now also works for long vectors.

packageDescription() tries harder to solve re-encoding issues, notably seen in some Windows locales. This fixes the citation() issue in PR#17291.

poly(<matrix>, 3) now works, thanks to prompting by Marc Schwartz.

readLines() no longer segfaults on very large files with embedded '\0' (aka ‘nul’) characters. (PR#17311)

ns() (package splines) now also works for a single observation. interpSpline() gives a more friendly error message when the number of points is less than four.

dist(x, method = "canberra") now uses the correct definition; the result may only differ when x contains values of differing signs, e.g. not for 0-1 data.

methods:::cbind() and methods:::rbind() avoid deep recursion, thanks to Suharto Anggono via PR#17300.

Arithmetic with zero-column data frames now works more consistently; issue raised by Bill Dunlap.

Arithmetic with data frames gives a data frame for ^ (which previously gave a numeric matrix).

pretty(x, n) for large n or large diff(range(x)) now works better (though it was never meant for large n); internally it uses the same rounding fuzz (1e-10) as seq.default() — as it did up to 2010-02-03 when both were 1e-7.

Internal C-level R_check_class_and_super() and hence R_check_class_etc() now also consider non-direct super classes and hence return a match in more cases. This e.g., fixes behaviour of derived classes in package Matrix.

Reverted unintended change in behavior of return calls in on.exit expressions introduced by stack unwinding changes in R 3.3.0.

Attributes on symbols are now detected and prevented; attempt to add an attribute to a symbol results in an error.

fisher.test(*, workspace = <n>) now may also increase the internal stack size which allows larger problem to be solved, fixing PR#1662.

The methods package no longer directly copies slots (attributes) into a prototype that is of an “abnormal” (reference) type, like a symbol.

The methods package no longer attempts to call length<-() on NULL (during the bootstrap process).

The methods package correctly shows methods when there are multiple methods with the same signature for the same generic (still not fully supported, but at least the user can see them).

sys.on.exit() is now always evaluated in the right frame. (From Lionel Henry.)

seq.POSIXt(*, by = "<n> DSTdays") now should work correctly in all cases and is faster. (PR#17342)

.C() when returning a logical vector now always maps values other than FALSE and NA to TRUE (as documented).

Subassignment with zero length vectors now coerces as documented (PR#17344).
Further, x <- numeric(); x[1] <- character() now signals an error ‘replacement has length zero’ (or a translation of that) instead of doing nothing.

(Package parallel.) mclapply(), pvec() and mcparallel() (when mccollect() is used to collect results) no longer leave zombie processes behind.

R CMD INSTALL <pkg> now produces the intended error message when, e.g., the LazyData field is invalid.

as.matrix(dd) now works when the data frame dd contains a column which is a data frame or matrix, including a 0-column matrix/d.f. .

mclapply(X, mc.cores) now follows its documentation and calls lapply() in case mc.cores = 1 also in the case mc.preschedule is false. (PR#17373)

aggregate(<data.frame>, drop=FALSE) no longer calls the function on <empty> parts but sets corresponding results to NA. (Thanks to Suharto Anggono's patches in PR#17280).

The duplicated() method for data frames is now based on the list method (instead of string coercion). Consequently unique() is better distinguishing data frame rows, fixing PR#17369 and PR#17381. The methods for matrices and arrays are changed accordingly.

Calling names() on an S4 object derived from "environment" behaves (by default) like calling names() on an ordinary environment.

read.table() with a non-default separator now supports quotes following a non-whitespace character, matching the behavior of scan().

parLapplyLB and parSapplyLB have been fixed to do load balancing (dynamic scheduling). This also means that results of computations depending on random number generators will now really be non-reproducible, as documented.

Indexing a list using dollar and empty string (l$"") returns NULL.

Using \usage{ data(<name>, package="<pkg>") } no longer produces R CMD check warnings.

match.arg() more carefully chooses the environment for constructing default choices, fixing PR#17401 as proposed by Duncan Murdoch.

Deparsing of consecutive ! calls is now consistent with deparsing unary - and + calls and creates code that can be reparsed exactly; thanks to a patch by Lionel Henry in PR#17397. (As a side effect, this uses fewer parentheses in some other deparsing involving ! calls.)

[R logo] CHANGES IN R 3.4.4
NEW FEATURES
Sys.timezone() tries more heuristics on Unix-alikes and so is more likely to succeed (especially on Linux). For the slowest method, a warning is given recommending that TZ is set to avoid the search.

The version of LAPACK included in the sources has been updated to 3.8.0 (for the routines used by R, a very minor bug-fix change).

parallel::detectCores(logical = FALSE) is ignored on Linux systems, since the information is not available with virtualized OSes.

INSTALLATION on a UNIX-ALIKE
configure will use pkg-config to find the flags to link to jpeg if available (as it should be for the recently-released jpeg-9c and libjpeg-turbo). (This amends the code added in R 3.3.0 as the module name in jpeg-9c is not what that tested for.)

DEPRECATED AND DEFUNCT
Sys.timezone(location = FALSE) (which was a stop-gap measure for Windows long ago) is deprecated. It no longer returns the value of environment variable TZ (usually a location).

Legacy support of make macros such as CXX1X is formally deprecated: use the CXX11 forms instead.

BUG FIXES
power.prop.test() now warns when it cannot solve the problem, typically because of impossible constraints. (PR#17345)

removeSource() no longer erroneously removes NULL in certain cases, thanks to Dénes Tóth.

nls(`NO [mol/l]` ~ f(t)) and nls(y ~ a) now work. (Partly from PR#17367)

R CMD build checks for GNU cp rather than assuming Linux has it. (PR#17370 says ‘Alpine Linux’ does not.)

Non-UTF-8 multibyte character handling fixed more permanently (PR#16732).

sum(<large ints>, <stuff>) is more consistent. (PR#17372)

rf() and rbeta() now also work correctly when ncp is not scalar, notably when (partly) NA. (PR#17375)

is.na(NULL) no longer warns. (PR#16107)

R CMD INSTALL now correctly sets C++ compiler flags when all source files are in sub-directories of ‘src’.

(wen)

Updated devel/R-Rcpp to 0.12.17

(wen)

Update to 0.12.17

Upstream changes:
Changes in Rcpp version 0.12.17 (2018-05-09)
Changes in Rcpp API:

The random number Generator class no longer inhreits from RNGScope (Kevin in #837 fixing #836).

A spurious parenthesis was removed to please gcc8 (Dirk fixing #841)

The optional Timer class header now undefines FALSE which was seen to have side-effects on some platforms (Romain in #847 fixing #846).

Optional StoragePolicy attributes now also work for string vectors (Romain in #850 fixing #849).

Changes in Rcpp Documentation:

A few old typesetting conventions from the prior Rnw format have been corrected (Peter Hickey in #831; Joris Meys; Dirk)

Two internal links to the introduction published in JSS have been updated to the changed filename given the newer TAS introduction.

Some remaining backticks were replaced with straight quotes (Ralf Stubner in #845).

A citation to the Rcpp introducion in the The American Statistician has been added to the introductory and FAQ vignettes.

Changes in Rcpp version 0.12.16 (2018-03-08)
Changes in Rcpp API:

Rcpp now sets and puts the RNG state upon each entry to an Rcpp function, ensuring that nested invocations of Rcpp functions manage the RNG state as expected (Kevin in #825 addressing #823).

The R::pythag wrapper has been commented out; the underlying function has been gone from R since 2.14.0, and ::hypot() (part of C99) is now used unconditionally for complex numbers (Dirk in #826).

The long long type can now be used on 64-bit Windows (Kevin in #811 and again in #829 addressing #804).

Changes in Rcpp Attributes:

Code generated with cppFunction() now uses .Call() directly (Kirill Mueller in #813 addressing #795).

Changes in Rcpp Documentation:

The Rcpp FAQ vignette is now indexed as 'Rcpp-FAQ'; a stale Gmane reference was removed and entry for getting compilers under Conda was added.

The top-level README.md now has a Support section.

The Rcpp.bib reference file was refreshed to current versions.

(wen)

Updated math/R-R2HTML to 2.3.2

(wen)

Update to 2.3.2

(wen)

Updated math/R-PHYLOGR to 1.0.10

(wen)

Update to 1.0.10

(wen)

Updated devel/R-IRdisplay to 0.5.0

(wen)

Update to 0.5.0

No upstream changelog found.

(wen)

Updated devel/R-repr to 0.15.0

(wen)

Update to 0.15.0

No upstream changelog found.

(wen)

Added converters/R-base64enc version 0.1.3

(wen)

Add R-base64enc

(wen)

Import base64enc_0.1-3 as converters/R-base64enc.

This package provides tools for handling base64 encoding. It is
more flexible than the orphaned base64 package.

(wen)

Updated math/R-DAAG to 1.22

(wen)

Update to 1.22

Upstream changes:
Changes in DAAG version 1.21

NEW FEATURES
  o The dataset hurricNamed (deaths from Atlantic hurricanes,
    damage assessments, and other measures, from 1950 onwards)
    has been added.
  o The dataset cps3 (a subset of cps10 has been added
  o The help file CVlm.Rd has been amended to give the correct
    return value for CVlm() and cv.lm().
  o An enhanced version of overlapDensity()
    (an alias is  overlap.density()), as assumed by the code for
    Figure 13.3 on page 421 of the 3rd edition of "Data Analysis
    and Graphics using R", has replaced the former version.

(wen)

Updated www/p5-Plack-Middleware-MethodOverride to 0.20

(wen)

Update to 0.20

Upstream changes:
0.20  2018-05-25 03:35:00 PDT
      - Repackage the module. Same as 0.15, except repositroy metadata updates

0.19  2018-05-25 03:12:21 PDT
      - Fix license and copyright that have been accidentally broken in
        0.16-8 META files

0.18  2018-05-24 23:01:57 PDT
      - Maintenance release to retroactively update Changes

0.17  2018-05-23 23:37:03 PDT
      - Fixed prereqs and META.json

0.16  2018-05-23 23:10:55 PDT
      - Module ownership transferred
      - Switch to Milla and Module::Build::Tiny

(wen)

Updated www/p5-Net-HTTP to 6.18

(wen)

Update to 6.18

Upstream changes:
6.18      2018-05-09 14:55:43Z
    - Move test that relies on http://httpbin.org to author tests

(wen)

Updated devel/p5-File-pushd to 1.016

(wen)

Update to 1.016

Upstream changes:
1.016    2018-05-20 13:31:17-04:00 America/New_York

    - No changes from 1.015-TRIAL

1.015    2018-05-13 19:26:26-04:00 America/New_York (TRIAL RELEASE)

    [FIXED]

    - Directories created with tempd will only be cleaned up by the
      PID that created them, not by forked children.

(wen)

Updated devel/p5-CPANPLUS to 0.9174

(wen)

Update to 0.9174

Upstream changes:
0.9174      Mon May 21 20:28:36 BST 2018

* Fix circular dependency on CPANPLUS::Dist::Build when
  'prefer_makefile' is set to 0

(wen)

Updated devel/p5-List-SomeUtils-XS to 0.57

(wen)

Update to 0.57

Upstream changes:
0.57    2018-05-26

- Fix C compiler warnings. Reported by Jim Keenan. GH #2.

(wen)

Updated net/p5-Net-Frame-Dump to 1.16

(wen)

Update to 1.16

Upstream changes:
1.16 Thu 17 May 13:57:30 CEST 2018
  - bugfix: RT#125315, typos in pod and code.

(wen)

Updated net/p5-Net-Frame to 1.19

(wen)

Update to 1.19

Upstream changes:
1.19 Thu 17 May 13:59:33 CEST 2018
  - bugfix: test 13 when DNS getHostIpv6Addr() returns an IPv4 mapped to an IPv6 address
  => https://rt.cpan.org/Ticket/Display.html?id=124780

(wen)

Updated net/p5-SOAP-Lite to 1.27

(wen)

Update toi 1.27

Upstream changes:
1.27 May 14, 2018
    ! Problem with compression in server mode

(wen)

Updated devel/p5-CPAN-Perl-Releases to 3.58

(wen)

Update to 3.58

Upstream changes:
version 3.58 at 2018-05-21 13:21:57 +0000
-----------------------------------------

  Change: 9c2d1a38ad555ec8f3bf1afacb5da13ccd8387d2
  Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
  Date : 2018-05-21 14:21:57 +0000

    Updated for v5.28.0-RC1

(wen)

Updated devel/p5-Error to 0.17026

(wen)

Update to 0.17026

Upstream changes:
May 23 2018 <shlomif@shlomifish.org> (Shlomi Fish)

  Error.pm #0.17026
  - Convert to Dist-Zilla.

(wen)

Updated devel/p5-Type-Tiny-XS to 0.014

(wen)

Update to 0.014

Upstream changes:
0.014  2018-05-20

[ Bug Fixes ]
- Bool was letting overloaded blessed objects pass the type constraint. It
  shouldn't.
  Gregory Oschwald++
  <https://github.com/tobyink/p5-type-tiny-xs/issues/5>

0.013  2018-05-15

[ Bug Fixes ]
- Strings like '00' and '000' shouldn't be accepted as PositiveInt.
  Will Storey++
  <https://github.com/tobyink/p5-type-tiny-xs/issues/7>

(wen)

Updated databases/p5-DBM-Deep to 2.0016

(wen)

Update to 2.0016

Upstream changes:
2.0016 May 20 15:50:00 2018 PDT
    - Fix mistake in the previous release that prevented indexing.

(wen)

Updated databases/p5-DBD-CSV to 0.53

(wen)

Update to 0.53

Upstream changes:
0.53    - 2018-05-20, H.Merijn Brand
    * No folder scanning during automated tests
    * Fix col_names set to empty [] incorrectly skipping first row
      (issue#6) (Ronald Schmidt)
    * Small doc fix (Ronald Schmidt)
    * Tested on FreeBSD

(wen)

Updated time/p5-DateTime to 1.49

(wen)

Update to 1.49

Upstream changes:
1.49  2018-05-20

- Updated the ppport.h with the latest version of Devel::PPPort. This fixes a
  compilation warning when compiling with 5.27.11. Reported by Jim
  Keenan. Fixed GH #81.

(wen)

Updated databases/p5-DBM-Deep to 2.0015

(wen)

Update to 2.0015

Upstream changes:
2.0015 May 20 14:45:00 2018 PDT
    - Fix for tests failing on 5.28 (Thanks, Slaven!)

(wen)

Updated devel/p5-Module-FromPerlVer to 0.030000

(wen)

Update to 0.030000

Upstream changes:
-  v0.3.0  Wed May  9 09:11:35 CDT 2018

    Bump version number to try and find out why this fails for
    lack of a module that isn't required. Failure notices call
    out File::Copy::Recursive, which is not used in the Makefile.PL,
    module, or test code. File::Copy::Recursive is not installed
    here and the tests all pass... but the tests fail for lack of
    it on the smoke systems. Only thing I can figure is that they
    are testing a stale copy of the module. Hopefully bumping the
    version will solve the issue.

    Add test to use File::Find to walk down the tree grep-ing
    "File::Copy::Recursive" and dumping the result along with
    the module version to try and see what is going on.

(wen)

Updated devel/p5-ExtUtils-ModuleMaker to 0.62

(wen)

Update to 0.62

Upstream changes:
0.62    May 15, 2018
Update content of generated Makefile.PL to include META_MERGE where the
system's ExtUtils::MakeMaker is new enough to accommodate that element.  While
this is a step on the road to a "git-ready" module, this does not amount to a
change in effective functionality.

0.61    May 14, 2018
Update content of MANIFEST.SKIP (when INCLUDE_MANIFEST_SKIP is set to true
value) to reflect contemporary best practices (e.g., account for Travis and
AppVeyor configuration files).

(wen)

Updated devel/p5-ExtUtils-Manifest to 1.71

(wen)

Update to 1.71

Upstream changes:
1.71  2018-05-15
  - also skip .travis.yml (thanks, Jerome Eteve!)
  - also skip .appveyor.yml (thanks, James E Keenan!)

(wen)

Updated devel/p5-ExtUtils-InstallPaths to 0.012

(wen)

Update to 0.012

Upstream changes:
0.012    2018-05-10 21:03:33+02:00 Europe/Amsterdam
          - Allow a argument to install_map with source dirs
          - Make tests prove and 5.6 friendly

(wen)

Updated devel/p5-experimental to 0.020

(wen)

Update to 0.020

Upstream changes:
0.020    2018-05-09 21:31:22+02:00 Europe/Amsterdam
          - Add the declared_refs feature

(wen)

Updated www/moodle to 3.5

(wen)

Update to 3.5

Upstream changes:
Moodle 3.5 release notes
Releases > Moodle 3.5 release notes  Release date: 17 May 2018

Here is the full list of fixed issues in 3.5.

See our New Features page for a more user-friendly introduction to Moodle 3.5 with screenshots.

If you are upgrading from previous version, make sure you read the Upgrading documentation.

CONTENTS [hide]
1 Server requirements
1.1 Database requirements
2 Client requirements
2.1 Browser support
3 Major features
3.1 GDPR
3.2 Question bank tagging improvements
3.3 UX: Usability improvements
3.4 LTI Advantage support
3.5 RecordRTC for Atto
3.6 Messaging database tables
4 Other Highlights
4.1 Global search
4.2 Functional changes
4.3 Security issues
4.4 For developers
4.4.1 Upgrading plugins
5 See also
Server requirements
These are just the minimum supported versions. We recommend keeping all of your software and operating systems up-to-date.

Moodle upgrade: Moodle 3.1 or later
PHP version: minimum PHP 7.0.0 Note: minimum PHP version has increased since Moodle 3.3. PHP 7.1.x and 7.2.x are supported too. PHP 7.x could have some engine limitations.
PHP extension intl is required since Moodle 3.4 (it was recommended in 2.0 onwards)
Database requirements
Moodle supports the following database servers. Again, version numbers are just the minimum supported version. We recommend running the latest stable version of any software.

Database Minimum version Recommended
PostgreSQL 9.3 Latest
MySQL 5.5.31 Latest
MariaDB 5.5.31 Latest
Microsoft SQL Server 2008 Latest
Oracle Database 10.2 Latest
Client requirements
Browser support
Moodle is compatible with any standards compliant web browser. We regularly test Moodle with the following browsers:

Desktop:

Chrome
Firefox
Safari
Edge
Internet Explorer
Mobile:

MobileSafari
Google Chrome
For the best experience and optimum security, we recommend that you keep your browser up to date. https://whatbrowser.org

Note: Legacy browsers with known compatibility issues with Moodle 3.5:

Internet Explorer 10 and below
Safari 7 and below

Major features
GDPR
MDL-61275 - GDPR Consenting of Minors and Managing, Versioning and Tracking Privacy Policies and User Consents
MDL-61292 - A new admin tool to manage policy documents
MDL-61423 - Add age and location verification to identify minors
MDL-61302 - Workflow to allow users to agree to all policies
MDL-61301 - Report of user agreed policies and their versions
MDL-61705 - Bulk accept of policies on behalf of users
MDL-61864 - Include policy tool in core
MDL-62286 - Add policy link to the site footer
MDL-61306 - GDPR Data Requests and Data Registry
MDL-59718 - A process to send a request to the data protection officer
MDL-59720 - Delete personal data when it is no longer required
MDL-61307 - Create a new privacy subsystem
MDL-61362 - Ability to create data categories and purposes
MDL-61486 - Data registry with purpose and retention period
MDL-61489 - Report of plugin/components implementing the Privacy API
MDL-61499 - Ability to set default purpose and retention periods for context levels
MDL-61785 - Ability to review and confirm which expired data can be deleted
MDL-61899 - Include data privacy tool in core
MDL-61935 - Ability to specify the lawful bases for the collection of personal data
Question bank tagging improvements
MDL-61066 - Expanded tagging functionality for question bank
MDL-61133 - New modal to add/edit/remove tags on questions
MDL-61135 - Filter questions by tag
MDL-61138 - Show the list of questions in the 'Add a random question' dialog
MDL-61363 - Ability to add question tags at a course level in the edit question form
MDL-61364 - Manage tags at a question and course context level
MDL-61380 - Allow filtering/adding random questions by tag for quizzes
MDL-61410 - Add import/export support for course level question tags
MDL-61444 - New capabilities for tagging questions
UX: Usability improvements
MDL-62021 - Boost 4.0 Migration
MDL-56511 - Update bootstrap 4 to final release
MDL-61657 - Add images to the course cards on the dashboard
LTI Advantage support
MDL-60416 - Add support for LTI Advantage 1.1
RecordRTC for Atto
MDL-60848 - Implement RecordRTC Atto plugin as core feature
MDL-61973 - Update RecordRTC Atto plugin buttons
Messaging database tables
MDL-61254 - Merge messaging database tables
MDL-36941 - Create new tables for messaging
MDL-61255 - Ad-hoc task to upgrade messages to merged table

Other Highlights
Global search
MDL-58885 - Add group support
MDL-59434 - Content aware searching / alternate results sort orders
MDL-60981 - Reindex a single area
MDL-61028 - Allow filtering search by user
MDL-61256 - Search of section titles, summaries
Functional changes
MDL-2051 - Inform student whether and how their selected choice will display
MDL-32585 - SCORM: option to force new attempts
MDL-53226 - Add Moodle DB search engine
MDL-55491 - Use cohort as badge criteria
MDL-56246 - Add site wide default for grade export: include feedback
MDL-59875 - Allow badges as criteria for other badges
MDL-60119 - Feedback - Multiple choice (rated) - remove weights from answer
MDL-61203 - Allow uploading of profile picture to be used as badge criteria
MDL-61601 - Allow cohort themes
MDL-61651 - LTI: line item definition within link to return gradable LTI links
MDL-60811 - Bulk delete self-registered enrolments on participants page
MDL-60682 - Ability to set date/time to nearest minute
MDL-60441 - Ability to add a link to glossary entries
MDL-58411 - Ability to apply file type restrictions for essay question type
MDL-56945 - Add easy return path from PDF grading screen to list of submissions
MDL-52811 - Add force language capability to course settings
MDL-41090 - Allow teachers to embed files when manually grading questions
Security issues
  A number of security related issues were resolved. Details of these issues will be released after a period of approximately one week to allow system administrators to safely update to the latest version.

For developers
MDL-61307 - All plugins must implement Privacy API to be compliant with GDPR requirements. They must implement the API to report on, export and delete stored user data
MDL-56511 - Bootstrap is upgraded to final release of version 4
MDL-61869 - Infer rendering of templatables with no render method
MDL-61298 - Boost: use navigation node icon
Upgrading plugins
1. Check for changes in core APIs

Read lib/upgrade.txt to check for the deprecations and core API changes, make sure you applied them to your plugin. Note that entries there are not sorted by priority but rather by integration time. Below is the list of upgrade.txt files that contain information about upgrading from Moodle 3.4 to Moodle 3.5 (note that if you upgrade from earlier versions there may be more files):

lib/upgrade.txt changes to various core APIs, deprecations, functions removal
calendar/upgrade.txt changes to Calendar API
search/upgrade.txt changes to Global search API
message/upgrade.txt changes to Messages API
course/upgrade.txt changes to Course API
2. Check for changes in the API of your plugin type

Below is the list of plugin types that had API changes between Moodle 3.4 and 3.5:

enrol/upgrade.txt Enrolment method plugins
mod/upgrade.txt Activity module plugins
auth/upgrade.txt Authentication plugins
course/format/upgrade.txt Course format plugins
question/type/upgrade.txt Question type plugins
3. Check for changes in the depended plugins

If your plugin depends on another plugin or calls methods from another plugin, read upgrade.txt in this plugin directory (if it exists). Below is the list of standard plugins that had changes between Moodle 3.4 and 3.5:

tool_mobile, tool_usertours, mod_assign, mod_feedback, mod_quiz, mod_scorm, theme_boost

4. Do a smoke test of your plugin with developer debugging mode

Make sure to check on both Boost and Clean themes. Bootstrap was upgraded in Moodle 3.5

5. Run all behat and phpunit tests

(wen)

Updated devel/p5-Moose to 2.2011

(wen)

Update to 2.2011

Upstream changes:
2.2011  2018-05-16

  [BUG FIXES]

  - Various native trait methods would refuse "0" where they expected a
    string.  They have been fixed to allow all defined, non-reference values,
    and all objects with string, number, or boolean overloads.

  [DOCUMENTATION]

  - Moose::Manual::Construction now notes that roles can modify the BUILD sub.
    (thanks, E. Choroba!)

  [OTHER]

  - adjustments have been made to the Makefile for 'bmake', so it now works
    when not run in compatibility mode (Leon Timmermans).

(wen)

Updated devel/netcdf to 4.6.1

(wen)

Update to 4.6.1

Upstream changes:
## 4.6.1 - March 15, 2018

* [Bug Fix] Corrected an issue which could result in a dap4 failure. See [Github #888](https://github.com/Unidata/netcdf-c/pull/888) for more information.
* [Bug Fix][Enhancement] Allow `nccopy` to control output filter suppresion.  See [Github #894](https://github.com/Unidata/netcdf-c/pull/894) for more information.
* [Enhancement] Reverted some new behaviors that, while in line with the netCDF specification, broke existing workflows.  See [Github #843](https://github.com/Unidata/netcdf-c/issues/843) for more information.
* [Bug Fix] Improved support for CRT builds with Visual Studio, improves zlib detection in hdf5 library. See [Github #853](https://github.com/Unidata/netcdf-c/pull/853) for more information.
* [Enhancement][Internal] Moved HDF4 into a distinct dispatch layer. See [Github #849](https://github.com/Unidata/netcdf-c/pull/849) for more information.

## 4.6.0 - January 24, 2018
* [Enhancement] Full support for using HDF5 dynamic filters, both for reading and writing. See the file docs/filters.md.
* [Enhancement] Added an option to enable strict null-byte padding for headers; this padding was specified in the spec but was not enforced.  Enabling this option will allow you to check your files, as it will return an E_NULLPAD error.  It is possible for these files to have been written by older versions of libnetcdf.  There is no effective problem caused by this lack of null padding, so enabling these options is informational only.  The options for `configure` and `cmake` are `--enable-strict-null-byte-header-padding` and `-DENABLE_STRICT_NULL_BYTE_HEADER_PADDING`, respectively.  See [Github #657](https://github.com/Unidata/netcdf-c/issues/657) for more information.
* [Enhancement] Reverted behavior/handling of out-of-range attribute values to pre-4.5.0 default. See [Github #512](https://github.com/Unidata/netcdf-c/issues/512) for more information.
* [Bug] Fixed error in tst_parallel2.c. See [Github #545](https://github.com/Unidata/netcdf-c/issues/545) for more information.
* [Bug] Fixed handling of corrupt files + proper offset handling for hdf5 files. See [Github #552](https://github.com/Unidata/netcdf-c/issues/552) for more information.
* [Bug] Corrected a memory overflow in `tst_h_dimscales`, see [Github #511](https://github.com/Unidata/netcdf-c/issues/511), [Github #505](https://github.com/Unidata/netcdf-c/issues/505), [Github #363](https://github.com/Unidata/netcdf-c/issues/363) and [Github #244](https://github.com/Unidata/netcdf-c/issues/244) for more information.

## 4.5.0 - October 20, 2017

* Corrected an issue which could potential result in a hang while using parallel file I/O. See [Github #449](https://github.com/Unidata/netcdf-c/pull/449) for more information.
* Addressed an issue with `ncdump` not properly handling dates on a 366 day calendar. See [GitHub #359](https://github.com/Unidata/netcdf-c/issues/359) for more information.

### 4.5.0-rc3 - September 29, 2017

* [Update] Due to ongoing issues, native CDF5 support has been disabled by **default**.  You can use the options mentioned below (`--enable-cdf5` or `-DENABLE_CDF5=TRUE` for `configure` or `cmake`, respectively).  Just be aware that for the time being, Reading/Writing CDF5 files on 32-bit platforms may result in unexpected behavior when using extremely large variables.  For 32-bit platforms it is best to continue using `NC_FORMAT_64BIT_OFFSET`.
* [Bug] Corrected an issue where older versions of curl might fail. See [GitHub #487](https://github.com/Unidata/netcdf-c/issues/487) for more information.
* [Enhancement] Added options to enable/disable `CDF5` support at configure time for autotools and cmake-based builds.  The options are `--enable/disable-cdf5` and `ENABLE_CDF5`, respectively.  See [Github #484](https://github.com/Unidata/netcdf-c/issues/484) for more information.
* [Bug Fix] Corrected an issue when subsetting a netcdf3 file via `nccopy -v/-V`. See [Github #425](https://github.com/Unidata/netcdf-c/issues/425) and [Github #463](https://github.com/Unidata/netcdf-c/issues/463) for more information.
* [Bug Fix] Corrected `--has-dap` and `--has-dap4` output for cmake-based builds. See [GitHub #473](https://github.com/Unidata/netcdf-c/pull/473) for more information.
* [Bug Fix] Corrected an issue where `NC_64BIT_DATA` files were being read incorrectly by ncdump, despite the data having been written correctly.  See [GitHub #457](https://github.com/Unidata/netcdf-c/issues/457) for more information.
* [Bug Fix] Corrected a potential stack buffer overflow.  See [GitHub #450](https://github.com/Unidata/netcdf-c/pull/450) for more information.

### 4.5.0-rc2 - August 7, 2017

* [Bug Fix] Addressed an issue with how cmake was implementing large file support on 32-bit systems. See [GitHub #385](https://github.com/Unidata/netcdf-c/issues/385) for more information.
* [Bug Fix] Addressed an issue where ncgen would not respect keyword case. See [GitHub #310](https://github.com/Unidata/netcdf-c/issues/310) for more information.

### 4.5.0-rc1 - June 5, 2017

* [Enhancement] DAP4 is now included. Since dap2 is the default for urls, dap4 must be specified by
(1) using "dap4:" as the url protocol, or
(2) appending "#protocol=dap4" to the end of the url, or
(3) appending "#dap4" to the end of the url
Note that dap4 is enabled by default but remote-testing is
disbled until the testserver situation is resolved.
* [Enhancement] The remote testing server can now be specified with the `--with-testserver` option to ./configure.
* [Enhancement] Modified netCDF4 to use ASCII for NC_CHAR.  See [Github Pull request #316](https://github.com/Unidata/netcdf-c/pull/316) for more information.
* [Bug Fix] Corrected an error with how dimsizes might be read. See [Github #410](https://github.com/unidata/netcdf-c/issues/410) for more information.
* [Bug Fix] Corrected an issue where 'make check' would fail if 'make' or 'make all' had not run first.  See [Github #339](https://github.com/Unidata/netcdf-c/issues/339) for more information.
* [Bug Fix] Corrected an issue on Windows with Large file tests. See [Github #385](https://github.com/Unidata/netcdf-c/issues/385]) for more information.
* [Bug Fix] Corrected an issue with diskless file access, see [Pull Request #400](https://github.com/Unidata/netcdf-c/issues/400) and [Pull Request #403](https://github.com/Unidata/netcdf-c/issues/403) for more information.
* [Upgrade] The bash based test scripts have been upgraded to use a common test_common.sh include file that isolates build specific information.
* [Upgrade] The bash based test scripts have been upgraded to use a common test_common.sh include file that isolates build specific information.
* [Refactor] the oc2 library is no longer independent of the main netcdf-c library. For example, it now uses ncuri, nclist, and ncbytes instead of its homegrown equivalents.
* [Bug Fix] `NC_EGLOBAL` is now properly returned when attempting to set a global `_FillValue` attribute. See [GitHub #388](https://github.com/Unidata/netcdf-c/issues/388) and [GitHub #389](https://github.com/Unidata/netcdf-c/issues/389) for more information.
* [Bug Fix] Corrected an issue where data loss would occur when `_FillValue` was mistakenly allowed to be redefined.  See [Github #390](https://github.com/Unidata/netcdf-c/issues/390), [GitHub #387](https://github.com/Unidata/netcdf-c/pull/387) for more information.
* [Upgrade][Bug] Corrected an issue regarding how "orphaned" DAS attributes were handled. See [GitHub #376](https://github.com/Unidata/netcdf-c/pull/376) for more information.
* [Upgrade] Update utf8proc.[ch] to use the version now maintained by the Julia Language project (https://github.com/JuliaLang/utf8proc/blob/master/LICENSE.md).
* [Bug] Addressed conversion problem with Windows sscanf.  This primarily affected some OPeNDAP URLs on Windows.  See [GitHub #365](https://github.com/Unidata/netcdf-c/issues/365) and [GitHub #366](https://github.com/Unidata/netcdf-c/issues/366) for more information.
* [Enhancement] Added support for HDF5 collective metadata operations when available. Patch submitted by Greg Sjaardema, see [Pull request #335](https://github.com/Unidata/netcdf-c/pull/335) for more information.
* [Bug] Addressed a potential type punning issue. See [GitHub #351](https://github.com/Unidata/netcdf-c/issues/351) for more information.
* [Bug] Addressed an issue where netCDF wouldn't build on Windows systems using MSVC 2012. See [GitHub #304](https://github.com/Unidata/netcdf-c/issues/304) for more information.
* [Bug] Fixed an issue related to potential type punning, see [GitHub #344](https://github.com/Unidata/netcdf-c/issues/344) for more information.
* [Enhancement] Incorporated an enhancement provided by Greg Sjaardema, which may improve read/write times for some complex files.  Basically, linked lists were replaced in some locations where it was safe to use an array/table.  See [Pull request #328](https://github.com/Unidata/netcdf-c/pull/328) for more information.

(wen)

Updated devel/hdf5 to 1.10.2

(wen)

Update to 1.10.2

Upstream changes:
HDF5 version 1.10.2 released on 2018-03-29
================================================================================
CONTENTS

- New Features
- Support for new platforms and languages
- Bug Fixes since HDF5-1.10.1
- Supported Platforms
- Tested Configuration Features Summary
- More Tested Platforms
- Known Problems

New Features
============

    Configuration and Build Systems:
    --------------------------------
    - CMake builds
    --------------

      - Changed minimum CMake required version to 3.10.

      This change removed the need to support a copy of the FindMPI.cmake module,
      which has been removed, along with its subfolder in the config/cmake_ext_mod
      location.

      (ADB - 2018/03/09)

      - Added pkg-config file generation

      Added pkg-config file generation for the C, C++, HL, and HL C++ libraries.
      In addition, builds on Linux will create h5cc, h5c++, h5hlcc, and h5hlc++ scripts in the bin
      directory that use the pkg-config files. The scripts can be used to build HDF5 C and C++
      applications (i.e, similar to the compiler scripts produced by the Autotools builds).

      (ADB - 2018/03/08, HDFFV-4359)

      - Refactored use of CMAKE_BUILD_TYPE for new variable, which understands
      the type of generator in use.

      Added new configuration macros to use new HDF_BUILD_TYPE variable. This
      variable is set correctly for the type of generator being used for the build.

      (ADB - 2018/01/08, HDFFV-10385, HDFFV-10296)

    - Autotools builds
    ------------------

      - Removed version-specific gcc/gfortran flags for version 4.0 (inclusive)
      and earlier.

      The config/gnu-flags file, which is sourced as a part of the configure
      process, adds version-specific flags for use when building HDF5. Most of
      these flags control warnings and do not affect the final product.

      Flags for older versions of the compiler were consolidated into the
      common flags section. Moving these flags simplifies maintenance of
      the file.

      The upshot of this is that building with ancient versions of gcc
      (<= 4.0) will possibly no longer work without hand-hacking the file
      to remove the flags not understood by that version of the compiler.
      Nothing should change when building with gcc >= 4.1.

      (DER - 2017/05/31, HDFFV-9937)

      - -fno-omit-frame-pointer was added when building with debugging symbols
      enabled.

      Debugging symbols can be enabled independently of the overall build
      mode in both the autotools and CMake. This allows (limited) debugging
      of optimized code. Since many debuggers rely on the frame pointer,
      we've disabled this optimization when debugging symbols are requested
      (e.g.: via building with --enable-symbols).

      (DER - 2017/05/31, HDFFV-10226)

    Library:
    --------
    - Added an enumerated value to H5F_libver_t for H5Pset_libver_bounds().

      Currently, the library defines two values for H5F_libver_t and supports
      only two pairs of (low, high) combinations as derived from these values.
      Thus the bounds setting via H5Pset_libver_bounds() is rather restricted.

      Added an enumerated value (H5F_LIBVER_V18) to H5F_libver_t and
      H5Pset_libver_bounds() now supports five pairs of (low, high) combinations
      as derived from these values.  This addition provides the user more
      flexibility in setting bounds for object creation.

      (VC - 2018/03/14)

    - Added prefix option to VDS files.

      Currently, VDS source files must be in the active directory to be
      found by the virtual file. Adding the option of a prefix to be set
      on the virtual file, using a data access property list (DAPL),
      allows the source files to locate at an absolute or relative path
      to the virtual file.
      Private utility functions in H5D and H5L packages merged into single
      function in H5F package.

      New public APIs:
            herr_t H5Pset_virtual_prefix(hid_t dapl_id, const char* prefix);
            ssize_t H5Pget_virtual_prefix(hid_t dapl_id, char* prefix /*out*/, size_t size);
      The prefix can also be set with an environment variable, HDF5_VDS_PREFIX.

      (ADB - 2017/12/12, HDFFV-9724, HDFFV-10361)

    - H5FDdriver_query() API call added to the C library.

      This new library call allows the user to query a virtual file driver
      (VFD) for the feature flags it supports (listed in H5FDpublic.h).
      This can be useful to determine if a VFD supports SWMR, for example.

      Note that some VFDs have feature flags that may only be present
      after a file has been created or opened (e.g.: the core VFD will
      have the H5FD_FEAT_POSIX_COMPAT_HANDLE flag set if the backing
      store is switched on). Since the new API call queries a generic VFD
      unassociated with a file, these flags will never be returned.

      (DER - 2017/05/31, HDFFV-10215)

    - H5FD_FEAT_DEFAULT_VFD_COMPATIBLE VFD feature flag added to the C library.

      This new feature flag indicates that the VFD is compatible with the
      default VFD. VFDs that set this flag create single files that follow
      the canonical HDF5 file format.

      (DER - 2017/05/31, HDFFV-10214)

    - The H5I_REFERENCE value in the H5I_type_t enum (defined in H5Ipublic.h)
      has been marked as deprectated.

      This ID type value is not used in the C library. i.e.: There are no
      hid_t values that are of ID type H5I_REFERENCE.

      This enum value will be removed in a future major version of the library.
      The code will remain unchanged in the HDF5 1.10.x  releases and branches.

      (DER - 2017/04/05, HDFFV-10252)

    Parallel Library:
    -----------------
    - Enabled compression for parallel applications.

      With this release parallel applications can create and write compressed
      datasets (or the datasets with the filters such as Fletcher32 applied).

      (EIP - 2018/03/29)

    - Addressed slow file close on some Lustre file systems.

      Slow file close has been reported on some Lustre file systems.
      While the ultimate cause is not understood fully, the proximate
      cause appears to be long delays in MPI_File_set_size() calls at
      file close and flush.

      To minimize this problem pending a definitive diagnosis and fix,
      PHDF5 has been modified to avoid MPI_File_set_size() calls when
      possible.  This is done by comparing the library's EOA (End of
      Allocation) with the file systems EOF, and skipping the
      MPI_File_set_size() call if the two match.

      (JRM - 2018/03/29)

    - Optimized parallel open/location of the HDF5 super-block.

      Previous releases of PHDF5 required all parallel ranks to
      search for the HDF5 superblock signature when opening the
      file. As this is accomplished more or less as a synchronous
      operation, a large number of processes can experience a
      slowdown in the file open due to filesystem contention.

      As a first step in improving the startup/file-open performance,
      we allow MPI rank 0 of the associated MPI communicator to locate
      the base offset of the super-block and then broadcast that result
      to the remaining ranks in the parallel group.  Note that this
      approach is utilized ONLY during file opens which employ the MPIO
      file driver in HDF5 by previously having called H5Pset_fapl_mpio().

      HDF5 parallel file operations which do not employ multiple ranks
      e.g. specifiying MPI_COMM_SELF (whose MPI_Comm_size == 1)
      as opposed to MPI_COMM_WORLD, will not be affected by this
      optimization.  Conversely, parallel file operations on subgroups
      of MPI_COMM_WORLD are allowed to be run in parallel with each
      subgroup operating as an independant collection of processes.

      (RAW - 2017/10/10, HDFFV-10294)

    - Added large (>2GB)  MPI-IO transfers.

      Previous releases of PHDF5 would fail when attempting to
      read or write greater than 2GB of data in a single IO operation.
      This issue stems principally from an MPI API whose definitions
      utilize 32 bit integers to describe the number of data elements
      and datatype that MPI should use to effect a data transfer.
      Historically, HDF5 has invoked MPI-IO with the number of
      elements in a contiguous buffer represented as the length
      of that buffer in bytes.

      Resolving the issue and thus enabling larger MPI-IO transfers
      is accomplished first, by detecting when a user IO request would
      exceed the 2GB limit as described above.  Once a transfer request
      is identified as requiring special handling, PHDF5 now creates a
      derived datatype consisting of a vector of fixed sized blocks
      which is in turn wrapped within a single MPI_Type_struct to
      contain the vector and any remaining data.  The newly created
      datatype is then used in place of MPI_BYTE and can be used to
      fulfill the original user request without encountering API
      errors.

      (RAW - 2017/09/10, HDFFV-8839)

    C++ Library:
    ------------
    - The following C++ API wrappers have been added to the C++ Library:
      + H5Lcreate_soft:
        // Creates a soft link from link_name to target_name.
        void link(const char *target_name, const char *link_name,...)
        void link(const H5std_string& target_name,...)

      + H5Lcreate_hard:
        // Creates a hard link from new_name to curr_name.
        void link(const char *curr_name, const Group& new_loc,...)
        void link(const H5std_string& curr_name, const Group& new_loc,...)

        // Creates a hard link from new_name to curr_name in same location.
        void link(const char *curr_name, const hid_t same_loc,...)
        void link(const H5std_string& curr_name, const hid_t same_loc,...)

        Note: previous version of H5Location::link will be deprecated.

      + H5Lcopy:
        // Copy an object from a group of file to another.
        void copyLink(const char *src_name, const Group& dst,...)
        void copyLink(const H5std_string& src_name, const Group& dst,...)

        // Copy an object from a group of file to the same location.
        void copyLink(const char *src_name, const char *dst_name,...)
        void copyLink(const H5std_string& src_name,...)

      + H5Lmove:
        // Rename an object in a group or file to a new location.
        void moveLink(const char* src_name, const Group& dst,...)
        void moveLink(const H5std_string& src_name, const Group& dst,...)

        // Rename an object in a group or file to the same location.
        void moveLink(const char* src_name, const char* dst_name,...)
        void moveLink(const H5std_string& src_name,...)

        Note: previous version H5Location::move will be deprecated.

      + H5Ldelete:
        // Removes the specified link from this location.
        void unlink(const char *link_name,
            const LinkAccPropList& lapl = LinkAccPropList::DEFAULT)
        void unlink(const H5std_string& link_name,
            const LinkAccPropList& lapl = LinkAccPropList::DEFAULT)

        Note: additional parameter is added to previous H5Location::unlink.

      + H5Tencode and H5Tdecode:
        // Creates a binary object description of this datatype.
        void DataType::encode() - C API H5Tencode()

        // Returns the decoded type from the binary object description.
        DataType::decode() - C API H5Tdecode()
        ArrayType::decode() - C API H5Tdecode()
        CompType::decode() - C API H5Tdecode()
        DataType::decode() - C API H5Tdecode()
        EnumType::decode() - C API H5Tdecode()
        FloatType::decode() - C API H5Tdecode()
        IntType::decode() - C API H5Tdecode()
        StrType::decode() - C API H5Tdecode()
        VarLenType::decode() - C API H5Tdecode()

      + H5Lget_info:
        // Returns the information of the named link.
        H5L_info_t getLinkInfo(const H5std_string& link_name,...)

      (BMR - 2018/03/11, HDFFV-10149)

    - Added class LinkCreatPropList for link create property list.

      (BMR - 2018/03/11, HDFFV-10149)

    - Added overloaded functions H5Location::createGroup to take a link
      creation property list.
        Group createGroup(const char* name, const LinkCreatPropList& lcpl)
        Group createGroup(const H5std_string& name, const LinkCreatPropList& lcpl)

      (BMR - 2018/03/11, HDFFV-10149)

    - A document is added to the HDF5 C++ API Reference Manual to show the
      mapping from a C API to C++ wrappers.  It can be found from the main
      page of the C++ API Reference Manual.

      (BMR - 2017/10/17, HDFFV-10151)

    Java Library:
    ----------------
    - Wrapper added for enabling the error stack.

      H5error_off would disable the error stack reporting. In order
      to re-enable the reporting, the error stack info needs to be
      saved so that H5error_on can revert state.

      (ADB - 2018/03/13, HDFFV-10412)

    - Wrappers were added for the following C APIs:
      H5Pset_evict_on_close
      H5Pget_evict_on_close
      H5Pset_chunk_opts
      H5Pget_chunk_opts
      H5Pset_efile_prefix
      H5Pget_efile_prefix
      H5Pset_virtual_prefix
      H5Pget_virtual_prefix

      (ADB - 2017/12/20)

    - The H5I_REFERENCE value in the H5I_type_t enum (defined in H5Ipublic.h)
      has been marked as deprectated.

      JNI code which refers to this value will be removed in a future
      major version of the library. The code will remain unchanged in the
      1.10.x releases and branches.

      See the C library section, above, for further information.

      (HDFFV-10252, DER, 2017/04/05)

    Tools:
    ------
    - h5diff has a new option to display error stack.

      Updated h5diff with the --enable-error-stack argument, which
      enables the display of the hdf5 error stack. This completes the
      improvement to the main tools: h5copy, h5diff, h5dump, h5ls and
      h5repack.

      (ADB - 2017/08/30, HDFFV-9774)

Support for new platforms, languages and compilers.
=======================================
    - None

(wen)

Updated net/py-smb to 1.1.23

(wen)

Update to 1.1.23
Add missing DEPENDS

Upstream changes:
Version 1.1.23, 2018-05-05 (Download)

Fix bug in listShares() method which fails when the remote server has many shares. Contributed by carlosefr (github).
Improve echo() method to test and fail if the provided data to echo is not a bytes object. Contributed by carlosefr (github).
Fix bug in listPath() method where the path to query is not properly terminated. Contributed by Yepoleb (github).
Version 1.1.22, 2017-09-17 (Download)

Fix bug in getAttributes() method which should return only the filename instead of the entire path for the filename property for the return result.
Version 1.1.21, 2017-09-09 (Download)

Fix bug where timestamp values for SMB1 getAttributes() response are not converted properly from FILETIME to epoch time values.
Version 1.1.20, 2017-08-13 (Download)

Add getSecurity() method to support security descriptors query via SMB2. Contributed by koniiiik (github).
Improve retrieveFile() and retrieveFileFromOffset() methods to allow file retrievals over SMB2 even when the file is being locked on the server.
Silently discards NMB SESSION_KEEPALIVE packets instead of raising warnings. Thanks to a-mushroom (github) for reporting this.
SMB sessionID will be sent in ECHO requests to conform to SMB2 specs. Thanks to divad (github) for reporting this.
Fix type errors for MD4 functions in python3. Contributed by viatoriche (github).
Version 1.1.19, 2016-11-13 (Download)

Ignore STATUS_PENDING during delete and file store operations
Version 1.1.18, 2016-04-09 (Download)

Rollback fixes to NTLMv2 response algorithm in pysmb 1.1.17. The fixes fail to work with some servers.
Add missing errno imports in SMBConnection.py
Fix UnboundLocalError raised when using type() in SMBConnection.py
Version 1.1.17, 2015-11-11 (Download)

Fix crashes in directory listing with keyerror 窶�support_dfs窶�.
Fix bugs in NTLMv2 response algorithm.
Fix bugs where client domain is not included as part of the session negotiation.
Version 1.1.16, 2015-05-10 (Download)

Fix typo errors in authentication error messages.
Improve share listings on SMB2 protocol by ignoring interim STATUS_PENDING responses.
Version 1.1.15, 2015-02-15 (Download)

Add new parameter to SMBConnection窶冱 storeFileFromOffset method to determine whether the remote file is to be truncated before writing.
Version 1.1.14, 2015-02-01 (Download)

Add support for DFS shares in listPath(). Thanks to humberry (github) for raising the issue and helping with the testing.
Fix bug in python3窶冱 SMB2 listPath implementation. Thanks to deejrose (github) for reporting the bug.
Version 1.1.13, 2014-10-18 (Download)

Add missing methods and improve compatibility with python3. Thanks to keisetsu (github) for submitting the patch.
Fix bug in SMB2 rename implementation which fails to rename directory. Thanks to Jayke Meijer for raising the bug and providing the packet capture.
Version 1.1.12, 2014-09-21 (Download)

Fix syntax error for python3 NBNSProtocol implementation
Fix bug in SMB1 implementation which results in access denied errors with Samba 3.0. Many thanks for John Sivak for his assistance and support in helping to troubleshoot and test the bug fixes.
Version 1.1.11, 2014-09-13 (Download)

Add support for unicode characters in domain, username and password.
Add storeFileFromOffset method to SMB API
Fix bug in getAttributes implementation for SMB1
Fix bug for NMB which uses broadcast flag for unicast queries
Update the Tree Connect Andx request implementation to support MS-SMB 2.2.4.7.1 extensions
Version 1.1.10, 2014-06-29 (Download)

Add getAttributes() method to SMBConnection and SMBProtocolFactory class.
Add isReadOnly property to SharedFile class.
Version 1.1.9, 2014-06-01 (Download)

Add support for domains in smb:// URLs. Contributed by Andy Piper.
Fix a bug which fails to test for the correct GSS security provider OID values. Thanks for Fanen for bug report and assistance in testing the fix.
Version 1.1.8, 2013-12-21 (Download)

Fix a bug in storeFile() method when the destination file is not overwritten if it exists in SMB1 communication. Thanks to Vaikar Amol for reporting this bug and helping to fix it.
Fix a SMB1 authentication problem when extended negotation is not carried out because the remote server has specified its support for extended security in the payload, instead of in the message flags2.
Version 1.1.7, 2013-09-27 (Download)

Improve listShares() function which can fail with the listing response is separated into multiple SMB packets for large number of shares. Thanks to Pieter De Clerck for reporting this bug and helping to test the bug fix.
Fix bug in python3 implementation where session connection can fail when remote server supports message signing. Thanks to Simon for reporting this bug.
Version 1.1.6, 2013-08-16 (Download)

Fix bug where the status of the SMB_COM_NEGOTIATE reply is not checked for error before allowing further processing. Thanks to Adrian Cox for discovering this bug and submitting the patch.
Version 1.1.5, 2013-06-19 (Download)

Add support for Direct hosting of SMB over TCP/IP (TCP port 445)
Version 1.1.4, 2013-05-31 (Download)

Improve query performance for query IP addresses for NetBIOS names in NetBIOSProtocol.py. Thanks to Ties de Kock for submitting the performance patch.
Fix bugs in SMBConnection when sending large data packets can result in AssertionError. Thanks to David K. Hess for submitting the bug fix.
Version 1.1.3, 2013-03-18 (Download)

Fix a bug which results in endless loop in SMBConnection when remote CIFS server closes the connection. Thanks to Nitin Garg for submitting the bug fixes.

(wen)

Updated misc/p5-Business-ISSN to 1.003

(wen)

Update to 1.003
Update LICENSE

Upstream changes:
1.003 2018-05-09T20:27:50Z
* Clarify that it's the Artistic License 2.0

1.002 2016-12-05T18:59:45Z
* Remove the prereq.t test. I'll do that locally.

1.001 2016-01-12T18:51:29Z
* Freshen the module

(wen)

Updated net/p5-SNMP-Info to 3.61

(wen)

Update to 3.61
Update DEPENDS

Upstream changes:
Version 3.61 (2018-05-09)

  [ENHANCEMENTS]

  * #255 IPv6 support - Set the transport-specifier if given an IPv6 address
  * #195 IP address table - IPv4 Address Table will use the
    IP-MIB::ipAddressTable if the deprecated IP-MIB::ipAddrTable doesn't
    return results

  [BUG FIXES]

  * #261 EIGRP Peer Neighbor Formatting / Munge
  * #252 Unpack binary MAC if present in cdp_port
  * Fix SNMPv1 cdp_run check

(wen)

Updated textproc/p5-YAML to 1.25

(wen)

Update to 1.25
Update BUILD_DEPENDS

Upstream changes:
Changes for version 1.25 - 2018-05-11
Applied several pull requests by tinita
Support trailing comments (PR #189, PR #190, PR #191)
Remove unused code (PR #192)
Use Test::Deep to actually test correctly for class names (PR #193)
Fix loading of mapping key which starts with `= ` (PR #194)
Fix loading strings with multiple spaces (PR #172)
Allow more characters in anchor name (PR #196)
Add $YAML::LoadBlessed for disabling loading objects (PR #197)
Applied several pull requests by tinita
Disable test with long string under certain conditions (PR #201)
Quote scalar if it equals '=' (PR #202)
Multiple regexp roundtrip does not grow (PR #203)
Add support for compact nested block sequences (PR #204)
Support reverse order of block scalar indicators (PR #205)
Support nested mappings in sequences (PR #206)

(wen)

Updated textproc/p5-XML-Compile-Tester to 0.91

(wen)

Update to 0.91

Upstream changes:
version 0.91: Fri May 11 01:51:33 CEST 2018

        Improvements:
        - change documentation style.
        - convert to GIT and GitHub

(wen)

Updated textproc/p5-XML-Compile-SOAP12 to 3.06

(wen)

Update to 3.06

Upstream changes:
version 3.06: Fri May 11 00:29:18 CEST 2018

        Improvements:
        - convert to GIT and GitHub

(wen)

Updated www/drupal8 to 8.5.3

(wen)

Update to 8.5.3

Upstream changes:
Drupal-8.5.3:

Release notes
Maintenance and security release of the Drupal 8 series.

This release fixes critical security vulnerabilities. Sites are urged to upgrade immediately after reading the notes below and the security announcement:

Drupal core - Critical - Remote Code Execution - SA-CORE-2018-004
No other fixes are included.

Drupal-8.5.2:
This release fixes security vulnerabilities. Sites are urged to upgrade immediately after reading the notes below and the security announcement:

Drupal core - Moderately Critical - Cross Site Scripting- SA-CORE-2018-003
No changes have been made to the .htaccess, web.config, robots.txt or default settings.php files in this release, so upgrading custom versions of those files is not necessary.

(wen)

Updated sysutils/coreutils to 8.29

(wen)

Update to 8.29

Upstream changes:
2017-12-27  Pádraig Brady  <P@draigBrady.com>

        version 8.29
        * NEWS: Record release date.

2017-12-23  Pádraig Brady  <P@draigBrady.com>

        tests: avoid false failure on AIX 7.2
        * tests/tail-2/pipe-f.sh: Close stdout in a subshell
        to ensure the current shell isn't impacted.  Subsequent
        piped commands like `echo foo | blah` were seen to fail
        due to the previous closing of stdout.
        Reported by Assaf Gordon.

        doc: describe recent build checks for 32 bit time_t
        * README: Document the new handling of 32 bit time_t,
        with examples of how to build in 64 bit mode on AIX.
        Also mention that GNU make is desired on AIX
        due to its mishandling of the "[" target.
        Suggested by Assaf Gordon.

2017-12-21  Pádraig Brady  <P@draigBrady.com>

        tests: fix recent portability issues on solaris 10
        * tests/misc/ptx.pl: Escape the '^' character which is
        otherwise considered as a line continuation character.
        * tests/misc/shred-remove.sh: sed doesn't support \n.

        maint: remove reference to excluded changelog item
        * build-aux/git-log-fix: Remove old entry.

2017-12-20  Pádraig Brady  <P@draigBrady.com>

        maint: add doc/coverage to .gitignore
        * .gitignore: Ignore the generated coverage report.

        doc: remove older ChangeLog items
        * Makefile.am: Update the oldest documented version
        to 8.20 which is now about 5 years old.

2017-12-18  Bernhard Voelker  <mail@bernhard-voelker.de>

        doc: mention which privileges are needed to chmod
        POSIX specification for chmod(1):
        https://pubs.opengroup.org/onlinepubs/9699919799/utilities/chmod.html

        * doc/coreutils.texi (chmod invocation): Add a sentence about who can
        change the file mode bits of a file - (almost) a copy from what POSIX
        requires.

        Fixes https://bugs.gnu.org/29207.

2017-12-16  Pádraig Brady  <P@draigBrady.com>

        tests: fix recent regressions with dash
        * tests/misc/timeout.sh: dash outputs the "Killed"
        message to stderr rather than the terminal.
        * tests/misc/usage_vs_getopt.sh: dash doesn't yet
        support the POSIX proposed $'...' shell quoting syntax.

        build: avoid a signed overflow warning in ptx
        * src/ptx.c (fix_output_parameters): GCC 6.3.1 with
        ./configure --enable-single-binary would give:
          error: assuming signed overflow does not occur
          when simplifying conditional to constant [-Werror=strict-overflow]
            if (file_index > 0)
        So change the type of file_index to signed (size_t).

2017-12-11  Bernhard Voelker  <mail@bernhard-voelker.de>

        maint: adjust for the renamed nstrfime gnulib module
        * bootstrap.conf: s/strftime/nstrfrime/.

2017-12-11  Pádraig Brady  <P@draigBrady.com>

        build: update gnulib submodule to latest
        * gnulib: Update with various build/test fixes.

        tests: fix false failure in new dd/nocache_eof test
        * test/dd/nocache_eof.sh: Also handle fadvise64_64 which is
        used on 32 bit x86.  Note strace internally maps fadvise64_64
        to {arm,xtensa}_fadvise64_64.

        tail: fix tailing non seekable files on certain systems
        * src/tail.c (tail_bytes): On systems were blksize_t is unsigned
        and the same size or wider than off_t (android for example),
        our initialized (off_t) -1 would be promoted to unsigned before
        comparison, and thus fail to follow the appropriate path.
        * tests/tail-2/tail-c.sh: Add a test case.
        * NEWS: Mention the fix.
        This issue was introduced in commit v8.23-47-g2662702
        Reported at https://github.com/termux/termux-app/issues/233

        build: avoid build failure without sys/mtio.h
        * m4/jm-macros.m4: Check for the header.
        * src/dd.c: Avoid the workaround where the header
        is not available (on non glibc systems).
        * src/shred.c: Likewise.

        doc: reorganize ls -k and --time-style help
        * src/ls.c (usage): Clarify -k only applies to -s usage
        and directory 'total' lines.  Move the description
        of TIME_STYLE out of the option section as it was awkward
        to read and write there within 80 columns.

2017-12-10  Pádraig Brady  <P@draigBrady.com>

        doc: clarify numeric setuid handling in chmod man page
        * man/chmod.x: Update the information to state one can
        clear the setuid and setgid bits for directories numerically
        using an additional leading '0' or a leading '='.
        That has been supported since v8.15-64-g8931cdb.
        Fixes https://bugs.gnu.org/29390

        doc: shred: change 'truncate' to the more descriptive 'deallocate'
        * doc/coreutils.texi (shred invocation): s/truncate/deallocate/.
        * src/shred.c (usage): Likewise.
        Fixes https://bugs.gnu.org/29317

        doc: clarify that cp --force may recreate files
        * doc/coreutils.texi (cp invocation): The language used
        to describe recreating the file was a little confusing
        as it mentioned opening a removed file.
        Fixes https://bugs.gnu.org/29315

2017-12-04  Kamil Dudka  <kdudka@redhat.com>

        doc: fix default QUOTING_STYLE for %N format of stat(1)
        * doc/coreutils.texi (stat invocation): The default value
        of QUOTING_STYLE for the %N format of 'stat --printf' is
        'shell-escape-always'.
        Fixes https://bugs.gnu.org/29563
        Reported by Christian Groessler at
        https://bugzilla.redhat.com/1520399#c3

2017-12-02  Jean Delvare  <jdelvare@suse.de>

        tests: make ls/block-size more readable
        * tests/ls/block-size.sh: The output of the test was hard to read. Add
        comments saying what we are testing to make it easier to understand.

2017-11-29  Bernhard Voelker  <mail@bernhard-voelker.de>
            Pádraig Brady  <P@draigBrady.com>

        tests: verify usage vs. getopt
        Verify that all options mentioned in usage are actually recognized
        by the program.

        * tests/misc/usage_vs_getopt.sh: Add test.
        * tests/local.mk (all_tests): Reference it.

2017-11-29  Pádraig Brady  <P@draigBrady.com>

        readlink: remove superfluous comma from usage output
        * src/readlink.c (usage): Remove ',' after --quiet option.

2017-11-29  Bernhard Voelker  <mail@bernhard-voelker.de>

        all: use consistent diagnostics for unknown long options
        Previously, e.g. cksum failed to output the offending unknown long
        option:
          $ cksum --unknown-opt
          cksum: invalid option -- '-'
          Try 'cksum --help' for more information.
        i.e., it tried to diagnose '-' as short option.
        Instead, it should diagnose the unknown long option:
          $ cksum --unknown-opt
          cksum: unrecognized option '--unknown-opt'
          Try 'cksum --help' for more information.

        * src/cksum.c (long_options): Add struct with null entry only.
        (main): Use it in the getopt_long call.
        * src/dd.c: Likewise.
        * src/hostid.c: Likewise.
        * src/hostname.c: Likewise.
        * src/link.c: Likewise.
        * src/logname.c: Likewise.
        * src/nohup.c: Likewise.
        * src/sleep.c: Likewise.
        * src/tsort.c: Likewise.
        * src/unlink.c: Likewise.
        * src/uptime.c: Likewise.
        * src/users.c: Likewise.
        * src/whoami.c: Likewise.
        * src/yes.c: Likewise.
        * NEWS (Improvements): Mention the fix.

2017-11-29  Pádraig Brady  <P@draigBrady.com>

        test: fix issues with tests/cp/preserve-mode.sh
        * tests/cp/preserve-mode.sh: This was the only use of awk,
        which may not be available on the system resulting
        in an ineffective test.  Also the permissions bits for
        directories were not being checked at all.

        build: update gnulib submodule to latest
        * gnulib: Update with various build/test fixes.

2017-11-28  Pádraig Brady  <P@draigBrady.com>

        build: update gnulib submodule to latest
        * gnulib: Update including various build fixes.

2017-11-27  Bernhard Voelker  <mail@bernhard-voelker.de>

        timeout: also support short -v option
        * src/timeout.c (main): Add short option character 'v' to getopt_long
        call.
        * tests/misc/timeout.sh: Run the test both for the long and the short
        option.

2017-11-25  Pádraig Brady  <P@draigBrady.com>

        dd: support iflag=direct with arbitrary sized files
        * src/dd.c (iread): Handle read error with a non-aligned
        file offset in the O_DIRECT case.  This is not an issue
        on XFS at least, but on EXT4 the final read will return
        EINVAL rather than the expected 0 to indicate EOF.
        * tests/dd/direct.sh: Test the iflag=direct case also.
        * NEWS: Mention the improvement.

2017-11-24  Pádraig Brady  <P@draigBrady.com>

        timeout: add --verbose to diagnose timeouts
        This is useful as handling in shell is complicated
        with the varying exit status in the --kill-after case.

        * src/timeout.c (main): Handle '-v' and store
        COMMAND for the diagnostic.
        (cleanup): Diagnose the signal name before sending.
        (usage): Document -v, --verbose.
        * doc/coreutils.texi (timeout invocation): Likewise.
        * tests/misc/timeout.sh: Add a test case.
        * NEWS: Mention the new feature
        Fixes https://bugs.gnu.org/21760

2017-11-19  Pádraig Brady  <P@draigBrady.com>

        tail: seek to the end of block devices
        * src/tail.c (tail_bytes): Try lseek(..., SEEK_END) when
        we can't determine the file size.
        * tests/tail-2/end-of-device.sh: Add a new root only test.
        * tests/local.mk: Reference the new test.
        * NEWS: Mention the improvement.
        Paul Eggert suggested using lseek() (rather than ioctl(BLKGETSIZE64)).
        Fixes https://bugs.gnu.org/29259

2017-11-14  Bernhard Voelker  <mail@bernhard-voelker.de>

        maint: include the module year2038 from gnulib
        * bootstrap.conf (gnulib_modules): Add 'year2038' to ensure that time_t
        is 64-bit (and thus works after 2038).

        Suggested by Bruno Haible in
        lib/2017-11/msg00022.html">https://lists.gnu.org/r/bug-gnulib/2017-11/msg00022.html

2017-11-14  Bernhard Voelker  <mail@bernhard-voelker.de>

        maint: update gnulib to latest
        * gnulib: Update - mainly for the recent year2038 changes.
        * tests/init.sh: Update from gnulib/tests/init.sh.

2017-11-09  Assaf Gordon  <assafgordon@gmail.com>

        doc: add github issue/pull-request templates
        These templates instruct contributors not to use github, and instead
        use the upstream GNU development resources. Discussed in
        http://lists.gnu.org/archive/html/coreutils/2017-11/msg00007.html .

        * .github/ISSUE_TEMPLATE.txt,
          .github/PULL_REQUEST_TEMPLATE.txt: New files.

2017-11-08  Jim Meyering  <meyering@fb.com>

        maint: make hook script reject "/archive/html" in lists.gnu.org URLS
        * scripts/git-hooks/commit-msg: Require the abbreviated "/r/"
        form in any log message URL.

        maint: shorten https://lists.gnu.org/archive/html/... links
        Each /archive/html/ part can be replace with /r/.
        Run this to induce the change:
        git grep -l archive/html|xargs perl -pi -e 's,/archive/html/,/r/,g'
        * TODO: Perform that substitution.
        * bootstrap: Likewise.
        * src/sort.c (sequential_sort): Likewise.
        * src/tail.c (tail_file): Likewise.
        * tests/misc/sort-merge-fdlimit.sh: Likewise.
        * tests/misc/stty-row-col.sh: Likewise.
        * tests/misc/unexpand.pl: Likewise.
        * tests/rm/readdir-bug.sh: Likewise.
        * tests/tail-2/inotify-rotate.sh: Likewise.

2017-11-07  Thomas Deutschmann  <whissi@gentoo.org>

        tests: avoid false failure with inaccessible mount points
        * tests/ls/readdir-mountpoint-inode.sh: Skip the test
        if any mount points are inaccessible by the current user.
        Fixes https://bugs.gnu.org/29167
        Reported at: https://bugs.gentoo.org/353164

2017-11-06  Bernhard Voelker  <mail@bernhard-voelker.de>

        doc: fix "Up" field of realpath usage examples
        Older versions of 'makeinfo' choke on a missing reference:

          ./doc/coreutils.texi:14177: `Realpath usage examples' has no Up field\
          (perhaps incorrect sectioning?).
          makeinfo: Removing output file `doc/coreutils.info' due to errors; \
          use --force to preserve.

        * doc/coreutils.texi (realpath invocation): Add a menu referencing
        the usage examples - introduced in v8.27-91-g7449f0d.

2017-11-06  Pádraig Brady  <P@draigBrady.com>

        maint: ensure https:// URLs are used in --help and man pages
        * configure.ac(AC_INIT): Specify the URL explicitly, so we're
        not dependent on unreleased autoconf.

2017-10-31  Assaf Gordon  <assafgordon@gmail.com>

        stat: output default formats for --terse in usage
        Suggested by L A Walsh in https://bugs.gnu.org/28763 .

        * src/stat.c (fmt_terse_fs): Define format for --terse -f here.
        (fmt_terse_regular): Define format for --terse here.
        (fmt_terse_selinux): Likewise for when SELinux is enabled.
        (default_format): Use the above constants.
        (usage): Output the formats for the terse modes.

2017-10-30  Pádraig Brady  <P@draigBrady.com>

        df: fix hang with fifo argument
        * src/df.c (main): stat() before open(), and avoid
        the optional open when given a fifo argument.
        * tests/df/unreadable.sh: Add a test case.
        * NEWS: Mention the fix.
        Fixes https://bugs.gnu.org/29038

2017-10-28  Jim Meyering  <meyering@fb.com>

        build: ls.c: apply _GL_ATTRIBUTE_PURE to more functions
        * src/ls.c (DEFINE_SORT_FUNCTIONS): Apply _GL_ATTRIBUTE_PURE
        to each strcmp-derived function definition, since GCC8 with
        -Wsuggest-attribute=pure now warns it is needed.

2017-10-26  Vincent Lefevre  <vincent@vinc17.net>

        doc: reference statfs(2) in the stat(1) man page
        * man/stat.x (SEE ALSO): Mention statfs(2) in addition to stat(2).
        Note statfs() is generally used rather than statvfs(),
        so we'll defer that reference to the SEE ALSO section of statfs(2).
        Fixes https://bugs.gnu.org/28989

2017-10-25  Pádraig Brady  <P@draigBrady.com>

        tests: avoid false failure when O_DIRECT isn't supported
        * tests/dd/nocache_eof.sh: Only run the O_DIRECT tests
        when 512 byte alignment is supported.  Otherwise with older
        XFS on systems with > 1MiB pages, or on file systems not
        supporting O_DIRECT, there would have been false failures.
        * tests/dd/direct.sh: Clarify the skip message.

2017-10-25  Pádraig Brady  <P@draigBrady.com>

        dd: fix nocache regions passed to posix_fadvise()
        Previously with oflag=direct the call to invalidate_cache()
        was not passed to the kernel, as it was less than a page size,
        and a subsequent call was not made to invalidate the pending space.
        Similarly with oflag=nocache the pending space at EOF was
        not invalidated.  Even though these amount to only a single page
        in the page cache it can be significant.  For example on
        XFS before kernel patch v4.9-rc1-4-g0ee7a3f, O_DIRECT files
        would have been read inefficiently if any pages were cached,
        even if they were already synced to storage.

        * src/dd.c (i_nocache_eof, o_nocache_eof): New bools used
        to control when we want invalidate_cache(,0) to clear to EOF.
        (cache_round): Use IO_BUFSIZE (currently 132KiB) to minimize
        calls to the relatively expensive advise function, rather
        than page_size.  This also makes it clear that while the
        kernel function operates on pages, this size is chosen for
        performance reasons.
        (invalidate_cache): Refactor to share more code between
        input and output paths. Use i_nocache_eof and o_nocache_eof
        rather than proxying off max_records.  Ensure we
        invalidate full pages when clearing to EOF as the kernel
        will ignore any non complete pages.  Fix the offset used
        for the output path.
        (dd_copy): Invalidate the cache of the input after the
        offset is updated, for consistency and so we don't try to
        invalidate before the start of the file.  When we read
        EOF on input, set flags so that we invalidate to EOF.
        (main): Invalidate to EOF in more cases, by depending
        on the i_nocache_eof and o_nocache_eof flags.
        * doc/coreutils.texi (dd invocation): Clarify the alignment
        and persisted caveats on the example applying "nocache"
        to part of a file.
        * tests/dd/nocache_eof.sh: A new test.
        * tests/local.mk: Reference the new test.
        * NEWS: Mention the bug fix.
        Issue reported by Eric Bergen.

2017-10-24  Michael Stone  <mstone@debian.org>

        doc: mention QUOTING_STYLE env var in ls man page
        * src/ls.c (usage): Mention QUOTING_STYLE with the --quoting-style
        option, and indicate it has lower precedence than that option.

2017-10-24  Pádraig Brady  <P@draigBrady.com>

        maint: apply suggested cleanup to recent stty.c change
        This should have been part of commit v8.28-17-gf926f7c
        * src/stty.c (check_argument): Align line continuation chars,
        and ensure the function macro is immune to usage with if/else.
        Suggested by Jim Meyering and Paul Eggert.

        b2sum: fix crash with --check and truncated input
        * src/md5sum.c (split_3): Ensure we don't walk off
        the end of the string.
        * tests/misc/b2sum.sh: Add test cases.
        Fixes https://bugs.gnu.org/28860

2017-10-24  Pádraig Brady  <P@draigBrady.com>

        stty: fix processing of options when -F is specified
        This was a latent issue that became significant with
        the addition of the -F option in FILEUTILS-3_16n-56-ge46a424

        * src/stty.c (apply_settings): Refactor argument checking
        to a function macro.  Augment the argument check to ignore
        NULLed out arguments (already processed -F).
        * NEWS: Mention the fix.
        * tests/misc/stty-invalid.sh: Add a test case.
        Fixes https://bugs.gnu.org/28859

2017-10-24  Pádraig Brady  <P@draigBrady.com>

        timeout: fix a small race that would ignore command exit
        This fixes a regression from commit v8.26-39-g2f69dba

        * src/timeout.c (block_cleanup_and_chld): Rename from block_cleanup
        to indicate we also block SIGCHLD to avoid the race where SIGCHLD
        fires between waitpid() polling and sigsuspend() waiting for a signal.
        * NEWS: Mention the fix.

2017-10-24  Thomas Jarosch  <thomas.jarosch@intra2net.com>

        timeout: fix regression when invoked with blocked SIGCHLD
        We inherit the signal mask from our parent process,
        therefore ensure SIGCHLD is not blocked.

        If SIGCHLD is blocked, sigsuspend() won't be interrupted
        when the child process exits and we hang until the timeout (SIGALRM).

        This fixes a regression from commit v8.26-39-g2f69dba

        * src/timeout.c (install_sigchld): Ensure SIGCHLD is unblocked.
        * NEWS: Mention the issue.

2017-10-02  Pádraig Brady  <P@draigBrady.com>

        build: reinstate distribution of man pages
        man pages change little between systems,
        so falling back to distributed pages make sense
        when cross compiling or lacking perl.

        * man/local.mk: Add all man pages to EXTRA_DIST
        so that they're distributed in the generated tarball.
        Use the dummy-man page generator if cross compiling.
        Set TZ to avoid a distcheck failure where man pages
        used a diffent month than those rebuilt (with a .timestamp).
        * man/dummy-man: Only fall back to generating a stub
        if copying an existing man page fails.
        * man/help2man: Sync portable TZ=UTC0 specification
        from upstream help2man.
        * NEWS: Mention the build-related change.
        Fixes https://bugs.gnu.org/28574

2017-10-02  Pádraig Brady  <P@draigBrady.com>

        maint: remove a duplicate entry from THANKS
        * .mailmap: Prefer Colin Watson's last used email address.

2017-09-25  Paul Eggert  <eggert@cs.ucla.edu>

        copy: revert recent patch for vulnerable dirs
        I plan to propose a better patch to catch vulnerable parent
        directories.
        * NEWS, doc/coreutils.texi (Target directory): Document this.
        * src/cp.c, src/install.c, src/ln.c, src/mv.c:
        Do not include targetdir.h.
        (target_directory_operand): Remove test for vulnerable parents.
        * src/cp.c (stat_target_operand): Remove.  All uses removed.
        * src/local.mk (noinst_HEADERS): Remove src/targetdir.h.
        (src_ginstall_SOURCES, src_cp_SOURCES, src_ln_SOURCES)
        (src_mv_SOURCES): Remove src/targetdir.c.
        * src/targetdir.c, src/targetdir.h: Remove.
        * tests/mv/vulnerable-target.sh: Remove.
        * tests/local.mk (all_root_tests): Remove it.

2017-09-24  Pádraig Brady  <P@draigBrady.com>

        tests: fix test hang on case insenitive file systems
        * tests/split/filter.sh: Due to an invalid 'FILE = zero.in'
        construct trying to initialize a FILE variable, it would
        instead try to run the FILE command which is present on
        macOS 10.13 with APFS.
        We also remove a redundant duplicate test clause introduced
        during a rebase, and simplify the piped timeout command,
        to avoid requiring a subshell and associated quoting.
        * THANKS.in: Add the reporter Jack Howarth.
        Fixes https://bugs.gnu.org/28506

2017-09-21  Pádraig Brady  <P@draigBrady.com>

        tests: avoid a false failure in expr test with UTF8
        * tests/misc/expr.pl: Skip the quote varying tests in
        the multi-byte locales as these tests aren't that interesting
        in those locales.  Also ERR_SUBST is already defined for
        some tests so awkward to redefine to munge UTF8 quotes to ASCII.

2017-09-20  Assaf Gordon  <assafgordon@gmail.com>

        expr: add detailed syntax error messages
        Show offending argument instead of a generic 'syntax error' message.
        Suggested by Bernhard Voelker in https://bugs.gnu.org/28461#13 .

        * src/expr.c (syntax_error): Remove.
        (required_more_args): New function.
        (eval7, main): Replace syntax_error call with detailed die message.
        * tests/misc/expr.pl: Add tests for new messages.

2017-09-20  Pádraig Brady  <P@draigBrady.com>

        maint: fix new syntax-check failures from HTTPS adjustments
        * cfg.mk [old_NEWS_hash]: update with `make update-NEWS-hash`.
        [sc_long_lines]: Avoid flagging (long) URLs in NEWS.
        * src/sort.c: Tweak to a shorter line.
        * src/tail.c: Likewise.
        Introduced in v8.28-4-gbe87d61

        maint: fix new syntax check failures from copy restrictions
        * doc/coreutils.texi: Remove doubled word.
        * src/targetdir.c: Explicitly mark exported function.
        * tests/local.mk: This is not a root only test.
        * tests/mv/vulnerable-target.sh: Use returns_.
        Introduced in v8.28-3-g44ccd1c

        shred: reinstate --remove file name length obfuscation
        This was unintentionally removed in v8.27-60-g2ae1460
        * src/shred.c (wipename): Interate through all name lengths.
        * tests/misc/shred-remove.sh: Add test cases.
        * NEWS: Mention the bug fix.
        Fixes https://bugs.gnu.org/28507

2017-09-19  Paul Eggert  <eggert@cs.ucla.edu>

        maint: copy bootstrap from Gnulib

        all: prefer HTTPS in URLs

        copy: check for vulnerable target dirs
        * NEWS, doc/coreutils.texi (Target directory): Document this.
        * src/cp.c, src/install.c, src/ln.c, src/mv.c: Include targetdir.h.
        (target_directory_operand): Use the new targetdir_operand_type
        function to check for vulnerable target directories.
        * src/cp.c (stat_target_operand): New function.
        (target_directory_operand, do_copy): Use it.
        * src/local.mk (noinst_HEADERS): Add src/targetdir.h.
        (src_ginstall_SOURCES, src_cp_SOURCES, src_ln_SOURCES)
        (src_mv_SOURCES): Add src/targetdir.c.
        * src/targetdir.c, src/targetdir.h: New files.
        * tests/mv/vulnerable-target.sh: New test.
        * tests/local.mk (all_root_tests): Add it.

2017-09-14  Bernhard Voelker  <mail@bernhard-voelker.de>

        ptx: avoid infloop due to zero-length matches with -S regex
        * src/ptx.c (find_occurs_in_text): Die with an appropriate error
        diagnostic when the given regular expression returns a match of
        length 0.
        * tests/misc/ptx.pl (S-infloop): Add a test.
        * NEWS (Bug fixes): Mention the fix.

        Fixes https://bugs.gnu.org/28417 which was detected using
        Symbolic Execution techniques developed in the course of the
        SYMBIOSYS research project at COMSYS, RWTH Aachen University.

2017-09-02  Pádraig Brady  <P@draigBrady.com>

        maint: post-release administrivia
        * NEWS: Add header line for next release.
        * .prev-version: Record previous version.
        * cfg.mk (old_NEWS_hash): Auto-update.

        version 8.28
        * NEWS: Record release date.

2017-09-01  Pádraig Brady  <P@draigBrady.com>

        tests: fix false failure in recent ls --hyperlink test
        * tests/ls/hyperlink.sh: If the hostname or any part of
        the absolute path would be changed due to URL encoding,
        the test would fail.  Therefore simplify to remove
        these components of the URL from consideration.

        maint: avoid a syntax-check failure
        * .gitignore: Remove lines indicated by sc_gitignore_redundant
        in a freshly checked out repo.

2017-08-31  Pádraig Brady  <P@draigBrady.com>

        tests: exclude the expensive gnulib fts-tests
        * gnulib: The only change in this gnulib update
        is the tagging of the fts-tests module as longrunning,
        which gnulib-tool currently implicitly excludes.
        This test was seen to take about 20s and 285MB.
        Reported by Assaf Gordon on space restricted VMs.

        tty: don't distinguish input errors
        * src/tty.c (main): Don't distinguish ENOTTY from other errors,
        because isatty() doesn't portably distinguish errors.
        Solaris returns ENOENT for all input errors for example.
        Musl also returns ENOENT, and ENODEV may be returned as disscussed at:
        http://openwall.com/lists/musl/2017/04/06/6
        * tests/misc/tty.sh: Adjust accordingly.

        tests: avoid printf '0*d' construct unsupported by ash
        * tests/ln/sf-1.sh: Generate specific length with space padding
        which is supported.
        Reported by Assaf Gordon on Alpine Linux.

2017-08-31  Pádraig Brady  <P@draigBrady.com>

        tests: skip tests upon failure to set SELinux context
        On some setups the root:object_r:tmp_t context is invalid.
        This does indicate a limitation in the test framework,
        but for now we'll relax this to skipping the tests.
        The tests still run on a Fedora 25 system for example.

        * tests/cp/cp-a-selinux.sh: Upon chcon error, skip rather than ERROR.
        * tests/install/install-Z-selinux.sh: Likewise.
        * tests/misc/chcon.sh: Likewise.
        * tests/misc/runcon-no-reorder.sh: Likewise.
        * tests/misc/selinux.sh: Likewise.
        * tests/mkdir/restorecon.sh: Likewise.

2017-08-30  Kamil Dudka  <kdudka@redhat.com>

        expr: fix a recently introduced memory leak
        * src/expr.c (eval6): Free memory allocated by mbs_logical_substr().

        Introduced in v8.27-47-ga9f2be5.  Detected by Coverity Analysis:

        Error: RESOURCE_LEAK:
        src/expr.c:851: leaked_storage: Variable "s" going out of scope
        leaks the storage it points to.
        849|            char *s = mbs_logical_substr (l->u.s, pos, len);
        850|            v = str_value (s);
        851|->        }
        852|        freev (l);
        853|        freev (i1);

2017-08-30  Pádraig Brady  <P@draigBrady.com>

        build: fix build of renameat2 on Alpine Linux
        * gnulib: The only change included in this update
        it the added check for the presence of <linux/fs.h>
        which is not present on Alpine Linux by default.

        tty: fix exit code with EINVAL
        * src/tty.c (main): All systems mention that isatty()
        man return EINVAL as well as (the POSIX compliant) ENOTTY.
        Also Centos 6 was seen to return EINVAL from ttyname().
        * tests/misc/tty.sh: Fix a test issue where we assume
        standard input is always a valid tty.
        Reported by Assaf Gordon on OpenSolaris 5.10 and 5.11,
        and Centos 6.5

2017-08-30  Pádraig Brady  <P@draigBrady.com>

        runcon: revert "disable use of the TIOCSTI ioctl"
        This reverts commit v8.27-97-g8cb06d4 because
        the setsid() fallback was not implemented correctly
        and disabling the ioctl was not a complete solution
        to the security issue of the child being passed
        the tty of the parent.

        Given runcon is not really a sandbox command,
        the advice is to use `runcon ... setsid ...`
        to avoid this particular issue.

2017-08-30  Pádraig Brady  <P@draigBrady.com>

        stat: fix determination of max name length on BSD systems
        We only use one of statfs or statvfs for `stat -f`
        and on the BSDs we use statfs which doesn't have the
        f_namelen member.  However on OpenBSD and later FreeBSD
        systems statfs does provide f_namemax, so use that.

        * NEWS: Mention the improvement for OpenBSD and FreeBSD.
        * m4/stat-prog.m4: Check for f_namemax in the statfs struct.
        * src/stat.c: Return '?' rather than '*' when we can't
        determine the max length of the file system.
        * tests/ln/sf-1.sh: This test was failing on all BSDs
        due to '*' being returned for the max length which
        caused the test to attempt to create 1Mi+1 names.
        The test now uses a short name when we can't determine
        the max name length to use.

        Reported by Assaf Gordon on various BSD based systems.

2017-08-29  Pádraig Brady  <P@draigBrady.com>

        stat,tail: support "AAFS" AppArmor file system
        * src/stat.c (human_fstype): This file system is used
        to manage AppArmor policy in the Linux kernel.

        all: update gnulib submodule to latest
        * bootstrap: Sync timestamp update.

2017-08-29  Pádraig Brady  <P@draigBrady.com>

        runcon: disable use of the TIOCSTI ioctl
        Similar to the issue with SELinux sandbox (CVE-2016-7545),
        children of runcon can inject arbitrary input to the terminal
        that would be run at the originating terminal privileges.

        The new libseccomp dependency is widely available and used
        on modern SELinux systems, but is not available by default
        on older systems like RHEL6 etc.

        * m4/jm-macros.m4: Check for libseccomp and
        warn if unavailable on selinux supporting systems.
        * src/local.mk: Link runcon with -lseccomp.
        * src/runcon.c (disable_tty_inject): A new function to
        disable use of the TIOCSTI using libseccomp, or with setsid()
        where libseccomp is unavailable.
        * tests/misc/runcon-no-inject.sh: A new test that uses
        python to make the TIOCSTI call, and ensure that doesn't succeed.
        * tests/local.mk: Reference the new test
        * NEWS: Mention the fix.
        Addresses http://bugs.gnu.org/24541

2017-08-29  Pádraig Brady  <P@draigBrady.com>

        ls: support --hyperlink to output file:// URIs
        Terminals such as iTerm2 and VTE based terminals
        (as of version 0.49.1), support hyperlinks when
        passed terminals codes as described at:
        https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda

        * src/ls.c (gobble_file): Allocate an absolute file name to output.
        (quote_name): Output the absolute name with the appropriate codes.
        (file_escape): A new function to encode file names as per rfc8089.
        (main): Handle the new option and call the file_escape_init() helper.
        Disable --dired when --hyperlink is specified.
        (print_dir): Get the absolute file name here too, so that the
        directory name can be linkified.
        * NEWS: Mention the new feature.
        * tests/ls/hyperlink.sh: Add a new test.
        * tests/local.mk: Reference the new test.
        * doc/coreutils.texi (ls invocation): Describe --hyperlink.

2017-08-29  Pádraig Brady  <P@draigBrady.com>

        doc: remove older ChangeLog items
        This saves about 0.5MB uncompressed from the tarball.

        * Makefile.am: Following on from v8.26-34-g2c64bc8
        update the oldest documented version to 8.18 which
        is now about 5 years old.  Also remove older ChangeLogs
        that were previously thought to be for changes not
        in the git history, but are adequately recorded upon review.
        * build-aux/ChangeLog-2007: Remove file.
        * lib/ChangeLog-2007: Likewise.
        * m4/ChangeLog-2007: Likewise.

2017-08-29  Colin Watson  <cjwatson@debian.org>

        env: add --chdir option
        This is useful when chaining with other commands that run commands in a
        different context, while avoiding using the shell to cd, and thus
        having to consider shell quoting the chained command.

        * NEWS (New features): Document the new option.
        * doc/coreutils.texi (env invocation): Likewise.
        * src/env.c (usage): Likewise.
        (main): Implement the new option.
        * tests/misc/env.sh: Test the new option.

2017-08-29  Pádraig Brady  <P@draigBrady.com>

        tests: don't fail tests when failing to write files
        * tests/sample-test: Use framework_error_ rather than fail=1
        * tests/chown/deref.sh: Likewise.
        * tests/chown/preserve-root.sh: Likewise.
        * tests/cp/src-base-dot.sh: Likewise.
        * tests/dd/unblock-sync.sh: Likewise.
        * tests/du/2g.sh: Likewise.
        * tests/du/inacc-dest.sh: Likewise.
        * tests/du/one-file-system.sh: Likewise.
        * tests/fmt/goal-option.sh: Likewise.
        * tests/ln/hard-backup.sh: Likewise.
        * tests/ls/color-dtype-dir.sh: Likewise.
        * tests/ls/m-option.sh: Likewise.
        * tests/ls/stat-dtype.sh: Likewise.
        * tests/ls/time-style-diag.sh: Likewise.
        * tests/ls/x-option.sh: Likewise.
        * tests/misc/chcon.sh: Likewise.
        * tests/misc/nohup.sh: Likewise.
        * tests/misc/od-N.sh: Likewise.
        * tests/misc/sort-compress.sh: Likewise.
        * tests/misc/tac-continue.sh: Likewise.
        * tests/misc/time-style.sh: Likewise.
        * tests/mv/backup-dir.sh: Likewise.
        * tests/mv/dir2dir.sh: Likewise.
        * tests/rm/dir-no-w.sh: Likewise.
        * tests/rm/dir-nonrecur.sh: Likewise.
        * tests/rm/inaccessible.sh: Likewise.
        * tests/rm/interactive-always.sh: Likewise.
        * tests/rm/interactive-once.sh: Likewise.
        * tests/rm/rm3.sh: Likewise.
        * tests/rm/v-slash.sh: Likewise.
        * tests/touch/relative.sh: Likewise.

2017-08-29  Josef Cejka  <jcejka@suse.com>
            Bernhard Voelker  <mail@bernhard-voelker.de>

        df: avoid stat() for dummy file systems with -l
        When systemd is configured to automount a remote file system - see
        'man systemd.automount(5)', then the mount point is initially
        mounted by systemd with the file system type "autofs".
        When the resource is used later on, then the wanted file system is
        mounted over that mount point on demand.
        'df -l' triggered systemd to mount the file system because it called
        stat() on the mount point.
        Instead of single-casing "autofs" targets, we can avoid stat()ing
        all dummy file systems (which includes "autofs"), because those are
        skipped later on in get_dev() anyway.

        *src/df.c (filter_mount_list): Also skip dummy file systems unless
        the -a option or a specific target are given.
        * NEWS: Mention the fix.

        Fixes http://bugzilla.suse.com/show_bug.cgi?id=1043059

2017-08-29  Assaf Gordon  <assafgordon@gmail.com>

        doc: add 'realpath usage examples' section
        * doc/coreutils.texi (Realpath usage examples): New section.

2017-08-29  Assaf Gordon  <assafgordon@gmail.com>

        doc: fix realpath index entry
        The 'readlink' node has '@findex realpath' in it. This results in
            info doc/coreutils.info realpath
        incorrectly jumping to the 'readlink' node (instead of the 'realpath'
        node). Change it to @cindex instead.

        * doc/coreutils.texi (readlink): Change '@findex realpath' to @cindex.

2017-08-29  Assaf Gordon  <assafgordon@gmail.com>

        realpath: improve usage description for --relative-{to,base}
        * src/realpath.c (usage): Explicitly say 'DIR' instead of 'FILE' for
        --relative-{to,base} parameters, to avoid giving the impression
        that regular files can be used as relative base.
        * doc/coreutils.texi (realpath): Same.

2017-08-25  Pádraig Brady  <P@draigBrady.com>

        ls: consistently quote symlink targets
        * src/ls.c (gobble_file): Disable the optimization to avoid quoting
        if the symlink target itself needs quoting.  This was introduced
        with the quoting alignment adjustments in v8.25-106-g01971c0
        * tests/ls/symlink-quote.sh: Add a test.
        * tests/local.mk: Reference the test.
        * NEWS: Mention the fix.

2017-08-25  Pádraig Brady  <P@draigBrady.com>

        tail: reinstate inotify use with FIFOs
        commit v8.27-44-g18f6b22 was too aggressive in
        only allowing inotify use with regular files. This will
        support responsive processing of `tail -f fifo | ...`

        * src/tail.c (any_non_regular): Adjust to allow FIFOs
        since inotify supports these well.
        * tests/tail-2/inotify-only-regular.sh: Adjust comment.

2017-08-19  Pádraig Brady  <P@draigBrady.com>

        maint: avoid a syntax check failure
        * src/sort.c: Don't include stdio--.h as fopen() is no longer used.

        tests: fix issues on alpine linux
        * tests/misc/seq-epipe.sh: Remove stale comment.
        * tests/misc/sort-debug-warn.sh: musl doesn't indicate a set_locale()
        failure with missing locales, so avoid a test portion in that case.
        * tests/misc/wc-files0.sh: Avoid a bug on older ash implementations.
        Addresses http://bugs.gnu.org/28054

2017-08-17  Paul Eggert  <eggert@cs.ucla.edu>

        ptx: fix some integer overflow bugs
        Problem reported by Lukas Zachar at:
        http://bugzilla.redhat.com/1482445
        * src/ptx.c (line_width, gap_size, maximum_word_length)
        (reference_max_width, half_line_width, before_max_width)
        (keyafter_max_width, truncation_string_length, compare_words)
        (compare_occurs, search_table, find_occurs_in_text, print_spaces)
        (fix_output_parameters, define_all_fields):
        Use ptrdiff_t, not int, for object offsets and sizes.
        (WORD, OCCURS): Use ptrdiff_t, not short int.
        (WORD_TABLE, number_of_occurs, generate_all_output):
        Prefer ptrdiff_t to size_t where either will do.
        (total_line_count, file_line_count, OCCURS, fix_output_parameters)
        (define_all_fields):
        Use intmax_t, not int, for line counts.
        (DELTA): Remove.  All uses changed.
        (OCCURS, find_occurs_in_text, fix_output_parameters):
        Use int, not size_t, for file indexes.
        (tail_truncation, before_truncation, keyafter_truncation)
        (head_truncation, search_table, define_all_fields)
        (generate_all_output):
        Use bool for booleans.
        (digest_word_file, find_occurs_in_text):
        Use x2nrealloc instead of checking for overflow by hand.
        (find_occurs_in_text, fix_output_parameters, define_all_fields):
        Omit unnecessary cast.
        (fix_output_parameters): Don�<80><99>t assume integers fit in 11 digits.
        (fix_output_parameters, define_all_fields):
        Use sprintf return value rather than calling strlen.
        (define_all_fields): Do not rely on sprintf to generate a string
        that may contain more than INT_MAX bytes.
        (main): Use xstrtoimax, not xstrtoul.
        Use xnmalloc to catch integer overflow.

        nohup: simplify by using fcntl
        * src/nohup.c: Do not include cloexec.h.
        (main): Use fcntl rather than dup + set_cloexec_flag.

        sort: use pthread_sigmask, not sigprocmask
        POSIX says sigprocmask has unspecified behavior in a multithreaded
        program like �<80><98>sort�<80><99>.
        * src/sort.c (pthread_sigmask) [GNULIB_defined_pthread_functions]:
        New macro, for use when �<80><98>sort�<80><99> is not multithreaded.
        (cs_enter, cs_leave): Use it.  Pass address, not value, as
        this is typically a tad faster.  All callers changed.

        sort: minor cleanups
        * src/sort.c (move_fd): Rename from move_fd_or_die,
        since it no longer can die.

        sort: file descriptor discipline
        Use O_CLOEXEC when creating file descriptors, so that subsidiary
        processes do not inherit file descriptors that they do not need.
        This is helpful for �<80><98>sort�<80><99>, as it is a multithreaded program that
        forks and execs.
        * bootstrap.conf (gnulib_modules): Add mkostemp, open, pipe2.
        * src/sort.c (create_temp_file): Open temporary file with O_CLOEXEC.
        (stream_open): Open the stream with O_CLOEXEC.
        (pipe_fork): Create the pipe with O_CLOEXEC.
        (check_output): Open the output file with O_CLOEXEC.
        (main): Use xfopen/xfclose to handle --files0-from, so that
        O_CLOEXEC is used properly.  This is simpler anyway.
        * tests/misc/sort-files0-from.pl: Adjust to change in diagnostic
        wording.

        build: update gnulib submodule to latest

2017-08-14  Pádraig Brady  <P@draigBrady.com>

        kill: fix signal number to name lookup on AIX
        * src/operand2sig.c (operand2sig): AIX uses a different bit pattern
        in the returned status from the wait() functions and from shells.
        Therefore hardcode the selection of the lower bits of the number.
        * NEWS: Mention the fix.

        build: use the appropriate single file include option with xlc
        * configure.ac: Set USE_XLC_INCLUDE when __xlc__ is defined.
        * src/local.mk: Use it to select the appropriate include option.
        Reported by Michael Felt.

        tests: avoid false failures on AIX
        * tests/ln/sf-1.sh: Limit the symlink size to 1MiB
        to avoid memory exhaustion seen on NFS on AIX, giving:
          + printf '%0*d' 4294967296 0
          + ./tests/ln/sf-1.sh: line 38: printf: warning: 0: Result too large
        * tests/id/setgid.sh: Skip the test when the adjusted gid
        would equal 4294967295, as that's reserved on AIX.
        Reported by Michael Felt.

        sort: handle musl locale differences in --debug reporting
        * src/sort.c (main): Don't assume hard_LC_COLLATE implies
        a successful setting of the locale as musl defaults to
        UTF8 when failing to set the specified locale.
        * tests/misc/sort-debug-warn.sh: Adjust for the now
        separated locale debug info and map the musl specific
        message back to the common case.
        Addresses https://bugs.gnu.org/28054

        seq: produce consistent error messages upon write error
        * src/seq.c (io_error): Use the same error message as would
        be generated at exit time when closing the stdout stream.
        The inconsistency was added with commit v8.25-26-gc92585b.
        This was noticed due to an inconsistency in the expected
        error message generated by seq on musl libc.
        Addresses https://bugs.gnu.org/28054

        tests: fix false failure with large printf formats
        * tests/misc/printf-surprise.sh: With musl libc the
        large printf format does succeed, outputting data.
        To avoid SIGPIPE being generated we ignore that signal
        and then handle the subsequent EPIPE error.
        Addresses https://bugs.gnu.org/28054

2017-08-12  Jim Meyering  <meyering@fb.com>

        build: adjust warning options to work with latest GCC
        * configure.ac: Disable some new warnings to avoid false positives.
        Building with warnings enabled and latest gcc would evoke build
        failure without these changes.  Disable the following in coreutils
        proper: -Wformat-overflow=2 -Wformat-truncation=2, and
        disable these for gnulib: -Wformat-truncation=2 -Wduplicated-branches

        gnulib: update to latest and adjust gl/modules/tempname.diff
        * gnulib: Update to latest.
        * gl/modules/tempname.diff: This patch failed to apply.
        Adjust it to reflect removal of the secure_getenv dependency.

        chroot: fix typo in preceding change: didn't compile
        * src/chroot.c (usage): Add backslashes.

2017-08-10  Jim Meyering  <meyering@fb.com>

        doc: correct technicality in chroot's --help output
        * src/chroot.c (usage): Use correct quoting in descriptive diagnostic.
        We would run `"$SHELL" -i`, not `${SHELL} -i`.

2017-08-09  Assaf Gordon  <assafgordon@gmail.com>

        doc: fix join example
        * doc/coreutils.texi (join invocation): Fix incorrect output in example.
        Reported by Phlosioneer in https://bugs.gnu.org/28014 .

2017-08-04  Paul Eggert  <eggert@cs.ucla.edu>

        build: update gnulib submodule to latest

2017-08-03  Paul Eggert  <eggert@cs.ucla.edu>

        copy: more-accurate warning about destruction
        * src/copy.c (copy_internal):
        * tests/cp/backup-is-src.sh, tests/mv/backup-is-src.sh:
        Say "might destroy", not "would destroy".

2017-08-03  Pádraig Brady  <P@draigBrady.com>

        maint: avoid a syntax-check failure
        * src/shred.c (wipename): As per the comment, the arguments
        to error() are sufficiently quoted, so split the call over
        multiple lines to avoid the syntax-check.

2017-08-02  Paul Eggert  <eggert@cs.ucla.edu>

        build: update gnulib submodule to latest

2017-08-01  Paul Eggert  <eggert@cs.ucla.edu>

        copy: go back to failing 'cp --backup a~ a'
        Suggested by Kamil Dudka in:
        http://lists.gnu.org/archive/html/coreutils/2017-07/msg00072.html
        * NEWS: Document the changed nature of the fix.
        * doc/coreutils.texi, tests/cp/backup-is-src.sh:
        * tests/mv/backup-is-src.sh: Revert previous change.
        * src/copy.c (source_is_dst_backup): New function.
        (copy_internal): Use it.  Fail instead of falling back on numbered
        backups when it looks like the backup will overwrite the source.
        Although this reintroduces a race, it's more compatible with
        previous behavior.

2017-07-31  Paul Eggert  <eggert@cs.ucla.edu>

        copy: sanity-check --suffix
        * src/cp.c, src/install.c, src/ln.c, src/mv.c (main):
        Use set_simple_backup_suffix, to sanity-check the user-supplied
        backup suffix.

        copy: make backup files more reliably
        * NEWS, doc/coreutils.texi (Backup options): Document the change.
        * bootstrap.conf (gnulib_modules): Add backup-rename.
        * src/copy.c (copy_internal): Silently switch to numbered backups
        if a simple backup might lose data.  Use backup_file_rename
        to avoid races with numbered backups.
        * tests/cp/backup-is-src.sh, tests/mv/backup-is-src.sh:
        Adjust to match new behavior.

        shred: avoid rename race
        Use renameat2 to avoid a rename race condition, on recent-enough
        GNU/Linux.
        * bootstrap.conf (gnulib_modules): Add renameat2.
        * src/shred.c: Include renameat2.h.
        (wipename): Use renameat2 instead of rename.

        build: update gnulib submodule to latest

2017-07-25  Jim Meyering  <meyering@fb.com>

        maint: fix grammar in a shred.c comment
        * src/shred.c: Remove spurious "to" in an old comment.

2017-07-23  Pádraig Brady  <P@draigBrady.com>

        maint: fix recent syntax-check failures
        * .gitignore: Add /lib/utime.h from the recent gnulib update.
        * src/nproc.c (usage): Adjust spacing to placate help2man.

        shred: remove redundant zeroing of freed memory
        * src/shred.c (dopass): shred used to read the input file,
        and so needed to ensure internal memory was cleared.
        This is no longer the case since SH-UTILS-1_16f-260-gf381610
        so avoid this redundant clearing.
        (do_wipefd): Likewise.
        * NEWS: Remove the recent mention of this issue.

        maint: resync with blake2 upstream
        * src/blake2/blake2-impl.h: Don't use the equivalent explicit_bzero().

        tests: avoid a false failure on AIX
        * tests/misc/sync.sh: Normalize the error messages
        when syncing a non read/write directory, as AIX
        gives the "Is a directory" error.
        Also ensure that sync(1) returns an error for this
        case on all systems.

2017-07-20  Paul Eggert  <eggert@cs.ucla.edu>

        shred: use explicit_bzero
        * NEWS: Document this.
        * bootstrap.conf (gnulib_modules): Add explicit_bzero.
        * gl/lib/randint.c (randint_free):
        * gl/lib/randread.c (randread_free):
        * src/blake2/blake2-impl.h (secure_zero_memory):
        * src/shred.c (dopass, do_wipefd):
        Prefer explicit_bzero to memset when erasing secrets.

        build: update gnulib submodule to latest

2017-07-10  Andreas Schwab  <schwab@linux-m68k.org>

        nproc: fix indentation of usage output
        * src/nproc.c (usage): Align output.

2017-07-10  Jim Meyering  <meyering@fb.com>

        groups: do not exit early
        Most programs take care to operate on all command-line-specified
        operands before exiting.  That is an important feature that allows
        to identify all problems with the first run.  However, groups would
        exit upon the first problematic user name.
        Bug introduced via commit v6.10-56-g167b8025ac.
        * src/groups.c (main): Do not exit immediately upon error.
        * tests/misc/groups-process-all.sh: New file. Test for this.
        * tests/local.mk (all_tests): Add it.
        * NEWS (Bug fixes): Mention this.

2017-07-08  Jim Meyering  <jim@meyering.net>

        tests: groups-dash.sh: avoid false failure
        * tests/misc/groups-dash.sh: Avoid false failure on a system for which
        "none" is a valid user name.  The first invocation would succeed, and
        the second would fail with "groups: �<80><98>--�<80><99>: no such user".
        Use a user name that cannot exist.

        doc: tweak wording
        * NEWS (Bug fixes): Tweak wording of the mv/cp-vs-symlink-ownership
        entry and the one about df.

2017-06-28  Assaf Gordon  <assafgordon@gmail.com>

        expr: add multibyte support
        Discussed in https://bugs.gnu.org/26779 .

        * NEWS: Mention the improvement.
        * bootstrap.conf: Add gnulib modules mbslen,mbschr.
        * src/expr.c (mbs_logical_substr): New function to return a substring
        based on logical character positions (instead of bytes).
        (mbs_logical_cspn): Similar to strcspn/mbscspn, but returns number of
        logical characters instead of byte offset.
        (mbs_offset_to_chars): New function to return number of logical
        characters fitting in a given byte offset.
        (docolon): Report matched logical characters instead of bytes.
        (eval6): For length/substr/index operations, use logical characters
        instead of bytes by calling the above new functions.
        * tests/misc/expr.pl: Repeat all tests with non-C locale to detect any
        regressions.
        * tests/misc/expr-multibyte.pl: New tests with multibyte input.
        * tests/local.mk: Add new test file.

2017-06-21  Jim Meyering  <meyering@fb.com>

        maint: avoid spurious "make distcheck" failure
        When the generated file, doc/constants.texi, happens to be older than
        doc/coreutils.info, it will not be updated until/unless its generated
        contents change.  This is due to way that rule is careful to update
        the file, to avoid provoking a pointless rerunning of makeinfo.

        Note that this does not happen when one first runs "make distclean",
        as recommended in README-release.  However, I sometimes run it as
        a more-rigorous "make check", and shouldn't have to manually run
        "make distclean" first, in that case.

        Before this change, one could reproduce the failure by running
        `touch -dyesterday doc/constants.texi && make distcheck`.  It would
        fail with "makeinfo: could not open ../../doc/coreutils.info-t
        for writing: Permission denied"
        * Makefile.am (dist-hook): Touch the two generated files, so that
        they cannot be out of date wrt doc/coreutils.texi.

2017-06-17  Pádraig Brady  <P@draigBrady.com>

        maint: use C99 for loop initial declarations where possible
        This results in a net reduction of about 120 lines.

        tail: only use inotify with regular files
        * src/tail.c (any_non_regular): A new function to check passed files.
        (main): Use the above to skip inotify if any non regular files passed
        like /dev/tty or /dev/ttyUSB0 etc.
        * tests/tail-2/inotify-only-regular.sh: A new test.
        * tests/local.mk: Reference the new test.
        * NEWS: Mention the bug fix.
        Fixes http://bugs.gnu.org/21265 and http://bugs.gnu.org/27368

        tail: with -f don't warn if doing a blocking read of a tty
        * src/tail.c: (main): Only issue the warning about -f being
        ineffective when we're not going into simple blocking mode.
        * tests/tail-2/follow-stdin.sh: Ensure the warning is output correctly.
        Fixes http://bugs.gnu.org/27368

2017-06-11  Pádraig Brady  <P@draigBrady.com>

        tail: exit promptly when output no longer writable
        This will support use cases like:

          tail -f file.log | grep -q trigger &&
          process_immediately

        * src/tail.c (check_output_alive): A new function that
        uses select on fifos or pipes to detect if they're broken.
        (tail_forever): Call check_output_alive() periodically.
        (tail_forever_inotify): Merge the select() call from
        check_output_alive() into the select() originally present
        for the --pid case, and adjust accordingly.
        * tests/tail-2/pipe-f.sh: Add test cases.
        * NEWS: Mention the improvement.

2017-06-11  Jim Meyering  <meyering@fb.com>

        maint: update to work with GCC7's -Werror=implicit-fallthrough=5
        * src/system.h (FALLTHROUGH): Define.
        * src/cp.c (main): Use new FALLTHROUGH macro in place of comments.
        * src/basename.c (main): Likewise.
        * src/dircolors.c (append_quoted): Likewise.
        * src/echo.c (main): Likewise.
        * src/fold.c (main): Likewise.
        * src/join.c (main): Likewise.
        * src/kill.c (main): Likewise.
        * src/ls.c (get_funky_string, gobble_file): Likewise.
        * src/sort.c (parse_field_count, main): Likewise.
        * src/stat.c (print_it): Likewise.
        * src/tail.c (parse_obsolete_option): Likewise.
        * src/test.c (posixtest): Likewise.
        * src/wc.c (wc): Likewise.
        * src/who.c (main): Likewise.

2017-06-07  Pádraig Brady  <P@draigBrady.com>

        tail: with --pid, ensure all inotify events are processed
        * NEWS: Mention the bug fix.
        * src/tail.c (tail_forever_inotify): With --pid, avoid waiting
        for new events if there are still events to process.
        * tests/tail-2/inotify-dir-recreate.sh: Adjust to trigger.

        tests: fix issues with recently added tail test
        * tests/tail-2/inotify-dir-recreate.sh: Skip when
        inotify is not usable.  Also remove a bash specific &> construct.

2017-06-03  Pádraig Brady  <P@draigBrady.com>

        copy: don't fail when unable to chown symlinks
        * src/copy.c (copy_internal): Honor the x->require_preserve flag
        for symlinks as we do for ordinary files, so we don't exit with
        failure upon failure to chown a symbolic link.
        * NEWS: Mention the bug fix.

2017-05-29  Sebastian Kisela  <skisela@redhat.com>

        doc: mention `setpriv --no-new-privs` feature in runcon info
        * doc/coreutils.texi (runcon invocation): Mention setpriv usage.
        Discussed at https://bugzilla.redhat.com/1360903

2017-05-18  Pádraig Brady  <P@draigBrady.com>

        mv: distinguish copy and rename operations with --verbose
        * src/copy.c (copy_internal): In x->move_mode distinguish
        whether we're copying, creating directory, or renaming.
        * tests/mv/backup-dir.sh: Adjust to new output.
        * tests/mv/mv-n.sh: Likewise.
        * tests/mv/mv-special-1.sh: Likewise.
        * NEWS: Mention the improvement.
        Fixes http://bugs.gnu.org/26971

2017-05-11  Prateek saxena  <prateeksaxena2@gmail.com>

        uptime: remove inconsistent AM/PM from current time
        * src/uptime.c (main): 00-23 was always used for the hour component
        of the current time, so remove the AM/PM output (which was only
        present in some locales anyway).  Also add seconds to the time
        to be more consistent with the usual procps-ng uptime implementation
        on GNU/Linux.
        * NEWS: Mention the fix.
        Fixes http://bugs.gnu.org/26783

2017-05-04  Pádraig Brady  <P@draigBrady.com>

        maint: fix various typos in recent commits
        * NEWS: Grammar fixes.
        * HACKING: Likewise.

2017-05-04  Jaak Ristioja  <jaak.ristioja@cyber.ee>

        doc: Fixed typo in timeout man page
        * man/timeout.x: Correct spelling of "currently".
        Fixes http://bugs.gnu.org/26762

2017-04-30  Pádraig Brady  <P@draigBrady.com>

        doc: update the instructions for generating a coverage report
        * HACKING: Change from explicit instructions to using gnulib
        provided coverage testing targets.  Also include instructions
        for adding root only tests to the report.
        Fixes http://bugs.gnu.org/26709

2017-04-27  Paul Eggert  <eggert@cs.ucla.edu>

        dd: simplify translator�<80><99>s jobs
        * src/dd.c (print_xfer_stats): Format the SI units directly,
        without translating them, to simplify the translators�<80><99> jobs.
        See Bug#26621.

2017-04-27  Pádraig Brady  <P@draigBrady.com>

        date,touch: test and document large TZ security issue
        Add a test for CVE-2017-7476 which was fixed in gnulib at:
        http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=commitdiff;h=94e01571

        * tests/misc/date-tz.sh: Add a new test which overwrites enough
        of the heap to trigger a segfault, even without ASAN enabled.
        * tests/local.mk: Reference the new test.
        * NEWS: Mention the bug fix.

2017-04-27  Pádraig Brady  <P@draigBrady.com>

        build: update gnulib submodule to latest
        * .gitignore: Add new entry as indicated by `make syntax-check`.

2017-04-24  Paul Eggert  <eggert@cs.ucla.edu>

        dd: status=progress outputs "6 s", not "6.00001 s"
        Problem reported by Benno Schulenberg (Bug#26621).
        * NEWS: Document this.
        * src/dd.c (print_xfer_stats): With status=progress,
        format times with %.0f rather than %g.  Improve
        translator comments.

2017-04-22  Paul Eggert  <eggert@cs.ucla.edu>

        build: update gnulib submodule to latest

        maint: remove unused functions and constants
        These were found by clang.
        * gl/lib/rand-isaac.c (min):
        * gl/lib/randint.c (shift_right):
        * src/md5sum.c (algorithm):
        Remove; unused.

        date: adjust to gnulib parse-datetime changes
        * doc/coreutils.texi (Options for date): Capitalize a sentence.
        * tests/misc/date-debug.sh: Adjust --debug output to match
        recent changes to Gnulib�<80><99>s parse-datetime module.

        build: update gnulib submodule to latest
        * gl/modules/tempname.diff: Update to match current Gnulib.

2017-04-18  Bogdan Drozdowski  <bogdandr@op.pl>

        shred: fix invalid pattern generation for certain sizes
        * src/shred.c (fillpattern): Fix the "off by one" issue when
        testing whether we have enough space to copy the already
        written portion of the buffer to the remainder of the buffer.
        Specifically for buffer sizes that are (3*(2^x))+1, i.e. 7,13,...
        we both use an uninitialized byte and invoke undefined
        behavior in memcpy() operation on overlapping memory regions.
        * tests/misc/shred-passes.sh: Add an invocation that will
        trigger either valgrind UMR, or ASAN like:
          ERROR: AddressSanitizer: memcpy-param-overlap: memory ranges
          #1 0x403065 in fillpattern src/shred.c:293
        A direct test is awkward due to the random writes surrounding
        the problematic pattern writes.
        Fixes http://bugs.gnu.org/26545

2017-04-17  Bo Rydberg  <bolry@hotmail.com>

        doc: fix awk example for getting penultimate field
        * doc/coreutils.texi (cut invocation): Add required brackets.
        Fixes http://bugs.gnu.org/26519

2017-04-06  Sebastian Kisela  <skisela@redhat.com>

        tail: revert to polling if a followed directory is replaced
        * src/tail.c (tail_forever_inotify): Add the IN_DELETE_SELF flag when
        creating watch for the parent directory.  After the parent directory
        is removed, an event is caught and then we switch from inotify to
        polling mode.  Till now, inotify has always frozen because it waited for
        an event from a watched dir, which has been already deleted and was not
        added again.
        * tests/tail-2/inotify-dir-recreate.sh: Add a test case.
        * tests/local.mk: Reference the new test.
        * NEWS: Mention the bug fix.
        Fixes http://bugs.gnu.org/26363
        Reported at https://bugzilla.redhat.com/1283760

2017-04-06  Pádraig Brady  <P@draigBrady.com>

        maint: fix syntax-check issues in previous tty commit
        * src/tty.c: Avoid EXIT_FAILURE to be more descriptive
        and to placate sc_some_programs_must_avoid_exit_failure.

2017-04-05  Paul Eggert  <eggert@cs.ucla.edu>

        tty: handle misconfigured namespaces
        On some platforms, isatty succeeds but ttyname fails.
        POSIX does not seem to allow this, but there it is.
        Problem reported by Christian Brauner (Bug#26371).
        While we�<80><99>re at it, check for errors more carefully and return a
        new exit status 4 if stdin is closed or a similar error occurs.
        * doc/coreutils.texi (tty invocation): Document new behavior.
        * init.cfg (stderr_fileno_):
        Don't assume have_input_tty is not in the environment.
        * src/tty.c (TTY_STDIN_ERROR): New constant.
        (main): Exit with nonzero status if there is a usage error,
        like other coreutils programs.
        Check for error in getting stdin type.
        * tests/misc/tty.sh: New file.
        * tests/local.mk (all_tests): Add it.

2017-04-03  Pádraig Brady  <P@draigBrady.com>

        doc: refactor and update expand and unexpand --help
        * src/expand-common.c (emit_tab_list_info): A new function to
        output the extended info on --tab=LIST, including the new
        '+' and '/' prefixes.
        * src/expand-common.h: Declare the above.
        * src/expand.c (usage:): Call emit_tab_list_info and
        match alignment with that used in unexpand --help.
        * src/unexpand.c (usage): Likewise.

2017-04-03  Jacob Keller  <jacob.e.keller@intel.com>

        expand,unexpand: add support for incremental tab stops
        Support --tabs="1,+8" which is equivalent to --tabs="1,9,17,..."
        useful for viewing unified diff output with its 1 character
        gutter for example.

        * doc/coreutils.texi ({expand,unexpand} invocation): Document,
        using diff processing as the example.
        * src/expand-common.c (set_increment_size): Update the new
        increment_size global.
        (parse_tab_stops): Handle the new '+' prefix.
        (finalize_tab_stops): Verify both '+' and '/' prefixes
        are not used together.
        * tests/misc/expand.pl: Add test cases.
        * NEWS: Mention the new feature.

2017-03-30  Paul Eggert  <eggert@cs.ucla.edu>

        sort: update comment
        * src/sort.c: Update identifiers in comment.

2017-03-30  Chris Davies  <chris@roaima.co.uk>

        doc: clarify in dd man page that bs= overrides [io]bs=
        * src/dd.c (usage): Add the extra info.
        Reported in https://bugs.debian.org/859021

2017-03-28  Ludovic Courtès  <ludo@gnu.org>

        tests: avoid false ulimit failure on some systems
        * tests/misc/cut-huge-range.sh: On some systems returns_ may
        use more memory, so incorporate that in the determination
        of the ulimit value to use.  Noticed on ARMv7 with bash-4.4.12,
        and x86_64 with bash-4.2.37.
        Fixes http://bugs.gnu.org/26253

2017-03-28  Pádraig Brady  <P@draigBrady.com>

        maint: avoid syntax check failure with wrapped returns_
        * cfg.mk (sc_prohibit_env_returns): Allow wrapped calls to
        return_ of the form: `wrapper_ returns_ ...` which is needed
        with the following commit.

2017-03-28  Michael Heimpold  <mhei@heimpold.de>

        split: add new --hex-suffixes option
        * doc/coreutils.texi (split invocation): Document the new option.
        * src/split.c (usage): Likewise.
        (main): Process the new option much like --numeric-suffixes,
        but with an adjusted alphabet.
        * tests/split/numeric.sh: Refactor to support --hex mode.
        * NEWS: Mention the new feature.

2017-03-28  Pádraig Brady  <P@draigBrady.com>

        md5sum,b2sum,sha*sum: don't erroneously trigger BSD reversed mode
        * src/md5sum.c (split_3): Verify hex digits internally before
        triggering the global bsd_reversed mode flag.
        (bsd_split_3): Likewise.
        * tests/misc/md5sum-bsd.sh: Add a test case.
        * NEWS: Mention the bug fix.
        Fixes http://bugs.gnu.org/26263

2017-03-27  Philipp Thomas  <pth@suse.de>

        df: avoid querying excluded file systems
        * src/df.c (filter_mount_list): Avoid stat() on
        explicitly excluded file systems, which is especially
        significant in cases like `-x nfs` which may hang.
        * NEWS: Mention the bug fix.

2017-03-26  Pádraig Brady  <P@draigBrady.com>

        maint: avoid a static analysis warning in expand-common
        * src/expand-common.c (next_file): We're dependent on calling
        this function with NULL to initialize things appropriately.
        So enforce this with assert(), which avoids a warning from
        clang-anaylzer.

        split: process more efficiently when filters exit early
        * src/split.c (bytes_split): Don't write to an existing filter
        if it has exited.  When filters exit early, skip input data if
        possible.  Refactor out 2 redundant variables.
        * tests/split/filter.sh: Improve test coverage given the
        new more efficient processing.  Also use a 10TB file to
        expand the file systems tested on.

2017-03-26  Pádraig Brady  <P@draigBrady.com>

        split: ensure input is processed when filters exit early
        commit v8.25-4-g62e7af0 introduced the issue as it
        broke out of the processing loop irrespective of
        the value of new_file_flag which was used to indicate
        a finite number of filters or not.

        For example, this ran forever (as it should):
          $ yes | split --filter="head -c1 >/dev/null" -b 1000
        However this exited immediately due to EPIPE being propagated
        back through cwrite and the loop not considering new filters:
          $ yes | split --filter="head -c1 >/dev/null" -b 100000

        Similarly processing would exit early for a bounded number of
        output files, resulting in empty data sent to all but the first:
          $ truncate -s10T big.in
          $ split --filter='head -c1 >$FILE' -n 2 big.in
          $ echo $(stat -c%s x??)
          1 0

        I was alerted to this code by clang-analyzer,
        which indicated dead assigments, which is often
        an indication of code that hasn't considered all cases.

        * src/split.c (bytes_split): Change the last condition in
        the processing loop to also consider the number of files
        before breaking out of the processing loop.
        * tests/split/filter.sh: Add a test case.
        * NEWS: Mention the bug fix.

2017-03-11  Pádraig Brady  <P@draigBrady.com>

        tests: avoid a false failure on OS X 10.5.8
        * tests/misc/sort-debug-keys.sh: Disparate LC_CTYPE and LC_MESSAGES
        are not supported, with the result LC_MESSAGES=C is used throughout.
        Therefore just set LC_ALL in the test, and normalize the message
        variants with sed.
        Reported and tested by J Rogowsky.

        build: fix missing renameat() on OS X 10.5.8
        * bootstrap.conf: Depend on renameat.
        Reported and tested by J Rogowsky.
        Fixes http://bugs.gnu.org/26044

2017-03-10  Paul Eggert  <eggert@cs.ucla.edu>

        tests: port to tzdb-2017a
        Problem reported by Bernhard Voelker in:
        http://lists.gnu.org/archive/html/coreutils/2017-03/msg00026.html
        * tests/misc/date-debug.sh: Port test to tzdb 2017a,
        and future-proof the America/Belize test.

2017-03-09  Pádraig Brady  <P@draigBrady.com>

        build: for factor use C in more cases for arm64 and ppc64
        * src/longlong.h: Sync from gmp repo incorporating:
        Use asm-free umul_ppmm() on arm64 and ppc64.

        doc: rearrange a recent bug entry to an improvement in NEWS
        * NEWS: The stat,tail change was an improvement, not a bug fix.
        * cfg.mk [old_NEWS_hash]: update with `make update-NEWS-hash`.

        maint: post-release administrivia
        * NEWS: Add header line for next release.
        * .prev-version: Record previous version.
        * cfg.mk (old_NEWS_hash): Auto-update.

        version 8.27
        * NEWS: Record release date.

        build: update gnulib submodule to latest
        Removes extraneous parse-datetime.c generated during build

        stat,tail: support "RDT" Linux kernel control file system
        * src/stat.c (human_fstype): This file system is the user interface
        for resource allocation in Intel's Resource Director Technology.

        doc: spelling fix for recent doc addition
        * doc/coreutils.texi (join invocation): s/preceeding/preceding/.

2017-03-08  Bernhard Voelker  <mail@bernhard-voelker.de>

        build: avoid redundant build of tr with --enable-single-binary
        * src/local.mk [check-duplicate-no-install]: Depend on the
        single-binary tr, or the system tr, as the edge case
        where these are not available only result in the sanity
        check being effectively ignored.

2017-03-08  Assaf Gordon  <assafgordon@gmail.com>

        build: fix 'install-html' target
        Switching to non-recursive makefiles broke the 'install-html' target:
        The gettext plumbing requires an 'install-html' target in po/Makefile.
        This was fixed in gettext v0.19.8.1-41-ge5a008a, but packages using
        older gettext need to manually patch po/Makefile.in.in.
        Reported (for 'sed') and suggested fix by Eric Blake in
        https://bugs.gnu.org/25690 .

        * bootstrap.conf (bootstrap_epilogue): Add 'install-{html,pdf,dvi,ps}'
          targets to po/Makefile.in.in (if needed).

2017-03-04  Pádraig Brady  <P@draigBrady.com>

        maint: bump makeinfo --version requirement to 6.1
        * bootstrap.conf: s/4.13/6.1/ as versions previous to that
        generated invalid html with interspersed <span> tags that
        were visible to the user.  Version 6.1 is available for a
        year now, and is available in most distros.

2017-03-04  Bernhard Voelker  <mail@bernhard-voelker.de>

        doc: avoid makeinfo warning
        makeinfo issues the following:

          doc/coreutils.texi:6568: warning: @sc argument all uppercase,\
            thus no effect.

        * doc/coreutils.texi (join invocation): Remove the @sc macro around
        the all uppercase "GNU".

2017-03-03  Pádraig Brady  <P@draigBrady.com>

        build: update gnulib submodule to latest
        Fixes a test-lock failure on MacOS

        timeout: handle multiple children on solaris
        * src/timeout.c (install_sigchld): A new function to
        install the SIGCHLD handler using sigaction() rather
        than signal(), because with the latter on solaris
        the signal handler is reset to default and thus
        sigsuspend() only returns for the first finished child.
        Reported by Assaf Gordon.

        tests: avoid a spurious failure on older debian
        * tests/misc/cut-huge-range.sh: Bump up the ulimit,
        to avoid a false failure due hitting the previously
        detected ulimit.
        Reported by Assaf Gordon.

        build: fix libstdbuf build on AIX 7
        * src/libstdbuf.c: undef malloc so as libstdbuf is
        not linked with gnulib, and anyway the replacement is
        never needed since we never malloc(0).
        Reported by Assaf Gordon.

2017-03-02  Assaf Gordon  <assafgordon@gmail.com>

        doc: expand 'join' info section
        * doc/coreutils.texi (join invocation): Expand section to
        add examples and more details.
        Suggested by Dan Jacobson in https://bugs.gnu.org/25870

2017-03-01  Pádraig Brady  <P@draigBrady.com>

        doc: give a stronger security warning in md5/sha1 man pages
        * man/md5sum.x: Give a more direct warning againt the use
        of this hash algorithm for security purposes.
        * man/sha1sum.x: Likewise.
        Suggested by Jim Meyering.

        expand: avoid an extraneous warning on 32 bit
        * src/expand-common (parse-tab-stops): Exit earlier upon overflow
        so another warning isn't issued (on 32 bit) in add_tab_stop().
        Flagged in https://hydra.nixos.org/build/49499970

        doc: indicate sha1 has the same limitations as md5
        * doc/coreutils.texi (sha1sum invocation): Given that a SHA-1
        preimage attack has occurred as documented at http://shattered.io/,
        document sha1sum as having the same limitations as md5sum.
        (md5sum): Parameterize the warning for use in both cases.
        * man/md5sum.x: Mention b2sum(1) as a more secure alternative.
        * man/sha1sum.x: Give the same warning as done for md5sum(1).

        maint: avoid a -Werror=null-dereference with GCC-6.3.1
        * src/stty.c (sane_mode): Assert to inform the compiler
        we know the pointer will be valid.

        expand,unexpand: support specifying a trailing tab size
        * doc/coreutils.texi (expand invocation): Document the feature.
        (unexpand invocation): Likewise.
        * src/expand-common.c (extend_size): A new global to use
        when the last tab stop is prefixed by '/'.
        (set_extend_size): A new function to validate and set
        the new extend_size global.
        (parse_tab_stops): Call set_extend_size() for '/' prefixes.
        (finalize_tab_stops): Ensure a single specified '/' is
        treated like a standard tabsize, but also ensure that
        when '/' is specified with a single other entry that
        we process as a list rather than a tab size.
        (get_next_tab_stop): Use the tab size if set,
        for items after the user specified tab position list.
        * tests/misc/expand.pl: Add test cases
        * NEWS: Mention the new feature.
        Fixes http://bugs.gnu.org/25540

2017-02-26  Pádraig Brady  <P@draigBrady.com>

        nproc: support OMP_THREAD_LIMIT to set a max value
        This comes from the latest gnulib.
        Also handling of OMP_NUM_THREADS has been adjusted
        to support comma separated values indicating a nesting level,
        in which case the first value is taken.  Also OMP_NUM_THREADS=0
        is now ignored instead of being treated as 1, to match
        the behavior of libgomp.

        * NEWS: Mention the OMP_THREAD_LIMIT improvement,
        and OMP_NUM_THREADS now handling nested values.
        * doc/coreutils.texi (nproc invocation): Describe OMP_THREAD_LIMIT
        as a way to set the max value, with OMP_THREAD_LIMIT setting the min.
        * tests/misc/nproc-override.sh: A new test to exercise the
        updated gnulib code with all combinations of these OMP variables.
        * tests/local.mk: Reference the new test.

2017-02-26  Pádraig Brady  <P@draigBrady.com>

        build: update gnulib submodule to latest

2017-02-23  Pádraig Brady  <P@draigBrady.com>

        doc: add NEWS for an improvement to dd in the last release
        * NEWS: Mention the avoidance of the gotcha with specifying
        a hex constant like count=0x1000 etc. as that previously
        was silently interpreted as 0.

2017-02-23  Pádraig Brady  <P@draigBrady.com>

        cp: set SELinux context for --parents directories
        * src/copy.c (set_process_security_ctx, set_file_security_ctx):
        Export for use in cp.c.
        * src/copy.h: Likewise.
        * src/cp.c (make_dir_parents_private): Call the exported functions
        to set the security context for new and updated directories.
        * tests/cp/cp-a-selinux.sh: Add a test case.

        Fixes http://bugs.gnu.org/25378

2017-02-18  Pádraig Brady  <P@draigBrady.com>

        maint: tweaks so syntax tests pass for previous commit
        * .gitignore: placate sc_gitignore_redundant.
        * po/POTFILES.in: placate sc_po_check.

2017-02-16  Bernhard Voelker  <mail@bernhard-voelker.de>

        doc: avoid makeinfo warning
        Commit v8.26-38-g99deaff introduced this warning:

            MAKEINFO doc/coreutils.info
          ./doc/coreutils.texi:10268: \
            warning: `.' or `,' must follow @xref, not `@'.

        * doc/coreutils.texi (readlink invocation): Add '.' after @xref macro.

2017-02-16  Paul Eggert  <eggert@cs.ucla.edu>

        maint: xsetmode renamed to xbinary-io
        * bootstrap.conf, src/base64.c, src/cat.c, src/cksum.c:
        * src/head.c, src/md5sum.c, src/od.c, src/split.c, src/sum.c:
        * src/tac.c, src/tail.c, src/tee.c, src/tr.c, src/wc.c:
        Adjust to renaming of the xsetmode module to xbinary-io,
        and of the xsetmode function to xset_binary_mode.

        build: update gnulib submodule to latest

2017-02-15  Paul Eggert  <eggert@cs.ucla.edu>

        maint: use xsetmode, not xfreopen
        This fixes a bug noted by Eric Blake.  Code was using xfreopen to
        change files to binary mode, but this fails for stdout when in
        append mode.  Such code should use xsetmode instead.  This affects
        only the port on platforms like MS-Windows which distiguish text
        from binary I/O.
        * bootstrap.conf (gnulib_modules):
        Remove xfreopen and add xsetmode.  Sort.
        * src/base64.c (main):
        * src/cat.c (main):
        * src/cksum.c (cksum):
        * src/head.c (head_file, main):
        * src/md5sum.c (digest_file):
        * src/od.c (open_next_file):
        * src/split.c (main):
        * src/sum.c (bsd_sum_file, sysv_sum_file):
        * src/tac.c (tac_file, main):
        * src/tail.c (tail_file):
        * src/tee.c (tee_files):
        * src/tr.c (main):
        * src/wc.c (wc_file): Use xsetmode, not xfreopen.

        build: update gnulib submodule to latest

2017-02-13  Pádraig Brady  <P@draigBrady.com>

        maint: tweaks so syntax tests pass for previous commit
        * src/force-link.h: Don't include headers already included by system.h
        * src/force-link.c: Likewise.  Also include system.h and
        explicitly mark extern functions as such.

2017-02-12  Paul Eggert  <eggert@cs.ucla.edu>

        ln: replace destination links more atomically
        If the file B already exists, commands like 'ln -f A B' and
        'cp -fl A B' no longer remove B before creating the new link.
        Instead, they arrange for the new link to replace B atomically.
        This should fix a race condition reported by Mike Crowe (Bug#25680).
        * NEWS, doc/coreutils.texi (cp invocation, ln invocation):
        Document this.
        * bootstrap.conf (gnulib_modules): Add symlinkat.
        * src/copy.c, src/ln.c: Include force-link.h.
        * src/copy.c (same_file_ok): It's also OK to remove a destination
        symlink when creating symbolic links, or when the source and
        destination are on the same file system and when creating hard links.
        * src/copy.c (create_hard_link, copy_internal):
        * src/ln.c (do_link):
        Rewrite using force_linkat and force_symlinkat, to close a window
        where the destination temporarily does not exist.
        * src/cp.c (main): Do not set x.unlink_dest_before_opening
        merely because we are in link-creation mode.
        * src/force-link.c, src/force-link.h: New files.
        * src/local.mk (copy_sources, src_ln_SOURCES): Add them.
        * tests/cp/same-file.sh: Adjust test case to match fixed behavior.

2017-02-10  Tobias Stoeckmann  <tobias@stoeckmann.org>

        timeout: fix race possibly terminating wrong process
        The race is unlikely, as timeout(1) needs to receive a signal
        in the few operations between waitpid() returning and exit().
        Also the system needs to have reallocated the just released pid
        in this time window.

        Previously we never disabled the signal handler that sent
        the termination signal to the "child" pid.  However once waitpid()
        has reaped the child, the system is free to allocate that pid,
        so we must ensure we don't process any further signals.

        * build-aux/gen-lists-of-programs.sh: Build timeout(1) optionally...
        * configure.ac: ...predicated on sigsuspend() being available.
        * src/timeout.c (block_cleanup): A new function to ensure the
        cleanup() handler is disabled after waitpid has returned.
        (main): Use sigsuspend() to wait with cleanup() enabled but
        disabled once it returns, and thus disabled for the waitpid() call.
        (monitored_pid): Change to the more accurate pid_t.
        * NEWS: Mention the fix.

        Fixes http://bugs.gnu.org/25624

2017-02-10  Pádraig Brady  <P@draigBrady.com>

        doc: note the relationship between realpath and readlink
        * doc/coreutils.texi (realpath invocation): Mention that realpath
        is the preferred command for canonicalization.
        (readlink invocation): Likewise.
        * man/readlink.x: Likewise.

2017-02-08  Janne Snabb  <snabb@epipe.com>

        tail: fix output of redundant headers when resuming
        * src/tail.c (check_fspec): Only enable printing of the file header
        if we've actually read some data and this is a new file.  Also
        move printing of the file header to...
        (dump_remainder): ...here, to allow printing only when data read.
        * tests/tail-2/overlay-headers.sh: A new test for suspension
        and resumption of tail.
        * tests/local.mk: Reference the new test.
        * NEWS: Mention the fix.
        Fixes http://bugs.gnu.org/23539

2017-02-08  Pádraig Brady  <P@draigBrady.com>

        tests: fix tail test race causing false failure
        * tests/tail-2/retry.sh: The replacement of the "missing" directory
        is not atomic, and therefore tail(1) can take a different path,
        especially if there is a delay between the rmdir(2) and creat(2).
        This is noticeable for example with `make coverage` because in
        that case the coverage files written by rmdir(1) on exit,
        induce a significant delay thus triggering the issue.

        tail: fix erroneous status about 'giving up' on file
        * src/tail.c (recheck): Set f->ignore before we
        use it to show the appropriate error.
        * tests/tail-2/retry.sh: Ensure the "giving up" message
        is not presented.

2017-02-08  Pádraig Brady  <P@draigBrady.com>

        doc: only distribute 5 years of ChangeLogs
        Remove old log files that have corresponding entries
        in the source code repository.
        This saves about 2.5MB uncompressed, 0.5M compressed.

        * Makefile.am (gen-ChangeLog): Adjust to taking all
        logs since a particular version (8.15 in this case).
        Also mention in the truncated log where to get older entries.
        (changelog_etc): Remove the no longer distributed files.
        * build-aux/git-log-fix: Remove now unused entries.
        * ChangeLog-200[5-8]: Delete.
        * doc/ChangeLog-2007: Likewise.
        * po/ChangeLog-2007: Likewise.
        * old/*: Likewise.

2017-02-07  Maxime de Roucy  <maxime.deroucy@gmail.com>

        maint: fix HACKING instructions to run a single test
        * HACKING: s/make TEST=/make check TEST=/
        The 'check' target was missing there since v8.20-57-geac397e.

2017-01-25  Pádraig Brady  <P@draigBrady.com>

        build: fix issue with HAVE_FALLOCATE on centos5
        * src/copy.c (punch_hole): Work around an empty definition
        of HAVE_FALLOCATE which leads to a build error of:
        "error: #if with no expression"
        That was triggered by the inclusion of <linux/fs.h> in
        commit v8.25-68-g89e1fef with kernel-headers-2.6.18.
        Reported by Nelson H. F. Beebe

2017-01-25  Manolis Ragkousis  <manolis837@gmail.com>  (tiny change)

        build: Properly expand cu_install_program when cross-compiling
        * src/local.mk (cu_install_program): Replace @INSTALL_PROGRAM@
        with @INSTALL@ when cross-compiling; missed in commit 477a1e8e.
        Message-Id: <20170125163329.5690-1-manolis837@gmail.com>

2017-01-21  Mike Swanson  <mikeonthecomputer@gmail.com>

        dircolors: highlight windows archive format
        * src/dircolors.hin: Match *.{wim,swm,dwn,esd}

2017-01-21  Pádraig Brady  <P@draigBrady.com>

        maint: mention the recent date time zone bug fix
        * NEWS: Add the bug fix from commit v8.26-27-gb14be50

2017-01-21  Assaf Gordon  <assafgordon@gmail.com>

        maint: appease syntax-check failures due to recent updates
        * bootstrap: s/time stamp/timestamp/.
        * old/fileutils/NEWS: Likewise.
        * src/tail.c: Avoided a long line.

2017-01-21  Paul Eggert  <eggert@cs.ucla.edu>

        date: fix TZ= regression
        Problem reported by Paul Wise for Debian, in:
        https://bugs.debian.org/851934
        This is fallout from the fix for GNU Bug#23035.
        * src/date.c (batch_convert): New args TZ and TZSTRING.
        All uses changed.
        (batch_convert, main): Adjust to parse_datetime2 API change.
        (main): Allocate time zone object.
        * tests/misc/date-debug.sh: Fix incorrect test case,
        caught by the fix.
        * tests/misc/date.pl: Test the fix.

        build: update gnulib submodule to latest

2017-01-15  Jim Meyering  <meyering@fb.com>

        maint: update README-hacking, now that vc-dwim accepts --init
        * README-hacking: I've just released vc-dwim-1.8, so we can improve
        the documentation to reference its --initialize option here.

2017-01-15  Paul Eggert  <eggert@cs.ucla.edu>

        date: output "-00" for indeterminate time zone
        * NEWS: Document this behavior, which comes with recent Gnulib.
        * doc/coreutils.texi (Formatting file timestamps, du invocation)
        (Time conversion specifiers, Setting the time, Options for date):
        Mention when -00 is output for numeric time zones.
        Be more careful about Internet RFC numbers, ISO 8601, etc.

        date: new option spelling --rfc-email
        * NEWS:
        * doc/coreutils.texi (Time conversion specifiers)
        (Options for date, Examples of date): Document this.
        * src/date.c (rfc_email_format): Rename from rfc_2822_format.
        All uses changed.
        (usage, long_options): Support --rfc-email.

        maint: modernize URLs
        A lot of this is converting http: to https:.
        Also, gmane went away, so remove URLs that no longer work and
        are not easy to figure out what they were.
        Some of this stuff is so old that it no longer matters anyway.

        build: update gnulib submodule to latest

2017-01-10  Bernhard Voelker  <mail@bernhard-voelker.de>

        doc: move "File timestamps" to a separate chapter
        The above new section looked a bit odd as the only general documentation
        in between the utility chapters.

        * doc/coreutils.texi (File timestamps): Move to a separate chapter.

2017-01-10  Eric Blake  <eblake@redhat.com>

        maint: fix recent syntax check failures
        Commit 4f650aad was incomplete; it changed NEWS but not the hash,
        and introduced a grammar error.

        * cfg.mk (old_NEWS_hash): Update via 'make update-NEWS-hash'.
        * doc/coreutils.texi (File timestamps): Fix doubled word.

2017-01-09  Paul Eggert  <eggert@cs.ucla.edu>

        doc: cover file timestamps better
        Prompted by a bug report from Scott Deerwester (Bug#25407).
        * doc/coreutils.texi (File timestamps): New section.
        Revamp other sections to use this new section, and
        use more-consistent terminology.

        maint: standardize on "timestamp" as per POSIX

2017-01-09  Pádraig Brady  <P@draigBrady.com>

        stty: ensure no side effects from invalid options
        * src/stty.c (apply_settings): A new function refactored
        from main() that is used to both check and apply options.
        (main): Call apply_settings before we open the device,
        so all validation is done before interacting with a device.
        * NEWS: Mention the improvement.
        * tests/misc/stty.sh: Add a test case.

2017-01-06  Assaf Gordon  <assafgordon@gmail.com>

        tests: improve 'date --debug' tests
        Update tests following improvements to gnulib's parse-datetime.y module.
        See lib/2017-01/msg00002.html">https://lists.gnu.org/archive/html/bug-gnulib/2017-01/msg00002.html

        * tests/misc/date-debug.sh: Add tests for each of the gnulib changes.

2017-01-06  Assaf Gordon  <assafgordon@gmail.com>

        build: update gnulib submodule to latest

2017-01-01  Pádraig Brady  <P@draigBrady.com>

        maint: update all copyright year number ranges
        Run "make update-copyright" and then...

        * gnulib: Update to latest with copyright year adjusted.
        * tests/init.sh: Sync with gnulib to pick up copyright year.
        * bootstrap: Likewise.
        * tests/sample-test: Adjust to use the single most recent year.

2016-12-28  Zooko  <zookog@gmail.com>

        doc: recommend b2sum as well as SHA2
        b2sum is faster, easier to use safely, and more future-proof

2016-12-28  Pádraig Brady  <P@draigBrady.com>

        tests: avoid false fails on NFS due to EPERM
        * tests/chgrp/basic.sh: On some NFS setups a user is
        not allowed to set a group on a file even if a member
        of that group.  Therefore skip this test on remote file systems.
        * tests/chgrp/default-no-deref.sh: Likewise.
        * tests/chgrp/no-x.sh: Likewise.
        * tests/chgrp/posix-H.sh: Likewise.
        * tests/chgrp/recurse.sh: Likewise.
        * tests/tail-2/inotify-rotate-resources.sh: Change to skipping
        on remote file systems in the standard way.

2016-12-27  Paul Eggert  <eggert@cs.ucla.edu>

        doc: Update POSIX part of README (Bug#25259)

2016-12-26  Pádraig Brady  <P@draigBrady.com>

        wc: with only --bytes, determine size more efficiently
        * src/wc.c (wc): Avoid reading the end of the file
        when the size is not a multiple of PAGE_SIZE,
        as the special case handling for files in /proc and /sys
        is only required when st_size is 0 or a multiple of PAGE_SIZE.
        * tests/misc/wc-proc.sh: Add a test case.

2016-12-20  Pádraig Brady  <P@draigBrady.com>

        maint: correct the version for the previous bug fix
        While st_size would have been incorrect for subsequent
        files since v7.1, it was only used since v8.24.

        * tests/misc/wc-files0.sh: s/7.1/8.24/
        * NEWS: Likewise.

        Reported by Bernhard Voelker

2016-12-19  William R. Fraser  <wfraser@codewise.org>

        wc: fix wrong byte counts when using --files-from0
        * src/wc.c (main): Reset fstatus[0].failed between files when reusing
        the fstatus[0] entry in --files-from0 mode.  This ensures a stat() is
        done for each file, avoiding incorrect counts and redundant reading.
        * NEWS: Mention the bug fix.
        * tests/misc/wc-files0.sh: Add a test case.
        Fixes http://bugs.gnu.org/23073

2016-12-18  Pádraig Brady  <P@draigBrady.com>

        tests: fix typos in previous commit
        * init.cfg (skip_if_mcstransd_is_running_): Fix typos
        _introduced my me_ in the previous commit.

2016-12-18  Nicolas Iooss  <nicolas.iooss@m4x.org>

        tests: support non-MLS enabled SELinux systems
        When running "make check" on a Linux system running SELinux with a
        non-MLS policy, tests/mkdir/restorecon.sh test fails with:

          chcon: invalid context: root:object_r:tmp_t:s0: Invalid argument

        Indeed in such a configuration, contexts cannot have ":s0" suffix.

        * init.cfg (get_selinux_type): Refactor this function to here
        from various tests.  Update to work with a non-MLS policy.
        (mls_enabled_): A new function to detect if MLS is enabled.
        (skip_if_mcstransd_is_running_): Update to not skip when
        MLS is not enabled.
        * tests/mkdir/restorecon.sh: Use a valid non-MLS context when needed.
        * tests/install/install-Z-selinux.sh: Likewise.
        * tests/cp/cp-a-selinux.sh: Likewise.
        * tests/misc/selinux.sh: Likewise.
        * tests/misc/chcon.sh: Skip if non-MLS as --range used throughout.
        Fixes http://bugs.gnu.org/22631

2016-12-08  Torbjörn Granlund  <tg@gmplib.org>

        factor: retry properly if Pollard rho gives a trivial factorization
        * src/factor.c (factor_using_pollard_rho): Handle trivial factor g = n.
        (factor_using_pollard_rho2): Handle trivial factor g1 = n1, g0 = n0.
        * tests/misc/factor.pl: Add a test case.
        Fixes http://bugs.gnu.org/25135

2016-12-08  Niels Möller  <nisse@lysator.liu.se>

        factor: fix infinite loop in gcd2_odd
        * src/factor.c (gcd2_odd): Fix the case a1 == 0, a0 == 0.
        * NEWS: Mention the bug fix.
        Fixes http://bugs.gnu.org/25135

2016-12-02  Pádraig Brady  <P@draigBrady.com>

        doc: fix --help for: od -t f[SIZE]
        * src/od.c (usage): SIZE is that of float, not integer.

(wen)

Updated editors/joe to 4.6

(wen)

Update to 4.6

Upstream changes:
JOE 4.6
Enhancements

Upgrade to Unicode 10.0.0. Add configure environment variable that allows you select Unicode 8.0.0, 9.0.0 or 10.0.0.

Install Gnome .desktop files for JOE. This allows you to use JOE in GUI "Open With" mouse actions.

Support strikeout and double-underline attributes for the few terminal emulators that support them. Use "stricken" and "dunderline" in the syntax files, or \s and \z in help screens and status line.

Bugs fixed

Fix bug where JOE would sometimes crash when editing shell scripts. This was due to an obscure bug in the syntax highlighter: 'reset' command (used for default state) was incorrectly messing with stack.

-highlighter_context was missing from many file types, which negated some improvements from the previous version.

Fixed signed char overflow with old style mouse events in large windows

Fix bug where path restart (//) was being applied to block filter command prompt. Strange things would happen if you had adjacent slashes in command arguments.

Allow ~ expansion but suppress path restart (//) in compiler error parsing.

Restore default handling of SIGPIPE and SIGINT for shell commands. This fixes an issue where SIGPIPE doesn't terminate a process as expected, for example by the head -n 10 in: ^K R !sh -c 'while :; do echo y; done' | head -n 10. This issue only occurred in read/write to !, JOE already did the right thing for shell windows and the filter region through shell command.

Improve screen update algorithm so that spaces at the ends of lines are always emitted. This allows them to be preserved when cutting text with the mouse from a terminal emulator window.

Improve efficiency of screen update algorithm: JOE had been resetting attributes such as background color before performing cursor motions (probably as work around for bugs in old terminal emulators). This made screen update slow when there were many attributes, as with syntax highlighting and color schemes. It was also repeatedly emitting ESC [ K.

Switch JOE to issue scrolling commands, even at high baud rates (before this, JOE issued scrolling commands only at 19200 and below because it used to be that simple screen refresh was faster than scrolling in terminal emulators. But this is no longer true with complex screens involving color schemes, unicode and highlighting).

Windows version

Fix file access checks. Inaccessible files due to ACL's weren't correctly detected. This was a regression in 4.5.

Check the correct (Windows) path in the home directory for color schemes. This was not re-implemented properly in 4.5.

JOE 4.5

Enhancements

New color scheme feature, which can be accessed with ^T S or by setting the -colors option.

24-bit color support can be enabled by setting the COLORTERM environment variable to truecolor or 24bit.

Several schemes included in-the-box:

gruvbox by Pavel Pertsev

ir_black by Todd Werth

molokai by Tomas Restrepo

solarized by Ethan Schoonover

wombat by Lars H. Nielsen

xoria by Dmitriy Y. Zotikov

zenburn by Jani Nurminen

The current line can be highlighted by pressing ^T U or by enabling the -hiline option.

The gutter containing line numbers has a dynamic size based on the length of the file, rather than a fixed size of 10.

Updated all language syntax files to use comment_todo and string/comment contexts where appropriate.

Now pass character which invoked a macro to each macro step and call. If a macro step happens to be the 'type' command, the character which invoked the macro will be typed in. For example, this macro will type three 'X's. Before this change you got three NULs.

type,type,type X

Bugs fixed

Fix exsave: (^K ^X) should close file when a block is present in the window, and the file is unmodified (regression from ^C change in 4.2).

Fix regex assertions: they were not working because the character before the search position was not being loaded.

For jmacs: ^Q^J now again inserts \n in the string replace prompt. This broke beginning with JOE 4.0.

Don't try to open files for writing to check file access. This behaves better on unionfs mounts.

Fix stray blocks created after find/replace.

Windows version

Fix inability to set indent step to 1 from menu.

Add support for math functions, fix engineering display.

Fix backslash escapes in find/replace.

Support italicized text.

JOE 4.4
Enhancements

Bugs fixed

Fix segfault due to buffer overrun. This happens if a line with many backslashes appears in the status line context display.

Fix jmacs: ^X ^F and ^X ^B were not working

Build fixes for Solaris

Improve php highlighter: allow numbers in substitution variable names

Unicode tweak: treat private use characters (Co) as printable

Dockerfile highlighter: Add Docker new commands from 1.12, mark bad strings in arrays

Fix loading external charmaps

Windows version

Fix crashing bug when using incremental search
JOE for Windows 4.3
Download

Bugs fixed

Fixed a missed merge that prevented some options menu items from being changed (tab width, tab char, etc).

Fixed bugs updating the title bar.

Minor memory leak/performance fix in subprocess communication.

(from newer version) Fix segfault due to buffer overrun. This happens if you a line with many backslashes appears in the status line context display.

JOE 4.3
Enhancements

Improve memory usage by shrinking buffer header size and highlighter state size.

Improve performance of status line context display (which shows the first line of the function that the cursor is currently in). This feature was making JOE very slow on extremely large files with auto indent enabled (typically JSON or XML data files). Now the syntax highlighter computes the context display (using a new syntax named context.jsf).

Add a mode 'title' to enable or disable the status line context display (previously autoindent mode was overloaded to do this).

Disable syntax highlighting and context display in very large files

Force more appropriate modes when we enter hex dump display: enable overtype, disable autoindent, wordwrap, ansi, picture.

Handle middle mouse button in "joe -mouse" mode (before it did nothing). It's treated as paste (copy region to mouse) as expected. (patch from Petr Olsak).

Bugs fixed

Do not kill region highlighting during incremental search (patch from Petr Olsak).

Negative numbers were not being recognized in blocks

PgUp/PgDn would try to scroll menu if the window above is a menu (it should do this only for completion menus associated with prompts)

Use 'LC_ALL=C sed' to get JOE to compile in OS X.

Forward direction delimiter matching where the delimiters do not begin with special characters (for example in Verilog "begin" / "end") was not working.

Get mouse to work in menus: this broke in 4.1

Character classes with ranges were not working for UTF-8 (as in \[a-z])

Apply spec highlighting to .spec files

Gracefully handle short terminals: fix segfaults which occur when trying to shrink terminal while many windows are on the screen or while turning on help with a short terminal. Fix similar bugs involving the skiptop option. JOE now works even if the terminal height is only one line.

JOE 4.2
New or improved syntax files for the following languages:

Dockerfile
Usability Enhancements

The top Google help searches for JOE include:

How do I save and exit? The startup copyright notice has been replaced with basic help for beginners: ^K Q to exit and ^K H for help.

How do I dismiss the region highlighting? The traditional way is to hit ^K B ^K K, but this is slightly non-obvious and has always been awkward. Now Ctrl-C will do it.

How do I close all files and exit? Now Ctrl-K Q does this. Previously ^K Q was the same as ^C: abort a single file.

Restyle the help screens:

Make it more obvious that there is more than one screen: put the help for help inline with the text instead of in the (seemingly invisible) header.

Mention Ctrl-Arrow region selection, status and goto matching delimiter commands on the first screen.

Remove time and "Ctrl-K H for help" message from status bar. Beginners often don't notice this help message, and it takes up valuable status bar space that power users want for the context display.

Add ^KH for help to search and replace prompts. Many JOE users do not know about this context sensitive help.

Provide aborthint and helphint options so that the ^C and ^K H hints can be customized depending on the rc file (so say ^G for abort in jmacs, for example).

Enable -noxon by default (disable ^S/^Q flow control). This allows us to bind ^Q to quote and ^S / ^R to incremental search.

Document ESC X (command prompt) in the help screens.

"joe --help" now prints all command line options.

Other Enhancements

Tags search now tries to find the tags file in parent directories if it does not exist in the current directory and if the TAGS environment variable was not set.

Built-in calculator can now print and accept numbers in binary, octal and engineering formats: dec 12_345 eng 12.345_0e3 bin 0b11_0000_0011_1001 oct 0o3_0071 hex 0x3039

Built-in calculator now prints and accepts separating underscores for clarity. For example, 4_294_967_296 instead of 4294967296.

Enhanced calculator statistics functions:

dev computes standard deviation with full population
samp computes standard deviation with sample of population
Linear regression analysis. Select a region of x and y values, then:
lr(x) provide estimate of y given x
rlr(y) provide estimate of x given y
Lr, lR, LR: log, exponential, power regression
Calculator region functions now assume the entire buffer if no region is set.

Tab completion now works at the calculator prompt (and in all prompts which allow numeric input, such as ^KL- go to line).

Make new regex engine (from JOE 4.1) more compatible with the classic engine. \y is now shorthand for \(\.\*\), so that it does what \* did in the old engine. Also:

\. no longer matches newline.
\* matches shortest match, not longest match.
Add -left and -right options to control the amount scrolling when the cursor moves past the left or right edge of the screen. When the baud rate is low these are automatically set to a large amount. Also, these now control the manual horizontal scrolling commands. When these are positive, they indicate number of columns. When they are negative, they indicate a fraction of the screen width (-2 is 1/2 the width).

Bugs fixed

Fix use after free bug which shows up as a crash in OpenBSD

Fix bug where indent step value was not shown on ^T menu

Fix bug where setting margin doesn't work on big-endian systems

Fix issue where highest valued Unicode character equivalent was not translating to its corresponding 8-bit character. Effect of this was that Delete key was not working in shell windows in ASCII character set.

Standard deviation calculator function was not producing correct results.

Allow koi8r and koi8-r for KOI8-R in joe_getcodeset (which is only used if there is no setlocale).

Guess_crlf forced UNIX line endings for new files even though crlf was set. Now crlf is left alone if guess_crlf can not determine the line ending.

If cursor was at end of a long line and you switched to hex dump display mode, then hex dump was scrolled. Now scroll offset is reset when you switch to hex display mode.

(wen)