Updated devel/p5-Future to 0.35

(wen)

Update to 0.35

Upstream changes:
0.35    2017/06/23 20:37:57
        [CHANGES]
        * Link to YAPC::EU talk video in SEE ALSO

        [BUGFIXES]
        * Handle cancelled/completed slots when fmap is cancelled (RT120469)
          (thanks tm604)

(wen)

Updated databases/p5-DBD-mysq to 4.043

(wen)

Update to 4.043

Upstream changes:
2017-06-29 Patrick Galbraith, Michiel Beijen, DBI/DBD community (4.043)

YOUR ATTENTION PLEASE, THIS IS A REVERT TO 4.041
This version is the same as 4.041 with all its bugs and
limitations. In version 4.042 there were some changes to Unicode handling
that turned out to be causing issues with existing implementations.
While it is possible to argue that the old behaviour was wrong and buggy,
lots of applications and scripts were depending on this behaviour so it
is NOT a good idea to change this.

There were lots of commits since 4.041, we'll add those back bit by bit
in a future release, excluding the ones which cause problems.

2017-??-?? Patrick Galbraith, Michiel Beijen, DBI/DBD community (4.042_01)
* Use Devel::CheckLib 1.09 or newer, fixes
  https://github.com/perl5-dbi/DBD-mysql/issues/109
* Improve CI testing on AppVeyor: caching, path to cpan, configure deps (pali)
* Specify bigint as test dependency.

2017-03-08 Patrick Galbraith, Michiel Beijen, DBI/DBD community (4.042)
* Full release to include development releases 4.041_2 and 4.041_1.

2017-02-28 Patrick Galbraith, Michiel Beijen, DBI/DBD community (4.041_2)
* Statement handle now also has mysql_sock attribute, just as database handle.
  (by Pali)
* Fix type conversions for magic types. Issue reported by Dmitriy Yatsenko and
  Giovanni Bechis, fix by Pali.
    https://lists.amavis.org/pipermail/amavis-users/2016-December/004674.html
    https://github.com/perl5-dbi/DBD-mysql/issues/78
* Fix UTF8-encoding of table names, field names, warnings and error messages.
  Reported by Tanabe Yoshinori, fix by Pali.
    https://rt.cpan.org/Public/Bug/Display.html?id=120141
* Fix mysql_auto_reconnect when using mysql_server_prepare (pali). Reported by
  Vladimir Marek.
    https://github.com/perl5-dbi/DBD-mysql/pull/95
* Improve regex for removing database from dsn (pali)
    https://rt.cpan.org/Public/Bug/Display.html?id=118837
* Locate MySQL libs using Devel::CheckLib (pali)
* Support async on Windows (pali)

* Fix test suite on range of older and newer MySQL and MariaDB versions
  (https://github.com/perl5-dbi/DBD-mysql/pull/87)
* Fix compilation on MySQL 4.1 (pali)
* Do not leak dangling pointer to mysql result (pali)
* Fix logic when assigning to variable bind_comment_placeholders (pali)
* mysql_fd() still returned file descriptor after closing connection.
  Reported by Larry Leszczynski, fixed by Pali Roh獺r.
  (https://rt.cpan.org/Public/Bug/Display.html?id=110983)
* Fix parsing configure libs from mysql_config --libs output in Makefile.PL
  Libraries in mysql_config --libs output can be specified by library name
  with the -l prefix or by absolute path to library name without any prefix.
  Parameters must start with a hyphen, so treat all options without leading
  hyphen in mysql_config --libs output as libraries with full path.
  Partially fixes bug https://rt.cpan.org/Public/Bug/Display.html?id=100898
  Fix by Pali Roh獺r.
* Fix support for magic scalars (pali)
  (https://github.com/perl5-dbi/DBD-mysql/pull/76)

2016-12-12 Patrick Galbraith, Michiel Beijen, DBI/DBD community (4.041_1)
* Unicode fixes: when using mysql_enable_utf8 or mysql_enable_utf8mb4,
  previous versions of DBD::mysql did not properly encode input statements
  to UTF-8 and retrieved columns were always UTF-8 decoded regardless of the
  column charset.
  Fix by Pali Roh獺r.
  Reported and feedback on fix by Marc Lehmann
  (https://rt.cpan.org/Public/Bug/Display.html?id=87428)
  Also, the UTF-8 flag was not set for decoded data:
  (https://rt.cpan.org/Public/Bug/Display.html?id=53130)
* Return INTs with ZEROFILL as strings. Reported by Knarf, fix by Pali Roh獺r.
  (https://rt.cpan.org/Public/Bug/Display.html?id=118977)

(wen)

Updated editors/vim-share to 8.0.0702

(wen)

Update to 8.0.0702

Upstream changes:
17435  8.0.0647  syntax highlighting can cause a freeze
  1629  8.0.0648  possible use of NULL pointer
  3822  8.0.0649  when opening a help file the filetype is set several times
  2039  8.0.0650  for extra help files the filetype is set more than once
  1534  8.0.0651  build failure without the auto command feature
10036  8.0.0652  unicode information is outdated
  2013  8.0.0653  the default highlight for QuickFixLine is not good
  5188  8.0.0654  no warning for text after :endfunction
  2677  8.0.0655  not easy to make sure a function does not exist
  2329  8.0.0656  cannot use ! after some user commands
13104  8.0.0657  cannot get and set quickfix list items
47863  8.0.0658  spell test is old style
  5884  8.0.0659  no test for conceal mode
  1676  8.0.0660  silent install on MS-Windows shows dialog
10750  8.0.0661  recognizing urxvt mouse codes does not work well
  1828  8.0.0662  stray FIXME for fixed problem
  5971  8.0.0663  unexpected error message only when 'verbose' is set
  2373  8.0.0664  mouse does not work in tmux
  1517  8.0.0665  warning for uninitialized variable
  4110  8.0.0666  dead for loop
  5896  8.0.0667  memory access error when command follows :endfunc
  1689  8.0.0668  nsis installer script does not work
  2927  8.0.0669  CTRL-N at start of the buffer does not work correctly
  3607  8.0.0670  can't use input() in a timer callback
  1600  8.0.0671  hang when typing CTRL-C in confirm() in timer
  6001  8.0.0672  third item of synconcealed() changes too often
  1260  8.0.0673  build failure without conceal feature
  1751  8.0.0674  cannot build with eval but without timers
  2638  8.0.0675  'colorcolumn' has a higher priority than 'hlsearch'
  8830  8.0.0676  crash when closing quickfix window in autocmd
  4177  8.0.0677  setting 'filetype' may switch buffers
  3772  8.0.0678  closing a window does not trigger resizing
  2223  8.0.0679  using freed memory
10068  8.0.0680  plugins in start packages are sourced twice
  2323  8.0.0681  unnamed register only contains the last deleted text
  2786  8.0.0682  no test for synIDtrans()
  4871  8.0.0683  visual bell flashes too quickly
13990  8.0.0684  old style tests are not nice
28918  8.0.0685  when conversion fails written file may be truncated
  1479  8.0.0686  extra redraw when using CTRL-L in second window
  6644  8.0.0687  minor issues related to quickfix
  4317  8.0.0688  cannot resize the window in a FileType autocommand
  2402  8.0.0689  ~ character not escaped when extending search pattern
  2150  8.0.0690  compiler warning on non-Unix system
  1267  8.0.0691  compiler warning without the linebreak feature
  2964  8.0.0692  CTRL-G with 'incsearch' and ? goes in the wrong direction
361923  8.0.0693  no terminal emulator support
  2396  8.0.0694  building in shadow directory does not work
  5003  8.0.0695  missing dependencies breaks parallel make
  1309  8.0.0696  .inc files missing in git
  9396  8.0.0697  recorded key sequences may become invalid
  4219  8.0.0698  crash on exit when using Python function in timer.
  1742  8.0.0699  checksum tests are not actually run
  2709  8.0.0700  segfault with QuitPre autocommand closes the window
  2900  8.0.0701  system test failing when using X11 forwarding
  5191  8.0.0702  an error in a timer can make Vim unusable
  3118  8.0.0703  illegal memory access with empty :doau command
  2145  8.0.0704  problems with autocommands when opening help

(wen)

Always stop libstdc++ from using PCH, not just on bootstrap.

It fails and hangs on netbsd, and people who have looked at the PCH
implementation blame it rather than the OS.

whitespace.

(maya)

Normalize patch filenames. No functional change.

(schmonz)

Updated textproc/xapian-omega to 1.4.4

(schmonz)

Updated textproc/xapian to 1.4.4

(schmonz)

Update to 1.4.4. From the changelog:

indexers:

* omindex:

  + 1.4.3 added a new --sample option, but contrary to the documentation
    the default behaviour was to take the sample from the meta description
    (which was the hard-wired behaviour in 1.4.2 and earlier).  The default
    has now been changed to take the sample from the body.

  + Index .shtm, .xhtml and .xhtm as HTML by default - .shtm is another
    extension used for server-parsed HTML (in addition to the more common
    .shtml), and .xhtm and .xhtml are XHTML.

  + Fix fallback lookup for extension containing upper case.  User mappings
    worked, but built-in extension to MIME type mappings were effectively being
    ignored (because the result of the function call was not being checked).
    Bug introduced in 1.3.4.

  + Fix term-based date ranges, broken by changes in 1.4.2.  Found and
    diagnosed by Gaurav Arora.

  + Handle date range with start after end better - with term-based ranges,
    this used to generate a bogus filter, but now just generates Dlatest.

  + Use Y-term when range starts/ends at year start/end.  Previously we used 12
    M-terms for these cases.

  + Use full leap-year check when constructing term-based date ranges -
    previous code was good until 2100, but even then it would only result
    in an extra term being included for a non-existent February 29th in
    rare cases.

  + Add support for indexing vCard files if Perl and its Text::vCard module
    are available.

  + Recognise application/x-rpm as alternative type since libmagic reports this
    rather than application/x-redhat-package-manager.

  + Use official MIME type application/vnd.debian.binary-package for debian
    packages.  We used to map .deb and .udeb to application/x-debian-package,
    but in 2014 (after we added that support for .deb) an official type was
    registered with IANA.  We now map extensions .deb and .udeb to the official
    type, but the unofficial type is still recognised (older versions of
    libmagic probably report it, and users may be mapping to it).

  + Handle PHP as MIME type text/x-php.  The main difference this makes is that
    PHP files which don't have extension '.php' (e.g. .phtml, .phps, .php5,
    .ph4, etc) get identified by libmagic as text/x-php and will now be indexed.
    It also means that the user can now more easily configure different filters
    for HTML and PHP.

  + Don't use meta description as sample by default.  Now we have dynamic
    snippets (via $snippet), the body text is a better default.  Also generated
    HTML sometimes has unhelpful content in the meta description.  To get the
    previous behaviour, use the new omindex command line option:
    --sample=description

omega:

* New OmegaScript command $cgiparams which returns a list of the parameter
  names.

* Handle tab in a CGI parameter name in the same way as space.  Mostly this is
  a way to avoid having tabs in CGI parameter names - they aren't useful, but
  if they could have tabs in we can't put CGI parameter names in a list.

templates:

* query: Fix highlighting of matching terms.  We were using both $snippet and
  $highlight, which results in double highlighting and HTML escaping, most
  noticeable by literal <strong> and </strong> appearing around matching terms
  in the rendered HTML snippet.  Reported by Mark Thomas on xapian-discuss.

build system:

* If gen-mimemap failed after creating mimemap.h, the rule wouldn't get rerun.

(schmonz)

Update to 1.4.4. From the changelog:

API:

* Database::check():

  + Fix checking a single table - changes in 1.4.2 broke such checks unless you
    specified the table without any extension.

  + Errors from failing to find the file specified are now thrown as
    DatabaseOpeningError (was DatabaseError, of which DatabaseOpeningError is
    a subclass so existing code should continue to work).  Also improved the
    error message when the file doesn't exist is better.

* Drop OP_SCALE_WEIGHT over OP_VALUE_RANGE, OP_VALUE_GE and OP_VALUE_LE in the
  Query constructor.  These operators always return weight 0 so OP_SCALE_WEIGHT
  over them has no effect.  Eliminating it at query construction time is cheap
  (we only need to check the type of the subquery), eliminates the confusing
  "0 * " from the query description, and means the OP_SCALE_WEIGHT Query object
  can be released sooner.  Inspired by Shivanshu Chauhan asking about the query
  description on IRC.

* Drop OP_SCALE_WEIGHT on the right side of OP_AND_NOT in the Query
  constructor.  OP_AND_NOT takes no weight from the right so OP_SCALE_WEIGHT
  has no effect there.  Eliminating it at query construction time is cheap
  (just need to check the subquery's type), eliminates the confusing "0 * "
  from the query description, and means the OP_SCALE_WEIGHT object can be
  released sooner.

* MSet::snippet(): Favour candidate snippets which contain more of a diversity
  of matching terms by discounting the relevance of repeated terms using an
  exponential decay.  A snippet which contains more terms from the query is
  likely to be better than one which contains the same term or terms multiple
  times, but a repeated term is still interesting, just less with each
  additional appearance.  Diversity issue highlighted by Robert Stepanek's
  patch in https://github.com/xapian/xapian/pull/117 - testcases taken from his
  patch.

* MSet::snippet(): New flag SNIPPET_EMPTY_WITHOUT_MATCH to get an empty snippet
  if there are no matches in the text passed in.  Implemented by Robert
  Stepanek.

* Round MSet::get_matches_estimated() to an appropriate number of significant
  figures.  The algorithm used looks at the lower and upper bound and where the
  estimate sits between them, and then picks an appropriate number of
  significant figures.  Thanks to S辿bastien Le Callonnec for help sorting out a
  portability issue on OS X.

* Add Database::locked() method - where possible this non-invasively checks if
  the database is currently open for writing, which can be useful for
  dashboards and other status reporting tools.

testsuite:

* Add more tests of Database::check().  Fixes #238, reported by Richard
  Boulton.

* Make apitest testcase nosuchdb1 fail if we manage to open the DB.

* Skip testcases which throw NetworkError with errno value ECHILD - this
  indicates system resource starvation rather than a Xapian bug.  Such failures
  are seen on Debian buildds from time to time, see:
  https://bugs.debian.org/681941

* Use terms that exist in the database for most snippet tests.  It's good to
  test that snippet highlighting works for terms that aren't in the database,
  but it's not good for all our snippet tests to feature such terms - it's
  not the common usage.

matcher:

* Fix incorrect results due to uninitialised memory.  The array holding max
  weight values in MultiAndPostList is never initialised if the operator is
  unweighted, but the values are still used to calculate the max weight to pass
  to subqueries, leading to incorrect results.  This can be observed with an OR
  under an unweighted AND (e.g. OR under AND on the right side of AND_NOT).
  The fix applied is to simply default initialise this array, which should lead
  to a max weight of 0.0 being passed on to subqueries.  Bug reported in
  notmuch by Kirill A. Shutemov, and forwarded by David Bremner.

* Improve value range upper bound and estimated matches.  The value slot
  frequency provides a tighter upper bound than Database::get_doccount().
  The estimate is now calculated by working out the proportion of possible
  values between the slot lower and upper bounds which the range covers
  (assuming a uniform distribution).  This seems to work fairly well in
  practice, and is certainly better than the crude estimate we were using:
  Database::get_doccount() / 2

* Handle arbitrary combinations of OP_OR under OP_NEAR/OP_PHRASE, partly
  addressing #508.  Thanks to Jean-Francois Dockes for motivation and testing.

* Only convert OP_PHRASE to OP_AND if full DB has no positions.  Until now the
  conversion was done independently for each sub-database, but being consistent
  with the results from a database containing all the same documents seems more
  useful.

* Avoid double get_wdf() call for first subquery of OP_NEAR and OP_PHRASE,
  which will speed them up by a small amount.

documentation:

* Correct "Query::feature_flag" -> "QueryParser::feature_flag".  Fixes #747,
  reported by James Aylett.

* Rename set_metadata() `value` parameter to `metadata`.  This change is
  particularly motivated by making it easier to map this case specially in SWIG
  bindings, but the new name is also clearer and better documents its purpose.

* Rename value range parameters.  The new names (`range_limit` instead of
  `limit`, `range_lower` instead of `begin` and `range_upper` instead of `end`)
  are particularly motivated by making it easier to map them specially in SWIG
  bindings, but they're also clearer names which better document their
  purposes.

* Change "(key, tag)" to "(key, value)" in user metadata docs.  The user
  metadata is essentially what's often called a "key-value store" so users
  are likely to be familiar with that terminology.

* Consistently name parameter of Weight::unserialise() overridden forms.
  In xapian/weight.h it was almost always named `serialised`, but LMWeight
  named it `s` and CoordWeight omitted the name.

* Fix various minor documentation comment typos.

* INSTALL: Update section about -Bsymbolic-functions which is not a new
  GNU ld feature at this point.

tools:

* xapian-delve: Uses new Database::locked() method to report if the database
  is currently locked.

portability:

* Fix configure probe for __builtin_exp10() to work around bug on mingw - there
  GCC generates a call to exp10() for __builtin_exp10() but there is no exp10()
  function in the C library, so we get a link failure.  Use a full link test
  instead to avoid this issue.  Reported by Mario Emmenlauer on xapian-devel.

* Fix configure probe for log2() which was failing on at least some platforms
  due to ambiguity between overloaded forms of log2().  Make the probe
  explicitly check for log2(double) to avoid this problem.

* Workaround the unhelpful semantics of AI_ADDRCONFIG on platforms which follow
  the old RFC instead of POSIX (such as Linux) - if only loopback networking is
  configured, localhost won't resolve by name or IP address, which causes
  testsuites using the remote backend over localhost to fail in auto-build
  environments which deliberately disable networking during builds.  The
  workaround implemented is to check if the hostname is "::1", "127.0.0.1" or
  "localhost" and disable AI_ADDRCONFIG for these.  This doesn't catch all
  possible ways to specify localhost, but should catch all the ways these might
  be specified in a testsuite.  Fixes https://bugs.debian.org/853107, reported
  by Daniel Schepler and the root cause uncovered by James Clarke.

* Fix build failure cross-compiling for android due to not pulling in header
  for errno.

* Fix compiler warnings.

debug code:

* Adjust assertion in InMemoryPostList.  Calling skip_to() is fine when the
  postlist hasn't been started yet (but the assertion was failing for a term
  not in the database).  Latent bug, triggered by testcases complexphrase1 and
  complexnear1 as updated for addition of support for OP_OR subqueries of
  OP_PHRASE/OP_NEAR.

(schmonz)

Updated mail/spamassassin to 3.4.1nb7

(schmonz)

Apply upstream patch (SVN rev 1791010) to fix "Unescaped left brace in
regex is deprecated here (and will be fatal in Perl 5.30)". While here,
remove references to LOCALBASE. Bump PKGREVISION.

(schmonz)

Updated www/py-paste to 2.0.3, www/py-waitress to 1.0.2

(adam)

1.0.2:

Features
- Python 3.6 is now officially supported in Waitress

Bugfixes
- Add a work-around for libc issue on Linux not following the documented
  standards. If getnameinfo() fails because of DNS not being available it
  should return the IP address instead of the reverse DNS entry, however
  instead getnameinfo() raises. We catch this, and ask getnameinfo()
  for the same information again, explicitly asking for IP address instead of
  reverse DNS hostname.

(adam)

2.0.3
-----
* 26: Change six requirement to >=1.4.0
* 28: Py3k fixes
* 29: paste.wsgilib.add_close: Add __next__ method to support using `add_close` objects as iterators on Python 3.
* 30: tox.ini: Add py35 to envlist
* 31: Enable testing with pypy
* 33: tox.ini: Measure test coveraage

(adam)

Simplify github usage.

(wiz)

Better shell quoting.

(wiz)

Fill in DESCR.

(wiz)

Updated graphics/libimagequant to 2.10.1

(adam)

2.10.1:
Bug fixes

(adam)

Updated graphics/GraphicsMagick to 1.3.26

(adam)

1.3.26:

Security Fixes:
---------------
DPX: Fix excessive use of memory (DOS issue) due to file header claiming large image dimensions but insufficient backing data. (CVE-2017-10799).
JNG: Fix memory leak when reading invalid JNG image (CVE-2017-8350).
MAT: Fix excessive use of memory (DOS issue) due to continuing processing with insufficient data and claimed large image size. Verify each file extent to make sure that it is within range of file size. (CVE-2017-10800).
META: Fix heap overflow while parsing 8BIM chunk (CVE-2016-7800).
PCX: Fix denial of service issue.
RLE: Fix abnomally slow operation (denial of service issue) with intentionally corrupt colormapped file.
PICT: Fix possible buffer overflow vulnerability given suitably truncated input file.
PNG: Enforce spec requirement that the dimensions of the JPEG embedded in a JDAT chunk must match the JHDR dimensions (CVE-2016-9830).
PNG: Avoid NULL dereference when MAGN chunk processing fails.
SCT: Fix stack-buffer read overflow (underflow?) while reading SCT header.
SGI: Fix denial of service issues. Delay large memory allocations until file header has fully passed sanity checks.
TIFF: Fix out of bounds read when reading CMYKA TIFF which claims to have only 2 samples per pixel (CVE-2017-6335).
TIFF: Fix out of bounds read when reading RGB TIFF which claims to have only 1 sample per pixel (CVE-2017-10794).
WPG: Fix heap overflow (CVE-2016-7996). Fix assertion crash (CVE-2016-7997).

Bug fixes:
----------
DifferenceImage(): Fix Fix all-black difference image if an input file is colormapped.
EXIF orientation was not being properly detected for some files.
-frame: The import command -frame handling was improperly implemented and was using already freed data.
GIF: Fixes for "Excessive LZW string data" problem.
Magick++: Bug fixes to PathSmoothCurvetoRel::operator() and PathSmoothCurvetoRel::operator().
PAM: Support writing GRAYSCALE PAM format.
PNG: Fix memory leaks.
SVG: Fixed a memory leak. Fixed a possible null pointer dereference.
TclMagick: Problem that TkMagick could not resolve functions from TclMagick under Linux is fixed.
TclMagick: Fix parser validatation in magickCmd() to avoid crash given a syntax error.
TIFF: Fix for reading old JPEG files (avoids "Improper call to JPEG library in state 0. (LibJpeg).").
TXT: Fixed memory leak.
XCF: Error checking is improved.

New Features:
-------------
EXIF rotation: Support is added such that the EXIF orientation tag is updated when the image is rotated.
MAT: Now support reading multiple images from Matlab V4 format.
Magick++: Orientation method now updates orientation in EXIF profile, if it exists.
Magick++: Added Image attribute method which accepts a 'char *' argument, and will remove the attribute if the value argument is NULL.
-orient: The -orient command line option now also updates the orientation in the EXIF profile, if it exists.
PGX: Support PGX JPEG 2000 format for reading and writing (within the bounds of what JasPer supports).
Wand API: Added MagickAutoOrientImage(), MagickGetImageOrientation(), MagickSetImageOrientation(), MagickRemoveImageOption(), and MagickClearException().

(adam)

Rely on PKG_HAVE_KQUEUE to tell us whether the OS supports kqueue(2).
Fixes build on Darwin 10.12.5 (and probably Bitrig and MirBSD, which had
been excluded from ONLY_FOR_PLATFORM).

(schmonz)

Define PKG_HAVE_KQUEUE where available.

(schmonz)

Fix build with PKGSRC_USE_SSP, and support PKGSRC_USE_FORTIFY.
While here, remove reference to LOCALBASE.

(schmonz)

Updated textproc/p5-XML-Tidy to 1.18

(wen)

Update to 1.18

Upstream changes:
- 1.18 H78M5qm1 Sat Jul 8 05:52:48:01 -0500 2017
      * fixed new() to check file or xml to detect standalone in
      declaration, from <HTTPS://RT.CPAN.Org/Ticket/Display.html?id=122389>
      (Thanks Alex!)

      * traced tidy() memory leak from
      <HTTPS://RT.CPAN.Org/Ticket/Display.html?id=120296> (Thanks Jozef!)
      which seems to come from every XPath->findnodes() call

      * aligned synopsis comments

      * updated write() to use output encoding UTF-8 since that's what
      almost all XML should rely on (with thanks to RJBS for teaching me
      much from his great talk at <HTTPS://YouTube.Com/watch?v=TmTeXcEixEg>)

      * collapsed trailing curly braces on code blocks

      * added croak for any failed file open attempt

(wen)

Updated textproc/p5-XML-Compile-SOAP to 3.22

(wen)

Update to 3.22

Upstream changes:
version 3.22: Fri 30 Jun 10:03:10 CEST 2017

Fixes:
- ::XOP::Include read from file always died
  rt.cpan.org#119955 [Pavel Trushkin]
- ::XOP::Include read should enforce raw mode
  rt.cpan.org#119955 [Pavel Trushkin]

(wen)

Updated textproc/p5-XML-Compile to 1.58

(wen)

Update to 1.58
Update DEPENDS

Upstream changes:
version 1.58: Tue 27 Jun 16:50:29 CEST 2017

Fixes:
- early facet on missing field [Bernhard Reutner-Fischer]

Improvements:
- move to Log::Report 1.20, which has considerable changes.

version 1.57: Wed 14 Jun 14:48:18 CEST 2017

Fixes:
- better separation between lexical- and value-space facets.
  rt.cpan.org#121946 [Nils Barkald]
- json_friendly changes broke some (semi-illegal) enumeration and
  pattern facets.  Now new solution with dualvar [Wesley Schwengle]

(wen)

Updated devel/p5-Log-Report to 1.21

(wen)

Update to 1.21
Update DEPENDS

Upstream changes:
version 1.21: Mon  3 Jul 15:31:19 CEST 2017

Fixes:
- _prepend and _append texts doubled.
  rt.cpan.org#122304 [Andreas Koenig]
- accidental stringification of exceptions
  rt.cpan.org#122324 [Slaven Rezic]

Improvements:
- also \n of msgid moves to _prepend or _append

version 1.20: Tue 27 Jun 16:41:36 CEST 2017

Fixes:
- Dancer2 sporadic missing request. [Andrew Beverley]
- attribute _lang overrules default locale in translation
- remove Log::Report::Lexicon dependency from ::Translator
- formatter PRINTP cannot be used, remove docs which tell that
- exceptions triggered translation too often

Improvements:
- add textdomain($name, 'EXISTS')
- add textdomain($domain_object);
- short-cut when translating without context
- use String::Print::printi() to interpolate: that code was forked-off
  earlier, and now mature.

(wen)

Update DEPENDS, which forgot in last commit

(wen)

Updated devel/p5-Log-Report-Optional to 1.03

(wen)

Updated devel/p5-Test-BDD-Cucumber to 0.53

(schmonz)

Update to 0.53. From the changelog:

- Moose -> Moo, thanks to https://github.com/vti

(schmonz)

Update to 1.03

Upstream changes:
version 1.03: Tue 27 Jun 16:39:41 CEST 2017

Fixes:
- manpage, change SYNOPSYS into SYNOPSIS
  rt.cpan.org#113351 [Shlomi Fish]
- formatter PRINTP cannot be used

Improvements:
- more flexible formatter construction
- textdomain compatible with full implementation
- changed formatter constructor to be more flexible

(wen)

Fix formatting of converters/unoconv addition (found by pkglint).

(schmonz)

Updated devel/p5-Ouch to 0.0500

(schmonz)

Update to 0.0500. From the changelog:

- Made :trytiny more Try::Tiny-sh thanks to Flavio Poletti

(schmonz)

Updated devel/p5-String-Print to 0.92

(wen)

Update to 0.92

Upstream changes:
version 0.92: Thu  6 Jul 12:31:36 CEST 2017

Fixes:
- %F and %T not supported on Windows [cpantesters]

version 0.91: Tue 27 Jun 17:10:58 CEST 2017

Improvements:
- add output encoding (encodeFor)
- add complex keys
- add modifier BYTES
- add time modifiers
- add default modifier '//'
- add new(missing_key) with warning

(wen)

Register support for PKGSRC_USE_STACK_CHECK

This adds -fstack-check to the CFLAGS (with GCC on NetBSD x86). Only
tested on NetBSD/amd64 so far.

Disabled by default.

(khorben)

Add converters/unoconv

unoconv converts any document format that OpenOffice can import, to any
document format that OpenOffice can export.

unoconv uses the OpenOffice's UNO bindings for non-interactive conversion
of documents.

(manu)

Make path information easily available for upcoming converters/unoconv package

(manu)

Updated net/corebird to 1.5.1

(ryoon)

Update to 1.5.1

Changelog:
- Stability improvements

(ryoon)

+ GraphicsMagick-1.3.26, abcm2ps-8.13.12, afl-2.45b, caff-2.6,
  convmv-2.03, cups-filters-1.14.1, fontconfig-2.12.4,
  libopenmpt-0.2.8461beta26, p5-Algorithm-Permute-0.14,
  p5-Mojolicious-7.35, p5-WWW-Mechanize-1.86, py-easyAI-1.0.0.4,
  py-mercurial-4.2.2, spice-protocol-0.12.13, thunderbird-enigmail-1.9.8.1,
  vim-8.0.0702, vim-share-8.0.0702, x264-devel-20170708, zstd-1.3.0.

(wiz)

Remove two ex-master sites.

(wiz)

Added PYTHON_VERSIONS_INCOMPATIBLE=27

(adam)

PkgSrc clean-ups

(adam)

Remove unused BUILD_DEPDNDS

(adam)

Added devel/py-clickclick version 1.2.1, textproc/py-inflection version 0.3.1

(adam)

Inflection is a string transformation library. It singularizes and pluralizes
English words, and transforms strings from CamelCase to underscored string.
Inflection is a port of Ruby on Rails' inflector to Python.

(adam)

Utility functions (Python 3 only) for the wonderful Click library. Click is
a Python package for creating beautiful command line interfaces in a composable
way with as little code as necessary.

(adam)

Added python to CATEGORIES

(adam)

Add auth-request option (--with-http_auth_request_module).

Approved by joerg@.

(alnsn)

firefox{,45,52}: bump pkgrevision with no change.

these packages pull in GCC_REQD+=4.9 via mozilla-common.mk, and
are very widely used (I suspect only www/firefox actually needs it)

this will take care of most of the fallout from major bumping
pkgsrc-gcc-libstdc++ to 7 on netbsd. these are the most widely
used packages setting GCC_REQD>4.8.

(maya)

Updated net/unbound to 1.6.4

(adam)

Changes 1.6.4:

Features:
* Implemented trust anchor signaling using key tag query.
* unbound-checkconf -o allows query of dnstap config variables. Also unbound-control get_option. Also for dnscrypt.
* unbound.h exports the shm stats structures. They use type long long and no ifdefs, and ub_ before the typenames.
* Implemented opportunistic IPsec support module (ipsecmod).
* Added redirect-bogus.patch to contrib directory.
* Support for the ED25519 algorithm with openssl (from openssl 1.1.1).
* renumbering B-Root's IPv6 address to 2001:500:200::b.
* Fix 1276: [dnscrypt] add XChaCha20-Poly1305 cipher.
* Fix 1277: disable domain ratelimit by setting value to 0.
* Added fastrpz patch to contrib

Bug Fixes:
* Added ECS unit test (from Manu Bretelle).
* ECS documentation fix (from Manu Bretelle).
* Fix 1252: more indentation inconsistencies.
* Fix 1253: unused variable in edns-subnet/addrtree.c:getbit().
* Fix 1254: clarify ratelimit-{for,below}-domain (from Manu Bretelle).
* iana portlist update
* Based on 1257: check parse limit before t increment in sldns RR string parse routine.
* Fix 1258: Windows 10 X64 unbound 1.6.2 service will not start. and fix that 64bit getting installed in C:\Program Files (x86).
* Fix 1259: "--disable-ecdsa" argument overwritten by "ifdef SHA256_DIGEST_LENGTH@daemon/remote.c".
* iana portlist update
* Added test for leak of stub information.
* Fix sldns wire2str printout of RR type CAA tags.
* Fix sldns int16_data parse.
* Fix sldns parse and printout of TSIG RRs.
* sldns SMIMEA and AVC definitions, same as getdns definitions.
* Fix tcp-mss failure printout text.
* Set SO_REUSEADDR on outgoing tcp connections to fix the bind before connect limited tcp connections. With the option tcp connections can share the same source port (for different destinations).
* Add 'c' to getopt() in testbound.
* Adjust servfail by iterator to not store in cache when serve-expired is enabled, to avoid overwriting useful information there.
* Fix queries for nameservers under a stub leaking to the internet.
* document trust-anchor-signaling in example config file.
* updated configure, dependencies and flex output.
* better module memory lookup, fix of unbound-control shm names for module memory printout of statistics.
* Fix type AVC sldns rrdef.
* Some whitespace fixup.
* Fix 1265: contrib/unbound.service contains hardcoded path.
* Fix 1265 to use /bin/kill.
* Fix 1267: Libunbound validator/val_secalgo.c uses obsolete APIs, and compatibility with BoringSSL.
* Fix 1268: SIGSEGV after log_reopen.
* exec_prefix is by default equal to prefix.
* printout localzone for duplicate local-zone warnings.
* Fix assertion for low buffer size and big edns payload when worker overrides udpsize.
* Support for openssl EVP_DigestVerify.
* Fix 1269: inconsistent use of built-in local zones with views.
* Add defaults for new local-zone trees added to views using unbound-control.
* Fix 1273: cachedb.c doesn't compile with -Wextra.
* If MSG_FASTOPEN gives EPIPE fallthrough to try normal tcp write.
* Also use global local-zones when there is a matching view that does not have any local-zone specified.
* Fix fastopen EPIPE fallthrough to perform connect.
* Fix 1274: automatically trim chroot path from dnscrypt key/cert paths (from Manu Bretelle).
* Fix 1275: cached data in cachedb is never used.
* Fix that unbound-control can set val_clean_additional and val_permissive_mode.
* Add dnscrypt XChaCha20 tests.
* Detect chacha for dnscrypt at configure time.
* dnscrypt unit tests with chacha.
* Added domain name based ECS whitelist.
* Fix 1278: Incomplete wildcard proof.
* Fix 1279: Memory leak on reload when python module is enabled.
* Fix 1280: Unbound fails assert when response from authoritative contains malformed qname. When 0x20 caps-for-id is enabled, when assertions are not enabled the malformed qname is handled correctly.
* More fixes in depth for buffer checks in 0x20 qname checks.
* Fix stub zone queries leaking to the internet for harden-referral-path ns checks.
* Fix query for refetch_glue of stub leaking to internet.
* Fix 1301: memory leak in respip and tests.
* Free callback in edns-subnetmod on exit and restart.
* Fix memory leak in sldns_buffer_new_frm_data.
* Fix memory leak in dnscrypt config read.
* Fix dnscrypt chacha cert support ifdefs.
* Fix dnscrypt chacha cert unit test escapes in grep.
* Fix to unlock view in view test.
* Fix warning in pythonmod under clang compiler.
* Fix lintian typo.
* Fix 1316: heap read buffer overflow in parse_edns_options.

(adam)

Updated mail/p5-Mail-DKIM to 0.41

(wen)

Update to 0.41

Upstream changes:
2017-04-14: Marc Bradshaw <marc@marcbradshaw.net>

* commit aac893fdbaa7f8ccd5d37fa7f20d1785406cda51
  Author: Marc Bradshaw <marc@marcbradshaw.net>
  Date:  Fri Mar 17 14:53:53 2017 +1100
    Avoid use of $_ in read loop
    RT 106485: Mail::DKIM::PrivateKey->load tampering $_ and <FILE>

* commit 06934f259e392b2a3cf94560e6051d9e522d0bf3
  Author: Marc Bradshaw <marc@marcbradshaw.net>
  Date:  Fri Mar 17 14:44:44 2017 +1100
    Ensure PrivateKey file is closed properly.
    Store PrivateKey file handle in lexical variable and close it
    once we are done.
    RT 120638: Mail::DKIM::PrivateKey does not close FILE

* commit 9e7c1c4cb78a6cb1cf396ece4379c7ed2c44c974
  Author: Marc Bradshaw <marc@marcbradshaw.net>
  Date:  Fri Feb 27 12:08:11 2015 +1100
    Allow greater control over signed headers

* commit 8291c034dc7db4394e9df80e70b8cbe8428a38c2
  Author: Marc Bradshaw <marc@marcbradshaw.net>
  Date:  Fri Jan 23 09:54:02 2015 +1100
    Allow greater control over which headers are signed by Signer

(wen)

Updated graphics/p5-SVG to 2.78

(wen)

Update to 2.78

Upstream changes:
2.78  2017.07.07 MANWAR
      - Merged PR #10 (fix entity escaping), thanks @haarg. This also resolved RT #121612.

2.77  2017.05.18 MANWAR
      - Proposed fix to the following test failure:
        http://www.cpantesters.org/cpan/report/ddc7eefc-3429-11e7-8430-9cfb106f656a

2.76  2017.05.08 MANWAR
      - Added AUTHOR ONLY test scripts (meta-json.t and meta-yml.t).

2.75  2017.05.07 MANWAR
      - Tidied up unit test scripts and removed '+x' attributes.
      - Tidied up SEE ALSO section in SVG::XML package pod.

2.74  2017.05.04 MANWAR
      - Added key 'provides' to the Makefile.PL script as recommended by CPANTS.

2.73  2017.05.03 MANWAR
      - Merged PR #8, thanks @CSSON. Modified slightly to keep tidyall happy.
      - Correct some spelling errors that were detected by codespell.
      - Don't insert whitespace between tag name and closing bracked when a tag has no attributes

2.72  2017.05.02 MANWAR
      - Added test script t/00-load.t
      - Fixed typo in the pod document.

2.71  2017.05.01 MANWAR
      - Fixed issue RT #57000 with regard to be explici about SVG instance.

2.70  2017.04.27 MANWAR
      - Added MIN_PERL_VERSION to the Makefile.PL script as suggested by CPANTS.

2.69  2017.04.20 MANWAR
      - Fixed issue RT #118091, skip test if Perl::Tidy is missing.

2.68  2017.04.19 MANWAR
      - Fixed issue RT #109521, adding min ver of Test::More v0.94.

2.67  2017.04.18 MANWAR
      - Fixed the issue with regard to github respository info in the Makefile.PL script, thanks @KENTNL.

2.66  2017.04.17 MANWAR
      - Addressed issue RT #121197 (incorrect meta spec), thanks @KENTNL.
      - Added 'clean' and 'dist' keys to the Makefile.PL script.
      - Tidied up .gitignore file.
      - Added MANIFEST file, no need to generate every time.

2.65  2017.04.16 MANWAR
      - Merged PR #9, correcting typo in the pod. Thanks @vti.
      - Tidied up Changes file.
      - Tidied up Makefile.PL script.
      - Tidied up README file.

(wen)

Added net/py-cares version 2.2.0

(adam)

pycares is a Python module which provides an interface to c-ares. c-ares is a
C library that performs DNS requests and name resolves asynchronously.

(adam)

Omit MESSAGE fragment about old versions of XFree86

(maya)

gcc

(maya)

gcc{48,49,5,7}: on netbsd only, bump libstdc++ major to 7.
bump pkgrevision.

while here, synchronize libssp comment to a clearer one.

bump gcc*-lib revision to be one higher.

(maya)

Updated textproc/p5-PDF-API2 to 2.033

(wen)

Update to 2.033

Upstream changes:
2.033    2017-07-06

    - [RT #122371] Remove a couple of improperly-placed weaken statements
      (reported by Phil Perry).

    - [RT #122372] Fix weakening when a page is added to the end of a multiple
      page document (reported by Phil Perry).

    - Fix Bank Gothic core font (reported by Phil Perry).

2.032    2017-07-02

    - PDF::API2 has many circular references, and the end() method doesn't clear
      them all, so memory is leaked.  This release uses Scalar::Util's weaken()
      function to improve garbage collection.  A significant number of circular
      references have been weakened, though many likely still remain.

    - [RT #120756] Eliminate a warning for an ambiguous call to CORE::open
      (first reported by Abdelbaki Brahmia).

    - $text->text_justified() and $text->text_fill_justified() now adjust the
      space between words rather than stretching individual characters in order
      to get the text to fit.

    - [RT #120397] Indirect references and indirect objects can have comments
      embedded in their whitespace, and their object number and generation may
      be split across multiple lines, which may not all be buffered (reported by
      SPROUT).

    - [RT #120450] Fix PDF::API2->open($filename)->stringify() (reported by
      SPROUT).

    - Fix an off-by-one error when calculating text width while charspace is
      non-zero.

    - [RT #120048] Fix PDF::API2->synfont() (broken in 2.029, fixed by Vadim
      Repin) and add basic testing.

(wen)

Updated net/ncftp3 to 3.2.6

(adam)

3.2.6:

If a recursive download operation is also requested with delete mode, attempt to remove empty directories after all files have completed successfully.

No longer trying to utime() after every single block on downloads, which could cause noticable performance degradation when the local filesystem was not local.

Changed behavior of resuming downloads where the timestamp wasn't preserved (because of the utime change, above). The new behavior is to resume the download when the local copy has a recent timestamp (less than a week).

You can now disable use of MFMT like you could similarly disable SITE UTIME (e.g., "-o noMFMT" and "-o noSITE_UTIME").

Now able to use sendfile() for uploads, on Linux/FreeBSD/Mac. Progress reports work too, with a small performance penalty. Ncftpput has a "-s" option to toggle whether it is used (defaults to on in ncftpput and ncftpbatch, off in ncftp).

Ncftpbatch/spooler now use larger buffers for pathnames, allowing for deeper directory trees.

Ncftpbatch/spooler now interpret a received SIGUSR1 as a hint to exit when the current file has finished.

Ncftpbatch/spooler now interpret a received SIGUSR2 to request it to stop sleeping and recheck the queue immediately.

Ncftpbatch/spooler's spool files now allow for you to specify that the local and/or remote file be renamed after a successful transfer.

Ncftpbatch/spooler now a little less chatty by reducing the number of PWD/CWD operations.

Ncftpbatch/spooler now log some xfer stats in its general log file, and ncftpspooler has a new "-x" option to specify a separate xfer log file.

Ncftpbatch/spooler now use a larger default maximum for its log file (10 MiB rather than 200 kB), and ncftpspooler has an -O command line option that can set this limit. Use "-O 0" for no maximum.

Ncftpbatch/spooler now try to present time in local timezone rather than UTC where possible.

Ncftpbatch/spooler now support multiple items per transaction (spool) file.

(adam)

Updated net/lftp to 4.7.7

(adam)

Version 4.7.7:
mirror: fixed coredump when source directory does not exist.
mirror: don't create target directory if can't enter to the source directory.
ftp: fixed a rare hang when a NOOP was sent between "transfer ok" reply and EOF on data socket.
fixed xfer:log setting (compatibility alias).
ftp: don't use EPSV with a proxy.

(adam)

gcc6 pkgrevision

(maya)

gcc6-lib: bump pkgrevision above gcc6

(maya)

gcc6: on netbsd only, bump libstdc++ major to 7 to match base version.

this stops packages using gcc6 as a compiler from linking to two libstdc++s
at once.

(maya)

Remove a bit-rotted inline prototype for a libc function, unbreaking
build on netbsd-8.

(hauke)

Updated textproc/p5-Text-CSV_XS to 1.31

(wen)

Update to 1.31

Upstream changes:
1.31 - 2017-06-13, H.Merijn Brand
    * Fix already decoded BOM in headers
    * New options in csv-check
    * Some perlcritic
    * "escape" is alias for "escape_char" for consistency.
    * Code cleanup and more tests (Devel::Cover)
    * Improve csv-check auto-sep-detection

1.30 - 2017-06-08, H.Merijn Brand
    * Add csv (..., out => ...) syntax examples (issue 7)
    * Disable escape_null for undefined escape_char
    * Fix ->say for bound columns (RT#121576)
    * Update to Devel::PPPort 3.36
    * Tested under 5.26.0 and 5.27.0
    * Doc changes and additions

(wen)

dtpstree: Fix build on NetBSD 8.99.1

Address <sys/user.h> removal.

(kamil)

Make freeimage and freeimageplus use the c++ copiler for linking and remove
linking to stdc++.
This will make them use the right runtime library, not all compilers
use stdc++.

(nros)

Updated textproc/p5-Text-CSV to 1.95

(wen)

Update to 1.95

Upstream changes:
1.95  2017-04-27
    - import "strict" attribute introduced in Text::CSV_XS 1.29

(wen)

Note update of net/bind99 package to 9.9.10pl3.

(taca)

Update bind99 to 9.9.10pl3 (BIND 9.9.10-P3).

--- 9.9.10-P3 released ---

4647. [bug] Change 4643 broke verification of TSIG signed TCP
message sequences where not all the messages contain
TSIG records.  These may be used in AXFR and IXFR
responses. [RT #45509]

(taca)

Note update of net/bind910 package to 9.10.5pl3.

(taca)

Update bind910 to 9.10.5pl3 (BIND 9.10.5-P3).

--- 9.10.5-P3 released ---

4647. [bug] Change 4643 broke verification of TSIG signed TCP
message sequences where not all the messages contain
TSIG records.  These may be used in AXFR and IXFR
responses. [RT #45509]

(taca)

Updated net/mikutter to 3.5.9

(tsutsui)

Update mikutter to 3.5.9.

pkgsrc changes:

- update dependencies per upstream Gemfile

Upstream changes:

- update twitter-text
- support the latest memoist
- fix crash with ruby-gnome2 3.1.1 or prior on opening profile tab
- wrong titles in Twitter Search Model and Web model

(tsutsui)

Updated net/ruby-twitter-text to 1.14.7

(tsutsui)

Update ruby-twitter-text to 1.14.7.

No upstream changelog.
(seems TLDs data updates only)

(tsutsui)

Updated devel/ruby-memoist to 0.16.0

(tsutsui)

Update ruby-memoist to 0.16.0.

No upstream changelog (except github log),
but there is an implicit API change that breaks pkgsrc/net/mikutter.
mikutter will be updated to the latest version which support newer memoist.

(tsutsui)

Disable SSP and FORTIFY - it doesn't seem to play well with stand-alone
tools!

Thanks wiz for identifying the build problem.

(pgoyette)