doc: Updated graphics/librsvg to 2.54.3

(gutteridge)

librsvg: update to 2.54.3

==============
Version 2.54.3
==============

- #866 - Fix detection of gi-docgen (David King, Jan Alexander Steffens).

- Install the generated documentation in the correct place so that
  Devhelp can find it (thanks to Jan Tojnar for pointing this out).

==============
Version 2.54.2
==============

- #823 - Fix regressions when computing element geometries.

- Add a --disable-gtk-doc option for the configure script, so people
  can disable generating documentation for cross-compiling (Matt
  Turner).

- MSVC: Support generating documentation, and passing introspection
  paths (Chun-wei Fan)

==============
Version 2.54.1
==============

This release just has some build-time fixes.

- Fix oversight in the Minimum Supported Rust Version (MSRV) - this
  release requires Rust 1.56 or later.  This has been true since
  librsvg 2.53.1, but it was not checked correctly. (Matt Turner)

- #859 - Make rst2man and gi-docgen optional. (Kleis Auke Wolthuizen)

- #856 - Fix documentation comments.

==============
Version 2.54.0
==============

This is a summary of the 2.53.x release notes; you can also read them
for more detail.  Not many new features; we've been working on
infrastructure and documentation.

- Librsvg now supports SVG2 geometry properties for these elements:
  rect, circle, ellipse, image, svg (Jeremias Ortega).

- #721 - Catch circular references when rendering patterns.

- The C API documentation now uses gi-docgen instead of gtk-doc.
  Rsvg-convert's man page is now converted to reStructuredText instead
  of troff.

- The "Recommendations for applications" chapter in the documentation
  is much improved!

==============
Version 2.53.2
==============

- #851 - Output filled text as text for PDF; fixes regression due to
  outputting all text as paths.

- Fix taller-than-wide proportional scaling and size limiting in
  rsvg-convert.

- #772, #773, #774, #775, #776 - Implement SVG2 geometry properties
  for these elements: rect, circle, ellipse, image, svg (Jeremias
  Ortega).

- #450 - Fix potential unaligned accesses in surface iterators
  (Michael Howell).

- Actually use GDK_PIXBUF_MODULEDIR when calling
  gdk-pixbuf-query-loaders (Fabrice Fontaine).

- Windows build fixes (Chun-wei Fan).

- Add links to functions and types throughout the C API documentation.

==============
Version 2.53.1
==============

While it is not a user-visible change, Jordan Petridis has ported
librsvg's Continuous Integration (CI) infrastructure to use
Freedesktop CI Templates.  This lets us test librsvg easily on
different environments.  If you want your environment tested, now is
the time to add it!

- #599 - Fix incorrect text rendering when text has different scales
  in the X/Y axes.  This regressed after librsvg 2.52.5, when Pango
  had to revert its fix for the same bug.  Now librsvg renders all
  text as paths, and does the scaling itself.  Please file a bug if
  you have evidence that this presents a performance problem for you.

- Update to the latest gtk-rs release (Bilal Elmoussaoui).

- Port to Freedesktop CI Templates (Jordan Petridis).
  of troff.

- The "Recommendations for applications" chapter in the documentation
  is much improved!

==============
Version 2.53.2
==============

- #851 - Output filled text as text for PDF; fixes regression due to
  outputting all text as paths.

- Fix taller-than-wide proportional scaling and size limiting in
  rsvg-convert.

- #772, #773, #774, #775, #776 - Implement SVG2 geometry properties
  for these elements: rect, circle, ellipse, image, svg (Jeremias
  Ortega).

- #450 - Fix potential unaligned accesses in surface iterators
  (Michael Howell).

- Actually use GDK_PIXBUF_MODULEDIR when calling
  gdk-pixbuf-query-loaders (Fabrice Fontaine).

- Windows build fixes (Chun-wei Fan).

- Add links to functions and types throughout the C API documentation.

==============
Version 2.53.1
==============

While it is not a user-visible change, Jordan Petridis has ported
librsvg's Continuous Integration (CI) infrastructure to use
Freedesktop CI Templates.  This lets us test librsvg easily on
different environments.  If you want your environment tested, now is
the time to add it!

- #599 - Fix incorrect text rendering when text has different scales
  in the X/Y axes.  This regressed after librsvg 2.52.5, when Pango
  had to revert its fix for the same bug.  Now librsvg renders all
  text as paths, and does the scaling itself.  Please file a bug if
  you have evidence that this presents a performance problem for you.

- Update to the latest gtk-rs release (Bilal Elmoussaoui).

- Port to Freedesktop CI Templates (Jordan Petridis).

- Visual Studio builds can work from a git checkout now (Chun-wei Fan)

- Windows build fixes (Martin Hertz, Chun-wei Fan)

==============
Version 2.53.0
==============

This is the first release in the new development series.  There are no
new features, just changes to how the documentation is built.

The man page for rsvg-convert is now generated from a reStructuredText
document, and the C API reference is generated using gi-docgen.

Please make sure you install python3-docutils (for rst2man) and
gi-docgen before compiling librsvg from a tarball.

- #755 - rsvg-convert's man page is now generated from reStructuredText.

- #827 - Move C API reference to gi-docgen.

- Test suite updates (Michael Howell).

- Link against bcrypt for Visual Studio builds (Martin Hertz).

(gutteridge)

doc: Updated sysutils/dusage to 0.3.0

(pin)

sysutils/dusage: update to 0.3.0

-Simplify ok, err and highlight.
-Autoclap: Bump to 0.2.2

(pin)

doc: Updated audio/librespot to 0.4.0

(pin)

audio/librespot: update to 0.4.0

0.4.0 - 2022-05-21
Changed
[chore] The MSRV is now 1.53
[contrib] Hardened security of the systemd service units
[core] Session: connect() now returns the long-term credentials
[core] Session: connect() now accepts a flag if the credentails should be
  stored via the cache
[main] Different option descriptions and error messages based on what
  backends are enabled at build time
[playback] More robust dynamic limiter for very wide dynamic range (breaking)
[playback] alsa: improve --device ? output for the Alsa backend
[playback] gstreamer: create own context, set correct states and use sync
  handler
[playback] pipe: create file if it doesn't already exist
[playback] Sink: write() now receives ownership of the packet (breaking)

Added
[main] Enforce reasonable ranges for option values (breaking)
[main] Add the ability to parse environment variables
[main] Log now emits warning when trying to use options that would otherwise
  have no effect
[main] Verbose logging now logs all parsed environment variables and command
  line arguments (credentials are redacted)
[main] Add a -q, --quiet option that changes the logging level to WARN
[main] Add disable-credential-cache flag (breaking)
[main] Add a short name for every flag and option
[playback] pulseaudio: set the PulseAudio name to match librespot's device
  name via PULSE_PROP_application.name environment variable (user set env var
  value takes precedence) (breaking)
[playback] pulseaudio: set icon to audio-x-generic so we get an icon instead
  of a placeholder via PULSE_PROP_application.icon_name environment variable
  (user set env var value takes precedence) (breaking)
[playback] pulseaudio: set values to: PULSE_PROP_application.version,
  PULSE_PROP_application.process.binary, PULSE_PROP_stream.description,
  PULSE_PROP_media.software and PULSE_PROP_media.role environment variables
  (user set env var values take precedence) (breaking)

Fixed
[connect] Don't panic when activating shuffle without previous interaction
[core] Removed unsafe code (breaking)
[main] Fix crash when built with Avahi support but Avahi is locally unavailable
[main] Prevent hang when discovery is disabled and there are no credentials
  or when bad credentials are given
[main] Don't panic when parsing options, instead list valid values and exit
[main] --alsa-mixer-device and --alsa-mixer-index now fallback to the card and
  index specified in --device.
[playback] Adhere to ReplayGain spec when calculating gain normalisation factor
[playback] alsa: make --volume-range overrides apply to Alsa softvol controls

Removed
[playback] alsamixer: previously deprecated options mixer-card, mixer-name and
  mixer-index have been removed

(pin)

doc: Updated audio/termusic to 0.6.16

(pin)

audio/termusic: update to 0.6.16

[V0.6.16] Released on: May 21, 2022.
-New: support loading of m3u,m3u8,xspf,pls,asx playlists.
Only local url supported.
-New: sqlite3 integration. Filter database by different criteria.
Triggered by 2 key.

(pin)

doc: Updated www/py-flask-flatpages to 0.8.1

(kleink)

py-flask-flatpages: Update to 0.8.1.

v0.8.1
------

Release Date: 2021-12-21

Bug Fixes

Small version bump to fix required Python versions

v0.8.0
------

Release Date: 2021-12-21

Prelude

This release includes three PRs. One adds a new attribute to the Page
class, one improves our documentation, and the third improves the
metadata parsing to make it consistent with other 'FlatPage' style
libraries, and less fussy for cases like no page metadata.

New Features {#Release Notes_v0.8.0_New Features}

Added Page.folder as a convencience attribute to the Page class.

Improved metadata parsing, documented above.

Upgrade Notes

This version adds a new method for parsing Metadata. While it passes
new and existing tests, there may still be bugs. As well as
reporting issues, you can use the FLATPAGES_LEGACY_META_PARSER
setting to revert to the old behaviour.
Bug Fixes

Resolves issue
59,
by updating our metadata parser.

(kleink)

librsvg: de-duplicate BUILDLINK_ABI_DEPENDS.librsvg

(gutteridge)

doc: Updated textproc/py-cssselect2 to 0.6.0

(kleink)

py-cssselect2: Update to 0.6.0.

Version 0.6.0
.............

Released on 2022-04-15.

**This version deprecates the ``iter_ancestors`` and
``iter_previous_siblings`` methods, that will be removed in 0.7.0.
Use the ``ancestors`` and ``previous_siblings`` properties instead.**

* Improve speed of ancestors and previous siblings

Version 0.5.0
.............

Released on 2022-02-27.

* Support Python 3.10
* Drop support of Python 3.6
* Handle many CSS4 selectors
* Ignore comments at the beginning of selectors

(kleink)

net/php-yaz: dose not support PHP 8 and later

This package dose not support php80 and php81.

(taca)

net/php-xmlrpc: is not supported PHP 8 and later

php80 and php81 droped xmlrpc extension.

(taca)

net/php-geoip: dose not support PHP 8 and later

This package dose not support php80 and php81.

(taca)

math/php-stats: dose not support PHP 8 and later

This package dose not support php80 and php81.

(taca)

misc/php-orangehrm: fix build problem

* Remove "*.orig" files before install.
* Replace interpreter of bash scripts.
* Do not check interpreter for Symfony.

(taca)

print/php-pdflib: dose not support PHP 8 and later

This package dose not support php80 and php81.

(taca)

doc: Updated math/R-mvtnorm to 1.1.3

(mef)

(math/R-mvtnorm) Updated 1.1.2 to 1.1.3 (test passes)

(from: inst/NEWS)
Changes in version 1.1-3 (2021-10-05)

  o  fix diffs reported on M1mac

(mef)

doc: Updated math/R-dplyr to 1.0.9

(mef)

(math/R-dplyr)  Updated 1.0.7 to 1.0.9

# dplyr 1.0.9

* New `rows_append()` which works like `rows_insert()` but ignores keys and
  allows you to insert arbitrary rows with a guarantee that the type of `x`
  won't change (#6249, thanks to @krlmlr for the implementation and @mgirlich
  for the idea).

* The `rows_*()` functions no longer require that the key values in `x` uniquely
  identify each row. Additionally, `rows_insert()` and `rows_delete()` no
  longer require that the key values in `y` uniquely identify each row. Relaxing
  this restriction should make these functions more practically useful for
  data frames, and alternative backends can enforce this in other ways as needed
  (i.e. through primary keys) (#5553).

* `rows_insert()` gained a new `conflict` argument allowing you greater control
  over rows in `y` with keys that conflict with keys in `x`. A conflict arises
  if a key in `y` already exists in `x`. By default, a conflict results in an
  error, but you can now also `"ignore"` these `y` rows. This is very similar to
  the `ON CONFLICT DO NOTHING` command from SQL (#5588, with helpful additions
  from @mgirlich and @krlmlr).

* `rows_update()`, `rows_patch()`, and `rows_delete()` gained a new `unmatched`
  argument allowing you greater control over rows in `y` with keys that are
  unmatched by the keys in `x`. By default, an unmatched key results in an
  error, but you can now also `"ignore"` these `y` rows (#5984, #5699).

* `rows_delete()` no longer requires that the columns of `y` be a strict subset
  of `x`. Only the columns specified through `by` will be utilized from `y`,
  all others will be dropped with a message.

* The `rows_*()` functions now always retain the column types of `x`. This
  behavior was documented, but previously wasn't being applied correctly
  (#6240).

* The `rows_*()` functions now fail elegantly if `y` is a zero column data frame
  and `by` isn't specified (#6179).

# dplyr 1.0.8

* Better display of error messages thanks to rlang 1.0.0.

* `mutate(.keep = "none")` is no longer identical to `transmute()`.
  `transmute()` has not been changed, and completely ignores the column ordering
  of the existing data, instead relying on the ordering of expressions
  supplied through `...`. `mutate(.keep = "none")` has been changed to ensure
  that pre-existing columns are never moved, which aligns more closely with the
  other `.keep` options (#6086).

* `filter()` forbids matrix results (#5973) and warns about data frame
  results, especially data frames created from `across()` with a hint
  to use `if_any()` or `if_all()`.

* `slice()` helpers (`slice_head()`, `slice_tail()`, `slice_min()`, `slice_max()`)
  now accept negative values for `n` and `prop` (#5961).

* `slice()` now indicates which group produces an error (#5931).

* `cur_data()` and `cur_data_all()` don't simplify list columns in rowwise data frames (#5901).

* dplyr now uses `rlang::check_installed()` to prompt you whether to install
  required packages that are missing.

* `storms` data updated to 2020 (@steveharoz, #5899).

* `coalesce()` accepts 1-D arrays (#5557).

* The deprecated `trunc_mat()` is no longer reexported from dplyr (#6141).

(mef)

Updated devel/py-llvmlite, security/py-gnupg

(adam)

py-gnupg: updated to 0.4.9

0.4.9

Added a status attribute to the returned object from gen_key() which is set to 'ok' if a key was successfully created, or 'key not created' if that was reported by gpg, or None in any other case.
Provided the ability to add subkeys. Thanks to Daniel Kilimnik for the feature request and patch.
Added keygrip values to the information collected when keys are listed. Thanks to Daniel Kilimnik for the feature request and patch.
Added extra_args to send_keys(), recv_keys() and search_keys() to allow passing options relating to key servers.

(adam)

py-llvmlite: updated to 0.38.1

v0.38.1
This is a maintenance release to support the Apple M1 architecture.

(adam)

doc/TODO: add some

+ ImageMagick-7.1.0.35, Mesa-22.1.0, gopls-0.8.4, grpc-1.46.3,
  harfbuzz-4.3, libxkbcommon-1.4.1, ninja-build-1.11.0, py-dill-0.3.5.1,
  py-feedparser-6.0.10, py-pip-audit-2.3.0, rust-analyzer-0.0.20220517
  [needs rust 1.60.0].

(wiz)

libcyaml: Update MAINTAINER

(charlotte)

error: Autoconf version 2.71 or higher is required for libcups-2.4.1

(rxg)

doc: Updated math/R-XML to 3.99.0.9

(mef)

(math/R-XML) Updated 3.99.0.8 to 3.99.0.9, ChangeLog not updated

(mef)

doc: Updated math/R-car to 3.0.13

(mef)

(math/R-car) Updated 3.0.11 to 3.0.13

Changes to Version 3.0-13

  o Handle singularities better in Anova.mlm() (suggestion of Marius Barth).

  o Fixed error reporting in logit() (bug reported by Collin Erickson).

  o Added pt.cex and cex sub-arguments to legend argument of
  scatterplotMatrix() (implicit suggestion of Duncan Mackay).

  o New vif.lm() method with options for handling linear models with interactions.

  o Fixed bug in cooks.distance.influence.lme() (in effect, incorrect
  division by the error variance twice rather than once), and a sign
  error in dfbeta.influence.lme() and dfbetas.influence.lme()
  (following report by Ben Bolker).

  o Added an error.df argument to the default methods of
  linearHypothesis() and Anova() for computing F-tests; if the
  argument is unspecified, then df.residual() is applied to the model,
  and if df.residual() returns NULL or NA, then an asymptotic
  chi-square test is substituted for the F-test (in response to a
  report by Adrian Olszewski that Anova() fails for gls models when an
  F-test is requested).

Changes to Version 3.0-12

  o The 'rio` package is now suggested, not required.  To use the
  Import() and Export() functions users may need to install rio using
  install.packages("rio")

  o hccm() now reports an interpretable error, except for type = "hc0"
  or "hc1", when there are cases with hatvalues = 1 (following problem
  reported by Peng Ding).

  o Fix bug in rownames in dfbetasPlots.lm() (reported by Shu Fai
  Cheun).

  o Package 'rio' was moved from Imports to Suggests. Both Import()
  and Export() now check for the availability of 'rio' first and
  report an error otherwise.

(mef)

mediawiki: fix pkglint warnings about redundant APACHE_USER

(rillig)

mediawiki: remove noop SUBST blocks

These blocks have been ignored since 2018-07-04, when the SUBST_STAGE
was changed from post-patch to pre-configure.  Replacing
"NO_CONFIGURE=yes" with "do-configure: # nothing" demonstrated that
these SUBST blocks were outdated anyway, at least one of them referred
to missing files.

(rillig)

php-gd: remove unknown configure option --with-xpm-dir

Add license, bump PKGREVISION.

(rillig)

elm-me: fix pkglint warnings

(rillig)

doc: Updated misc/most to 5.1.0

(fcambus)

most: update to 5.1.0.

Changes since 5.0.0
1. src/Makefile.in: Install docs in $prefix/share/doc
2. src/most.h,sysdep.c: win32-specific patches (Andrew Wu)
3. most.1: Typo corrected (Stephane Blondon)
4. src/line.c: removed redundant switch in the most_forward_columns
  function.
5. src/buffer.c: avoid call to SLutf8_skip_char if the character is
  ascii.
6. src/file.c: Added an option to most_read_file_dsc to skip the line
  count and made use of it in search.c to speed up the searches a bit.
7. src/display.c: When wrapping a line, the color was not "wrapping".
8. src/line.c: When moving forward n columns, skip over trailing 0
  length color escape sequences.
9. src/buffer.c: If a color escape sequence occurs at the end of a
  line line whose display width is equal to the terminal width, do
  not wrap the line.
10. src/keym.c: Limit the amount of loops when going to the end of the
    buffer.  It is possible for a run-away process to write to a
    logfile that most is viewing causing it to loop forever seeking the
    end of the file, which may never come.
11. src/help.c: Change DELETE to BACKSPACE.  file.c: Indicate that
    chdir return value is being ignored.
12. configure,autoconf/*: Updated configure script and related files.
13. src/most.c: when acting as `cat`, a file pointer was not being
    closed (George Danchev).
14. src/*.c: Removed some unused variables.
15. configure: Use /etc/ld.so.conf to located system lib paths.
16. autoconf/*: updates
17. src/line.c, buffer.c: Fixed an issue with the display of
    double-width characters to address
    <https://bugs.launchpad.net/ubuntu/+source/most/+bug/312454>.
18. src/*.[ch]: Updated copyright year
19. src/search.c: Rewrote the search functions from scratch.  Both
    ordinary and regexp searches are now supported.  Use -r to startup
    with regular-expression searches.
20. src/keym.c: Added support for Home/End keys (top/bottom of buffer).
21. src/line.c: Added a work-around for programs that try colorize the
    output using the clear-to-end-of-line escape sequence (ESC[K)
    without regard for the value of isatty(fileno(sdout)).  Most will
    ignore ESC[K unless invoked with -v.
22. configure: Updated to use aclocal.m4 v0.3.3-1
23. src/Makefile.in: Added LDFLAGS to chkslang linker line (Denis Pronin)
24. *: Converted the man page source most.1 to text-macro, and
    regenerated it.  Updated copyrights for upcoming release.
25. Updated manifest and copyrights updated for 2019 release.

(fcambus)

doc: Updated sysutils/xplr to 0.18.0

(pin)

Asysutils/xplr: update to 0.18.0

Changes & fixes

    Added new command-line arguments --read0, --write0 and -0/--null to
    support reading and printing null character delimited file paths.
    Added support for filtering with regex using the new filters:
        RelativePathDoesMatchRegex
        RelativePathDoesNotMatchRegex
        IRelativePathDoesMatchRegex
        IRelativePathDoesNotMatchRegex
        AbsolutePathDoesMatchRegex
        AbsolutePathDoesNotMatchRegex
        IAbsolutePathDoesMatchRegex
        IAbsolutePathDoesNotMatchRegex
    Key binding f r and f R will now filter using regex.
    Added key binding f backspace to remove the last filter.
    Search mode now defaults to regex search.
    Added a new message SetInputPrompt to set the input prompt dynamically.
    Added new node properties (available via the Lua API):
        created
        last_modified
    Added support for sorting with timestamp using the new sorters:
        ByCreated
        ByLastModified
        ByCanonicalCreated
        ByCanonicalLastModified
        BySymlinkCreated
        BySymlinkLastModified
    The last column in the files table now displays the last modification time.
    Fixed a bug causing crashes when handling unicode characters in the input
    buffer.
    Improved docs.

(pin)

doc: Updated textproc/tuc to 0.7.0

(pin)

textproc/tuc: update to 0.7.0

[0.7.0] - 2022-05-21
    Add support for --join
    Add support for --lines
    Add support for --complement
    Add support for --zero-terminated

(pin)

security/php-ssdeep: restrict php56 and php74

* Resrict this package for php56 and php74.
* Update HOMEPAGE.

(Not updated more than 4 years.)

(taca)

doc: Updated lang/php81 to 8.1.6nb1

(taca)

lang/php80: forgot to add a file

Oops, forgot to add a patch file.

(taca)

php81: allow copying files with size zero

Allow copying files with size zero.  This behavior differ from PHP 7.4
and break a pear package which contains files with size zero.

Bump PKGREVISION.

XXX: pullup-2022Q1

(taca)

doc: Updated lang/php80 to 8.0.19nb1

(taca)

php80: allow copying files with size zero

Allow copying files with size zero.  This behavior differ from PHP 7.4
and break a pear package which contains files with size zero.

Bump PKGREVISION.

XXX: pullup-2022Q1

(taca)

doc/pkgsrc.*: regen

(nia)

guide: remove __unix, pointed out by wiz, it's actually __unix__, but not
defined by macOS

(nia)

doc: Updated graphics/R-effects to 4.2.1

(mef)

(graphics/R-effects) Updated 4.1.1 to 4.2.1

Version 4.2-1

o Allow the vcov. argument in the call to one of the effect methods
to be a matrix or a function.  Previously only a function was
permitted.

o Fixed bug in plot.eff() for multiline plots in which the x.var
argument is set explicitly (reported by Gerrit Eichner).

o Small improvements to docs.

Version 4.2-0

o Adding the use of the insight package to get formulas.

o Added Effect.glmmPQL() for fitting generalized linear mixed models
using the glmmPQL() function in the MASS package for penalized
quasi-likelihood estimation.

o Don't ignore lines lty sub-argument to plot.effpoly() (fixing a bug
reported by Laura Adamovicz).

o One-column matrix predictors trigger a warning and correction
(following problem reported by Carla Hendricks).

o Added regression tests for classes of model objects
supported. Tests must be enabled in the effects sources at
tests/regression-tests.R.

o Rewrite of defining effect method vignette.

o Minor changes to predictor effects gallery vignette.

Version 4.1-4

o Change class(x) == "y" to inherits(x, "y") in code.

Version 4.1-3

o Handle xlevels=n argument correctly in Effect() (fixing bug
reported by Gerrit Eichner).

o Add col argument to effectsTheme(), mostly to accommodate
colorblind-friendly palette (suggestion of Sercan Kahveci).

Version 4.1-2

o Effect methods for clm, clm2, and clmm models were not correctly
passing the estimated theshholds to polr for computing effects (error
reported by Stefan Th. Gries).

o Updated "Defining Effects Methods ..." vignette to reflect
revisions to Effect.clm, Effect.clm2 and Effect.clmm.

o Make sure that Effect() and plot() methods work correctly with
character and logical predictors (following bug report by Julian
Gaviria Lopez).

(mef)

doc: Updated math/R-insight to 0.17.1

(mef)

(math/R-insight) Updated 0.14.4 to 0.17.1

# Insight 0.17.1

## New supported model classes

* `deltaMethod` (*car*), `marginaleffects`, `marginaleffects.summary`
  (*marginaleffects*)

## General

* `get_predicted()` now supports models of class `iv_robust` and `ivreg`.

* For `get_predicted()`, when both `type` and `predict` are given, `type`
  will overwrite `predict`. Note that this will print a message, because
  `predict` is the preferred argument.

* `get_varcov()` gains `vcov` and `vcov_args` arguments, to specify the
  variance-covariance matrix used to compute uncertainty estimates (e.g., for
  robust standard errors).

* `get_loglikehood()` improved handling of models from package *estimator*.

## Bug fixes

* Fixed bug in `get_data()` for model objects whose data needs to be recovered
  from the environment, and where the data name was a reserved word (e.g., named
  like an R function).

* The matrix returned by `get_varcov()` for models of class *bife* now returns
  row and column names.

* `find_offset()` did not find offset-terms for `merMod` objects when the
  offset was specified as `offset` argument in the function call.

# insight 0.17.0

## Breaking changes

* Arguments `vcov_estimation` and `vcov_type` in `get_predicted()`,
  `get_predicted_se()` and `get_predicted_ci()` are replaced by `vcov` and
  `vcov_args`, to have a more simplified and common interface to control
  robust covariance matrix estimation.

## General

* Improved performance for various functions, in particular `get_data()` and
  `model_info()`.

## New functions

* To check for names: `object_has_names()` and `object_has_rownames()`

* To work with lists: `is_empty_object()` and `compact_list()`

* To work with strings: `compact_character()`

* Further utility functions are `safe_deparse()`, `trim_ws()` and `n_unique()`.

## Changes to functions

* `export_table()` now better checks for invalid values of caption and footer
  for tables in HTML format, and silently removes, e.g., ansi-colour codes that
  only work for text-format.

* `get_data.coxph()` returns the original data frame instead of data with type
  coercion.

* `get_loglikelihood()` gets a `check_response` argument, to check if a model
  has a transformed response variable (like `log()` or `sqrt()` transformation),
  and if so, returns a corrected log-likelihood.

* `get_modelmatrix()` now supports *BayesFactor* models.

* `get_loglikelihood()` and `get_df()` now support more model classes.

* `get_predicted()` was improved for multinomial models from *brms*.

* `get_variance()` was improved to cover more edge cases of (more complex)
  random effect structures.

* `get_data()` now includes variables in the returned data frame that were
  used in the `subset` argument of regression functions (like `lm()`).

* In some edge cases, where `get_data()` is unable to retrieve the data that
  was used to fit the model, now a more informative error is printed.

* `ellipses_info()` now also accepts a list of model objects, is more stable
  and returns more information about the provided models (like if all fixed
  or random effects are the same across models, if all models are mixed models
  or null-models, etc.)

* `check_if_installed()` now works interactively and lets the user prompt
  whether to automatically update or install packages.

## Bug fixes

* Fixed incorrect column name conversion in `standardize_names()` for certain
  columns returned by `broom::glance()`.

* Fixed issue with correctly detecting Tweedie-models in `model_info()`.

* Fixed issue with `get_datagrid()` for *brms* models with monotonic factors.

* Fixed issue in `find_formula()` when argument `correlation` was defined
  outside of `lme()` and `gls()` (@etiennebacher, #525).

* Fixed issue with `get_data()` when back-transforming data from predictors
  that used `cos()`, `sin()` or `tan()` transformations.

# insight 0.16.0

## New functions

* `get_datagrid()`, to generate a reference grid, usually used when computing
  adjusted predictions or marginal means from regression models.

## Changes to functions

### `get_predicted()`

* `get_predicted()` was revised. Beside the four core options for the `predict`
  argument, it is now also possible to use any value that is valid for the
  model's `predict()` method's `type` argument.

* `get_predicted()` now supports more models (e.g., from packages like
  _GLMMadaptive_ or _survival_).

* `get_predicted()` is now more robust when calculating standard errors of
  predictions.

### Other functions

* `get_statistic()` and `find_statistic()` now support *htest* objects.

## General

* Various minor improvements.

# insight 0.15.1

## General

* Improved speed performance, especially for `get_data()`.

## Changes to functions

* `get_data()` for `coxph` models now returns the original factor levels for
  variables transformed with `strata()` inside formulas.

# insight 0.15.0

## Breaking changes

* Data management functions (like `reshape_longer()`, or `data_match()`) have
  been moved to the *datawizard* package.

* `get_data()` no longer returns factor types for numeric variables that have
  been converted to factors on-the-fly within formulas (like `y ~ as.factor(x)`).
  Instead, for each numeric variable that was coerced to factor within a formula
  gets a `factor` attribute (set to `TRUE`), and the returned data frame gets
  a `factors` attribute including all names of affected variables.

## New supported model classes

* Support for `bfsl` (*bfsl*)

## New functions

* New `standardize_column_order()` function can be used to standardize the
  column order in output dataframes.

## General

* Improved speed performance for some functions.

* Improved handling of table captions and footers in `export_table()`. See also
  the new vignette on exporting data frames into human readable tables here:
  https://easystats.github.io/insight/articles/export.html

* Revised `width` argument in `export_table()`, which now allows to set
  different column widths across table columns. See examples in
  `?export_table`.

* `export_table()` gets a `table_width` argument to split wide tables into
  two parts.

* `get_varcov()` for `MixMod` (package *GLMMadaptive*) was revised, and now
  allows to return a robust variance-covariance matrix.

* Added more `get_df()` methods.

## Bug fixes

* Fixed issues with manual sigma computation to handle dispersion models in
  `get_sigma()`.

* Fixed issue in `find_formula()` for `BayesFactor::lmBF()` with multiple random
  effects.

* Fixed issue in `get_parameters.BFBayesFactor()` with wrong sign of difference
  estimate for t-tests.

* Argument `width` in `format_value()` was ignored when formatting integer
  values and `protect_integers` was set to `TRUE`.

# insight 0.14.5

## New functions

* `find_transformation()` and `get_transformation()` to find or get any function
  that was used to transform the response variable in a regression model.

## General

* Improved support for models of class `sampleSelection`.

* Improved documentation.

* `get_modelmatrix()` now supports: `rms::lrm`

* `get_predicted()` supports: `MASS::polr`, `MASS::rlm`, `rms::lrm`, `fixest`,
  `bife::bife`, `ordinal::clm`.

* `get_predicted()` standard errors are often much faster to compute.

* `get_predicted()` supports models with "grouped" or "level" outcomes (e.g.,
  multinomial logit).

* `get_predicted()` handles factors better.

* Improved documentation

## Changes to functions

* `check_if_installed()` gains a `quietly` argument, if neither stopping nor a
  warning message for non-installed packages is requested.

* `get_predicted()`'s `predict` argument now accepts these values: "link",
  "expectation", "prediction", "classification", or NULL.

* `get_predicted()` accepts `predict=NULL`, which allows users to push a `type`
  argument through the `...` ellipsis, forward to the `predict()` method of the
  modelling package.

## Bug fixes

* Fixed issue with parameter names from *emmeans* objects in
  `get_parameters()`.

* Fixed issues with unknown arguments in `get_predicted()`.

(mef)

doc: Updated math/R-robustbase to 0.95.0

(mef)

(math/R-robustness) Updated 0.93.8 to 0.95.0

(inst/NEWS.Rd or https://cran.r-project.org/web/packages/robustbase/news.html)

NEWS            R Documentation

News for R Package robustbase

CHANGES in robustbase VERSION 0.95-0 (2022-04-02, svn r894)

BUG FIXES

  * huberM(x, k) now also works in small sample cases with +/-Inf, e.g. for x =
    c(-Inf, 1).

NEW FEATURES

  * New huberize() utility, notably by default used in mc().

  * mc() now by default uses doScale=FALSE in order to guarantee convergence
    also in extreme cases when used together with new c.huberize = 1e11. Both
    are not back compatible but clearly better in extreme cases, as assessed by
    Lukas Graz' BSc thesis. Note that the behavior of the new default doScale=
    FALSE has been amended in the C code to be more compatible in its eps
    checks with (the previous default) doScale=TRUE.

    Lastly, a message() is printed currently, at most once per R session,
    whenever mc() is called without an explicit specification of doScale=*,
    alerting users to the changed default. This message can be turned off by
    setting options(mc_doScale_quiet = TRUE).

  * New dataset data(x30o50) as an example of potentially problematic data, not
    just for mc().

Misc

  * rather "truncating" +/-Inf to <large>, use NA.OK = TRUE in Qn() and Sn()'s
    .C() call.

  * Our C code no longer uses the DOUBLE_* constants from S, but rather
    standard <float.h>.

CHANGES in robustbase VERSION 0.93-9 (2021-09-27, svn r888)

BUG FIXES

  * Qn(x) and Sn(x) gain an na.rm = FALSE option and now work when x has NAs
    and contains +/- Inf.

Misc

  * No longer include 'Rdefines.h' as it is somewhat deprecated.

  * In 'src/lmrob.c' add USE_FC_LEN_T and use FCONE rather than the older FCLEN
    approach.

(mef)

py-PyRTL: Add upstream pull request URL to a patch

(ryoon)

doc: Updated math/R-psych to 2.2.5

(mef)

(math/R-psych) Updated 2.1.6 to 2.2.5, explict NEWS.md unknown

(mef)

doc: Updated www/firefox-l10n to 100.0.2

(ryoon)

firefox-l10n: Update to 100.0.2

* Sync with www/firefox-100.0.2.

(ryoon)

doc: Updated www/firefox to 100.0.2

(ryoon)

firefox: Update to 100.0.2

Changelog:
Security fixes:
#CVE-2022-1802: Prototype pollution in Top-Level Await implementation
#CVE-2022-1529: Untrusted input used in JavaScript object indexing, leading to
prototype pollution

(ryoon)

doc: Updated math/R-bbmle to 1.0.25

(mef)

(math/R-bbml) Updated again to 1.0.25, sorry

(mef)

Added math/R-bdsmatrix version 1.3.4

(mef)

math/R-bdsmatrix: import R-bdsmatrix-1.3.4

This is a special case of sparse matrices, used by coxme.

(mef)

(math/R-bbml) Revert to 1.0.20, dependency not available yet

(mef)

(math/R-bbml) Revert to 1.0.20, dependency not available yet

(mef)

doc: Updated math/R-bbmle to 1.0.25

(mef)

(math/R-bbml) Updated 1.0.20 to 1.0.25, explicit NEWS.md unknown

(mef)

doc: Updated math/R-VGAM to 1.1.6

(mef)

(math/R-VGAM) Updated 1.1.1 to 1.1.6

                CHANGES IN VGAM VERSION 1.1-6

NEW FEATURES

    o  Most of the GAIT regression code have new arguments 'd.mlm'
        and 'd.mix' but mainly only the dpqr-type functions work!
        Some other functions should work such as gaitdpoisson() and
        dgaitplot(). Also, is.deflated() and goffset() are new.
    o  New family function: gaitdnbinomial().
    o  Upgraded family functions: gaitpoisson() becomes gaitdpoisson(),
        gaitlog() becomes gaitdlog(), gaitzeta() becomes gaitdzeta().
    o  New generic and/or methods functions: KLD(), Pheapseep().
    o  Some colours are pre-defined for GAITD regression, and the default
        colour scheme has changed.
    o  trim.constraints() has a new 'intercepts = TRUE' argument.
    o  Many family functions have a 'dpqrfun' component in the @infos slot.
        Useful to tie in the modelling function with its density, etc.
    o  binormal() has a new 'rho.arg' argument.
    o  Some link functions handle deriv = 4: e.g., loglink(), logitlink().
    o  vglmfit@misc$history gives the convergence history.
    o  Improvements to clogloglink() at the boundary 1.
    o  hdeff.vglm() has a new argument 'theta0'.
    o  wald.stat() cleaned up internally.
    o  Tested okay on R 4.1.2.

BUG FIXES and CHANGES

    o  All GAIT regression code has arguments 'alt.mlm' changed to
        'a.mlm', 'inf.mix' changed to 'i.mix', etc. Colours have changed.
        [dpqr]gaitpois() have been renamed to [dpqr]gaitdpois(),
        and gaitpoisson() to gaitdpoisson().
    o  [dpqr]zanegbin() no longer supports argument 'prob'.
    o  hdeff() makes a recursive call if any NAs are in the answer and
        fd.only == FALSE. So the answer should then be a 'proper' one.
    o  Code involving anova() for VGLMs/VGAMs changed to avoid an
        error when loading in \pkg{semTools}.
    o  spikeplot(..., use.table = TRUE) failed wrt x.use and y.use.
    o  vcov() slightly faster for "vglm" objects.  Thanks to Frank Harrell
        for comments that led to this.
    o  linkfun.vglm() changed to linkfunvglm().

                CHANGES IN VGAM VERSION 1.1-5

NEW FEATURES

    o  New functions: meangait(), spikeplot(), trim.constraints().
    o  iam(i, 1:M, M = M) now works, for selecting a whole row,
        and similarly for a column.
    o  Tested okay on R 4.0.3.

BUG FIXES and CHANGES

    o  gait[log,poisson,zeta](): arguments llambda.[ai] and lshape.[ai]
        equal llambda.p and lshape.p by default. This means only one
        argument needs be assigned in order to change them all.
    o  [dpqr]genpois[12]() has argument 'mean' changed to 'meanpar'.
    o  gait[log,zeta]() bug fixed wrt missing 'allprobs'.
    o  qfoldnorm() now uses the bisection method. Thanks to
        Mauricio Romero for a suggestion that led to this.
    o  gait[log,poisson,zeta]() compute the sandwich formula
        explicitly for the working weights; this has speed gains.
        Their @deriv avoids nested for() loops too.
        Their initial value grids have become coarser.
    o  Functions moved to VGAMdata 1.1-5: yip88, bellff(dr), tikuv(dpqr).

                CHANGES IN VGAM VERSION 1.1-4

NEW FEATURES

    o  Functions score.stat() and wald.stat() offer greater flexibility
        by some new arguments, and 'as.summary' changed to 'orig.SE'.
    o  Influence() computes the influence functions of a "vglm" object.
    o  Methods functions hdeff.matrix() and hdeff.numeric() are new.
    o  Family function slot genpoisson()@ weight was updated with
        the EIM derived by Easton Huch, and after major changes,
        genpoisson() was renamed to genpoisson0().
    o  New functions: [dpqr]gaitnbinom(), [dpqr]gaitgenpois[012](),
        [dpqr]gaitbinom(), dgaitplot(), plotdgait.vglm(), Trunc().
    o  New family functions: genpoisson[12](dpqr), gaitpoisson()
        so that gaitpoisson.mix() and gaitpoisson.mlm() were removed,
        gaitlog() so that gaitlog.mix() was removed,
        gaitzeta() so that gaitzeta.mix() was removed.
    o  Tested okay on R 4.0.3.

BUG FIXES and CHANGES

    o  Deprecated: dgenpois(), genpoisson(), [dpqr]gaitnbinom.m[ix,lm]().
        gatnbinomial.m[ix,lm](dpqr), [dpqr]gaitbinom.mlm(),
        gaitpoisson.mix(), gaitpoisson.mlm(), gabinomial.mlm(),
        gtbinomial().
    o  'lhs.prob' changed to 'cdf.max.s'.
    o  Argument changes for all GAIT code, e.g., 'alter' changed to
        'alt.mix' or 'alt.mlm', 'byrow.arg' to 'byrow.ai', etc.
    o  Functions moved to VGAMdata 1.1-4: oalog, oilog, otlog,
        oapospois, oipospois, otpospois, oazeta, oizeta, otzeta, oizipf,
        oiposbinomial(dpqr), [dpqr]posbinom, [dpqr]pospois, [dpqr]posnegbin.
    o  Pearson residuals no longer printed by default in a summary
        of a "vglm" object: presid = FALSE.
    o  Some changes made to plot(vgam()) code for changes in R 4.0.0,
        especially if the covariate is a character/factor.
        Now avoids "coercing argument of type 'double' to logical" warning.
    o  negbinomial(lmu = "nbcanlink") had an error relating to "newemu".

                CHANGES IN VGAM VERSION 1.1-3

NEW FEATURES

    o  New functions: [dpqr]gaitpois(), [dpqr]gaitlog(), [dpqr]gaitzeta(),
        dextlogF().
    o  New family functions: gatnbinomial.mix(dpqr),
        gaitlog.mix(dpqr), gaitpoisson.mix(dpqr), gaitpoisson.mlm(dpqr),
        extlogF1(d).
    o  New generic and/or methods functions: altered(),
        inflated(), truncated(), specialsvglm(),
        fix.crossing(), is.crossing().
    o  Coef.vlm() handles models with contiguous "multilogitlink"s,
        e.g., gaitpoisson.mlm().
    o  get.offset() is now defined as a generic function (returns the
        n x M matrix of offsets by default).
    o  Some .f files have been pruned of unnecessary code.
    o  New data set: V2.
    o  Tested okay on R 4.0.0.

BUG FIXES and CHANGES

    o  Another attempt to .Deprecated()  all link function names
        which do not end in "link", e.g., loglink() replaces loge(),
        logitlink() replaces logit().
    o  gatpoisson.mix() (now merged into gaitpoisson.mix())
        reorders its linear/additive predictors and now handles
        length(alter) == 0 (default, so equivalent to poissonff())
        and length(alter) == 1.
        Ditto for gatnbinomial.mlm(), etc.
    o  Three argument mismatches found in .f and .c code.
    o  qlog() and qyules() buggy wrt interval selection prior to
        applying bisection.
    o  'Rank mismatches' flagged by gfortran10 fixed, by not
        calling two subroutines.
    o  [dpqr]gaitpois.mlm() and [dpqr]gaitpois.mix() have been combined into
        [dpqr]gaitpois().

                CHANGES IN VGAM VERSION 1.1-2

NEW FEATURES

    o  New generic and/or methods functions: add1.vglm(),
        dfterms(), drop1.vglm(), extractAIC.vglm(),
        rootogram4vglm() [based on \pkg{countreg}],
        step4vglm().
    o  New family functions: gabinomial.mlm(), gatnbinomial.mlm(dpqr),
        gatpoisson.mix(dpqr), gatpoisson.mlm(dpqr).
    o  New functions: [dpqr]zipfmb(),
        [dpqr]gaitbinom.mlm(), [dpqr]gaitnbinom.mlm(),
        [dpqr]gaitpois.mlm(),  [dpqr]gaitnbinom.mix(),
        [dpqr]gaitpois.mix().
        Note that the naming conventions have changed somewhat, especially
        regarding the negative binomial distribution.
    o  Decommissioned functions:
        gapoisson(), gtpoisson(),
        [dpqr]gabinom(), [dpqr]gapois(),
        [dpqr]gibinom(), [dpqr]gipois(),
        [dpqr]gtbinom(), [dpqr]gtpois().
    o  eta2theta() should be able to handle links for, e.g.,
        gatpoisson.mlm(), gibinomial().
    o  posbernoulli.t() has a 'type.fitted' argument.
    o  gtpoisson.mlm() supports type.fitted = "Pobs.a",
        gipoisson.mlm() supports type.fitted = "Pstr.i".
    o  New data set: backPain2.
    o  Cleve Moler added to DESCRIPTION as a contributor, to reflect
        the use of LINPACK.
    o  Tested okay on R 3.6.1.

BUG FIXES and CHANGES

    o  Use of old link function names now result in a call to
        .Deprecated(), except for a selected few such as logit() and
        probitlink()---this will only be allowed for a limited time.
    o  negbinomial()@initialize bug fixed for "nbcanlink()" and
        multiple responses; thanks to Victor Miranda.
    o  cens.poisson() gave a warning upon a summary().
        Thanks to Jens Heumann for picking this up.
    o  bisection.basic() returns a warning, not an error, if
        it cannot find an interval which covers the root.
    o  Rank() has an improved implementation and documentation.
    o  anova.vglm() had some bugs---thanks to Thamron Keowmani for
        finding them.
    o  posnormal() failed when 'constraints' was inputted---thanks
        to Michael Zamo for picking this up.
    o  Improvements to yules(dq).

(mef)

doc: Updated security/R-openssl to 2.0.1

(mef)

(security/R-openssl) Updated 2.0.0 to 2.0.1

2.0.1
- Fix a unit test for a changed error message in openssl 3.0.2

(mef)

gcc6: workaround: get this at least building by disabling RELRO

(nia)

doc/pkgsrc.*: regen

(nia)

guide: More configuration notes

(nia)

very outdated info

(nia)

More arch preprocessor defines

(nia)

Ensure definitely-not-HTML is definitely escaped

(nia)

State of the art on OS testing macros

(nia)

Update state of the art on autoconf

(nia)

Add a Rust section

(nia)

guide: Notes on C++ standards support in pkgsrc

(nia)

doc: Updated databases/mariadb105-server to 10.5.16

(nia)

doc: Updated databases/mariadb105-client to 10.5.16

(nia)

mariadb105: Update to 10.5.16

                        MariaDB 10.5.16 Release Notes

Notable Items

  InnoDB

    * innodb_disallow_writes removed (MDEV-25975)
    * InnoDB gap locking fixes (MDEV-20605, MDEV-28422)
    * InnoDB performance improvements (MDEV-27557, MDEV-28185)

  Replication

    * Server initialization time gtid_slave_pos purge related reason of
      crashing in binlog background thread is removed (MDEV-26473)
    * Shutdown of the semisync master can't produce inconsistent state
      anymore (MDEV-11853)
    * Binlogs disappear after rsync IST (MDEV-28583)
    * autocommit=0 slave hang is eliminated (DBAAS-7828)
    * master crash is eliminated in compressed semisync replication protocol
      with packet counting amendment (MDEV-25580)
    * OPTIMIZE on a sequence does not cause counterfactual
      ER_BINLOG_UNSAFE_STATEMENT anymore (MDEV-24617)
    * Automatically generated Gtid_log_list_event is made to recognize
      within replication event group as a formal member (MDEV-28550)
    * Replication unsafe INSERT .. ON DUPLICATE KEY UPDATE using two or more
      unique key values at a time with MIXED format binlogging is corrected
      (MDEV-28310)
    * Replication unsafe INSERT .. ON DUPLICATE KEY UPDATE stops issuing
      unnessary "Unsafe statement" with MIXED binlog format (MDEV-21810)
    * Incomplete replication event groups are detected to error out by the
      slave IO thread (MDEV-27697)
    * mysqlbinlog --stop-never --raw now flushes the result file to disk
      after each processed event so the file can be listed with the actual
      bytes (MDEV-14608)

  Backup

    * Incorrect binlogs after Galera SST using rsync and mariabackup
      (MDEV-27524)
    * mariabackup does not detect multi-source replication slave
      (MDEV-21037)
    * Useless warning "InnoDB: Allocated tablespace ID <id> for <tablename>,
      old maximum was 0" during backup stage (MDEV-27343)
    * mariabackup prepare fails for incrementals if a new schema is created
      after full backup is taken (MDEV-28446)

  Optimizer

    * A SEGV in Item_field::used_tables/update_depend_map_for_order...
      (MDEV-26402)
    * ANALYZE FORMAT=JSON fields are incorrect for UNION ALL queries
      (MDEV-27699)
    * Subquery in an UPDATE query uses full scan instead of range
      (MDEV-22377)
    * Assertion `item1->type() == Item::FIELD_ITEM ... (MDEV-19398)
    * Server crashes in Expression_cache_tracker::fetch_current_stats
      (MDEV-28268)
    * MariaDB server crash at Item_subselect::init_expr_cache_tracker
      (MDEV-26164, MDEV-26047)
    * Crash with union of my_decimal type in ORDER BY clause (MDEV-25994)
    * SIGSEGV in st_join_table::cleanup (MDEV-24560)
    * Assertion `!eliminated' failed in Item_subselect::exec (MDEV-28437)

  General

    * Server error messages are now available in Chinese (MDEV-28227)
    * For RHEL/CentOS 7, non x86_64 architectures are no longer supported
      upstream and so our support will also be dropped with this release

  Security

    * Fixes for the following security vulnerabilities:
          * CVE-2022-27376
          * CVE-2022-27377
          * CVE-2022-27378
          * CVE-2022-27379
          * CVE-2022-27380
          * CVE-2022-27381
          * CVE-2022-27382
          * CVE-2022-27383
          * CVE-2022-27384
          * CVE-2022-27386
          * CVE-2022-27387
          * CVE-2022-27444
          * CVE-2022-27445
          * CVE-2022-27446
          * CVE-2022-27447
          * CVE-2022-27448
          * CVE-2022-27449
          * CVE-2022-27451
          * CVE-2022-27452
          * CVE-2022-27455
          * CVE-2022-27456
          * CVE-2022-27457
          * CVE-2022-27458

(nia)

doc: Updated databases/mariadb106-server to 10.6.8

(nia)

doc: Updated databases/mariadb106-client to 10.6.8

(nia)

mariadb106: update to 10.6.8

                          MariaDB 10.6.8 Release Notes

Notable Items

  InnoDB

    * innodb_disallow_writes removed (MDEV-25975)
    * InnoDB gap locking fixes (MDEV-20605, MDEV-28422)
    * InnoDB performance improvements (MDEV-27557, MDEV-28185, MDEV-27767,
      MDEV-28313, MDEV-28137, MDEV-28465, MDEV-26789)
    * Backup regression fixes (MDEV-27919)
    * InnoDB portability: FreeBSD futexes (MDEV-26476), POWER and s390x
      transactional memory (MDEV-27956)
    * ALTER TABLE: Fixed bogus duplicate key errors (MDEV-15250)
    * DDL and crash recovery fixes (MDEV-27274, MDEV-27234, MDEV-27817)
    * Requests to recalculate persistent statistics were sometimes lost
      (MDEV-27805)

  Replication

    * Semisync-slave server recovery is refined to correctly rollback
      prepared transaction (MDEV-28461)
    * Circular semisync setup endless event circulation is handled
      (MDEV-27760)
    * Semisync-slave server recovery is extended to work on new server_id
      server (MDEV-27342)
    * Server initialization time gtid_slave_pos purge related reason of
      crashing in binlog background thread is removed (MDEV-26473)
    * Shutdown of the semisync master can't produce inconsistent state
      anymore (MDEV-11853)
    * Binlogs disappear after rsync IST (MDEV-28583)
    * autocommit=0 slave hang is eliminated (DBAAS-7828)
    * master crash is eliminated in compressed semisync replication protocol
      with packet counting amendment (MDEV-25580)
    * OPTIMIZE on a sequence does not cause counterfactual
      ER_BINLOG_UNSAFE_STATEMENT anymore (MDEV-24617)
    * Automatically generated Gtid_log_list_event is made to recognize
      within replication event group as a formal member (MDEV-28550)
    * Replication unsafe INSERT .. ON DUPLICATE KEY UPDATE using two or more
      unique key values at a time with MIXED format binlogging is corrected
      (MDEV-28310)
    * Replication unsafe INSERT .. ON DUPLICATE KEY UPDATE stops issuing
      unnessary "Unsafe statement" with MIXED binlog format (MDEV-21810)
    * Incomplete replication event groups are detected to error out by the
      slave IO thread (MDEV-27697)
    * mysqlbinlog --stop-never --raw now flushes the result file to disk
      after each processed event so the file can be listed with the actual
      bytes (MDEV-14608)

  Backup

    * Incorrect binlogs after Galera SST using rsync and mariabackup
      (MDEV-27524)
    * mariabackup does not detect multi-source replication slave
      (MDEV-21037)
    * Useless warning "InnoDB: Allocated tablespace ID <id> for <tablename>,
      old maximum was 0" during backup stage (MDEV-27343)
    * mariabackup prepare fails for incrementals if a new schema is created
      after full backup is taken (MDEV-28446)

  Optimizer

    * Query performance degradation in newer MariaDB versions when using
      many tables (MDEV-28073)
    * A SEGV in Item_field::used_tables/update_depend_map_for_order...
      (MDEV-26402)
    * ANALYZE FORMAT=JSON fields are incorrect for UNION ALL queries
      (MDEV-27699)
    * Subquery in an UPDATE query uses full scan instead of range
      (MDEV-22377)
    * Assertion `item1->type() == Item::FIELD_ITEM ... (MDEV-19398)
    * Server crashes in Expression_cache_tracker::fetch_current_stats
      (MDEV-28268)
    * MariaDB server crash at Item_subselect::init_expr_cache_tracker
      (MDEV-26164, MDEV-26047)
    * Crash with union of my_decimal type in ORDER BY clause (MDEV-25994)
    * SIGSEGV in st_join_table::cleanup (MDEV-24560)
    * Assertion `!eliminated' failed in Item_subselect::exec (MDEV-28437)

  General

    * Server error messages are now available in Chinese (MDEV-28227)
    * For RHEL/CentOS 7, non x86_64 architectures are no longer supported
      upstream and so our support will also be dropped with this release
    * Packages for Ubuntu 22.04 LTS "Jammy" and Fedora 36 are not yet
      available pending the resolution of MDEV-28133: Backport OpenSSL-3.0
      compatibility to 10.6 branch

  Security

    * Fixes for the following security vulnerabilities:
          * CVE-2022-27376
          * CVE-2022-27377
          * CVE-2022-27378
          * CVE-2022-27379
          * CVE-2022-27380
          * CVE-2022-27381
          * CVE-2022-27382
          * CVE-2022-27383
          * CVE-2022-27384
          * CVE-2022-27386
          * CVE-2022-27387
          * CVE-2022-27444
          * CVE-2022-27445
          * CVE-2022-27446
          * CVE-2022-27447
          * CVE-2022-27448
          * CVE-2022-27449
          * CVE-2022-27451
          * CVE-2022-27452
          * CVE-2022-27455
          * CVE-2022-27456
          * CVE-2022-27457
          * CVE-2022-27458

(nia)