[cairo] Release candidate for 1.4.12 (testing requested!)

[Carl Worth]

I've just done some work to assemble a release candidate for
1.4.12. Things you will find:

A tar file snapshot at:

http://cairographics.org/snapshots/cairo-1.4.11-20071120-e76d05.tar.gz

And a corresponding new branch named "1.4" in cairo's central
repository.

I constructed this branch by cherry-picking 58 separate bug-fix
commits from the 1.5 branch, (all but the last two of which were
already present in the 1.5.2 snapshot). I chose commits by reading the
log of all changes from 1.4.10..master looking for the following kinds
of things:

  * Things that sounded like bug fixes

  * Things that look self-contained

  * Things that applied cleanly to the 1.4 branch

And avoiding these kinds of things:

  * Things that introduced new API/significant features

  * Things that were stretched over many commits, (especially of the
    "oops, fixup regression introduced by that last commit" variety)

  * Things that primarily touched experimental backends, (though I did
    take a lot of quartz/ATSUI fixes)

  * Things that introduced significant conflicts when cherry-picked

So I didn't bring over most of Chris Wilson's error-handling
series. And I didn't bring over some PS/PDF things that could be
considered strictly bug fixes since they were too dependent on the
large changes to the fancy new fine-grained fallback rework.

So the hope is that this is a nice release, of similar robustness to
1.4.10 but with many additional fixes. Some people doing packages for
very conservative distributions might avoid using this new 1.4.12
directly, but might use it as a starting point for looking for the
most significant bug fixes to patch onto 1.4.10.

Before I release this officially, I'd like some review of the work
I've done. Here are some of the things I'd like to hear:

  * Reports of a smoketest on win32

  * Reports of a smoketest on quartz

  * Reports of any testing that people would like to do, (test suite
    or application-level testing)

  * Any commits in the 1.4 branch that anyone thinks should be removed

  * Any commits already on master that anyone things should be added

The list of commits included in this series is below.

I'll definitely wait over 24 hours before I make an official 1.4.12
release, but if anyone wants me to wait longer for specific testing,
please let me know. Otherwise I'll likely make the release on the
morning of Thursday, Nov. 22.

Thanks in advance. And have fun with cairo, everybody!

-Carl

Adrian Johnson (8):
      CFF Subsetting: fix integer encoding bug
      PDF: Fix gradients regression
      PDF: Fix regression in pdf_surface_set_size()
      Fix PDF gradients bug
      Eliminate unclipped CLEAR from meta-surface playback
      Create meta-surface source image with same content type as the
      Truetype Subsetting: Avoid failing when fonts are missing optional tables
      CFF Subetting: Fix for #10849

Behdad Esfahbod (6):
      [configure.in] Make PS backend require zlib too (#12210)
      [configure.in] Fix non-pkgconfig Xrender detection path (#4724)
      =?utf-8?q?[ChangeLog.mk]=20Make=20make=20rule=20depend=20on=20.git/HEAD,=20not=20.git
      [cairo-ft-font] Ignore FT_Load_Glyph errors other than out-of-memory
      [ChangeLog.mk] Fix changelog generation rules
      [cairo-xlib] Release glyph surfaces if we made them to be generated

Brian Ewins (15):
      [type1-subset] only subset ft fonts
      do not ignore errors in glyph paths
      do not ignore ATSUI errors.
      do not ignore errors when creating styles.
      do not ignore errors in setup_source
      do not ignore errors, return NULL instead (moz#874315)
      do not return a cairo_status_t of UNSUPPORTED
      call _cairo_error on failure
      Implement EXTEND_NONE for gradients.
      fix dash-zero-length test
      implement clipping with surface masks.
      [quartz] return status instead of CGShadingRef
      [quartz] return status instead of cairo_quartz_surface_t
      [quartz] return status instead of CGPatternRef
      [configure.in] fix atsui experimental warning

Carl Worth (8):
      * Update version to 1.4.11 after the 1.4.10 release
      Ensure that a copied pattern gets its own user_data array
      Ensure the Render extension is initialized before calling XESetCloseDisplay
      Fix typo in _cairo_gstate_fini
      Limit miters for small angles
      Fix degenerate-pen test case by removing the triggering assertion
      * Update two PS reference images for mysterious single-pixel change
      * Mark bitmap-font as XFAIL

      * These comits are new on the 1.4 branch. Everything else was
        already on master and simply cherry-picked.

Chris Heath (1):
      [autogen.sh] Tolerate tput failure (#12757)

Chris Wilson (10):
      [cairo-ft-font] Decrement lock count on error.
      [cairo-ft-font] Early detection of a zero sized bitmap.
      [cairo-path] Don't raise an error when attempting to create an empty path.
      [malloc] Check for integer overflow when realloc'ing.
      [malloc] Take advantage of calloc() argument checking.
      [cairo-path] Check for an empty path in cairo_append_path().
      [cairo-scaled-font] Acquire mutex around _cairo_scaled_font_glyph_path().
      [cairo-xlib-surface] Match content to xrender_format using the channel masks.
      [cairo-xlib] Check source for XRender support.
      [cairo-surface] Fix typo in doc.

Keith Packard (1):
      Force non-AA text when using a bitmap strike with only scaling transform.

Richard Hult (1):
      [atsui] Use ascent+descent+leading to get the distance between baselines

Vladimir Vukicevic (10):
      [fix] Avoid int overflow when allocating large buffers
      [win32] check correct surface for BitBlt support in get_subimage
      Implement fallback for clone_similar
      [win32] call free() and not scaled_font_destroy() if scaled init fails
      Fix previous create_similar fallback patch
      [win32] call scaled_font_fini in create before returning NULL
      Make NO_MUTEX really mean no mutexes
      Avoid divide-by-zero when trying to allocate a 0-sized array
      [win32] report error correctly if CreateCompatibleBitmap fails
      [win32] return a nil surface, not NULL

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.cairographics.org/archives/cairo/attachments/20071120/5af31db1/attachment.pgp