739096 - Update cairo to version 1.17.4

Status: RESOLVED FIXED

(Core :: Graphics, enhancement)

Description

[Boaz Dodin]

A tracking bug for updating the cairo to the just released 1.12 version:
http://cairographics.org/news/cairo-1.12.0/

It's a major stable version that brings many speed improvements (but only):
http://cairographics.org/news/cairo-1.11.4/

Comment 1

[Loic]

Could you add 1.12.0 in the bug title please? (to improve search)

Comment 2

[NVD]

http://ickle.wordpress.com/2012/05/30/cairo-1-12-4-is-just-around-the-corner/

Cairo 1.12.4 is coming soon, can someone update the bug title to 1.12.4?

Comment 3

[xunxun]

Cairo 1.12.2 has been released.

Comment 4

[NVD]

http://www.cairographics.org/news/cairo-1.12.4/

Cairo 1.12.4 has been released, can someone update the bug title to 1.12.4?

Comment 5

[xunxun]

Cairo 1.12.8 has been released.

Comment 6

[Boaz Dodin]

(In reply to xunxun from comment #5)
> Cairo 1.12.8 has been released.

There is no official news release or changelog about this version.

For now the title is updated for 1.12.6
http://www.cairographics.org/news/cairo-1.12.6/

Comment 7

[xunxun]

(In reply to Boaz Dodin from comment #6)
> (In reply to xunxun from comment #5)
> > Cairo 1.12.8 has been released.
> 
> There is no official news release or changelog about this version.
> 
> For now the title is updated for 1.12.6
> http://www.cairographics.org/news/cairo-1.12.6/

Well, I saw http://www.cairographics.org/releases/?C=M;O=A

Comment 8

[NVD]

Cairo 1.12.8 released, please update the bug title.

http://lists.cairographics.org/archives/cairo-announce/2012-November/000206.html

Complete list of changes since 1.12.6
-------------------------------------

Adrian Johnson (5):
      type1-subset: parse all operators
      type1-subset: restore correct callothersub behavior
      type1-subset: ensure subroutine numnber is an integer
      type1-subset: fix memory leak
      type1-subset: remove unused variable

Chris Wilson (19):
      version: Post release bump to 1.12.7
      xlib/shm: Sanity check that the server handles XSendEvent with ShmCompletion
      xlib: Check for both X.org and Xorg ServerVendors
      xlib/shm: Check for XShm headers
      xlib/shm: Use shmstr.h instead of shmproto.h if available
      xlib: Apply the image offsets to the destination rather the source
      pen: First check whether the in/out edges lie within the single pen vertex
      xlib/shm: Fix bogus assertion without shm available
      image: Add a couple of tracepoints for spans fallbacks
      stroke: Precompute the line half-width
      util/show-polygon: Show the limited range of each edge
      spans: Do not assume that we manage to perform the clip geometrically
      xlib: Fixup standalone header compilation for 'make check'
      gl: Tune the default VBO size to reduce overhead on embedded devices
      pen: Tighten checking for bevel (start==stop) joins
      test: Add stroke-clipped
      stroke: Fix calling '_cairo_spline_intersect' for in-bounds checking of splines
      xlib/shm: Need IncludeInferiors when creating the source fallback
      1.12.8 release

Kevin Tardif (1):
      type1-subset, cff-subset: Plugged 2 memory leaks

Comment 9

[Clarke Wixon]

This needs to happen!

...if for no other reason than to address bug 691661, which is a freetype subpixel rendering regression present in cairo 1.10.2 (the version I believe mozilla integrates) but fixed in upstream cairo since 1.11.4 (2012-03-12).

I can't seem to get any attention paid to that bug -- despite it being a clear and obvious one-liner fix -- so perhaps upgrading cairo wholesale will get it resolved.

Comment 10

[randyzie]

Guessing we wont see this happen?

Comment 11

[NVD]

Cairo 1.12.10 released, please update the bug title.

http://www.cairographics.org/news/cairo-1.12.10/

http://lists.cairographics.org/archives/cairo-announce/2013-January/000209.html

Complete list of changes from 1.12.8 to 1.12.10
-----------------------------------------------

Adam Jackson (1):
      xlib/shm: Fix memory leak

Adrian Johnson (1):
      doc: Add CAIRO_MIME_TYPE_UNIQUE_ID to list of supported mime types

Alejandro G. Castro (1):
      gl/msaa: Avoid the stencil buffer when possible during masking

Behdad Esfahbod (3):
      [Minor] Improve logging
      [ft] Remove ancient check for FT_Bitmap_Size.y_ppem
      [ft] Fix resizing of bitmap fonts

Chris Wilson (50):
      version: Post-release bump to 1.12.9
      trace: Fix operand emission for map-to-image and unmap-image
      trace: Do not forcibly add surfaces to the dictionary
      script: Fix map-to-image/unmap stack manipulations
      mempool: Reduce the assertion into an alignment adjustment for the base
      xlib/shm: Populate send_event and serial
      xlib/shm: Rate-limit events and only use as necessary
      xlib/shm: Do not trigger a surplus event from XShmPutImage
      Revert "xlib/shm: Do not trigger a surplus event from XShmPutImage"
      perf/chart: Show the geometric average as an extra column
      perf/chart: Contract the default output filenames
      gl: Use vfunc for vertex emission
      gl: Provide a fast emitter for solid glyphs
      gl: Provide a fast emitter for solid spans
      xlib/shm: Fix typo in creation of a SHM image
      xlib: Use SHM transport for ordinary image uploads
      stroke: Make the incremental trapezoid stroker optionally available again
      xlib: Avoid copying the source twice if it is an image
      scaled-font: Mention ownership of returned object from get_font_face()
      Add missing local slim proto for cairo_recording_surface_create
      gobject: Fix my typo s/TEST/TEXT/ in the previous commit
      script: Recompress strings using LZO whilst binding traces
      xlib/shm: Only mark the shm pixmap as active if we upload into it
      xlib: Simplify source creation by use of map-to-image
      image: Call pixman without a mask for opaque regions of inplace_spans
      script: Attempt to decompress images in place
      script: Simply exchange source/dest images for _set_source_image
      script: Thaw the scaled font cache on the error path
      scaled-font: Always hold the mutex even for single glyph probes
      scaled-font: Free the cached glyphs from the font before taking the global lock
      scaled-font: Assert if attempting to finish a frozen font
      scaled-font: Hold the scaled font mutex whilst reaping from the global cache
      xlib/shm: Discard damage upon shm finish
      xlib/shm: Only destroy an existing damage
      scaled-font: Remove a non-threadsafe double-freeze assert
      image: Allocate a temporary buffer for inline span composition
      scaled-font: Make reset-font-cache threadsafe
      scaled-font: Fix use after free when clearing the glyph cache
      gstate: Use the polygon intermediate for geometry queries
      stroke: Flip the dev slope as well for computing the cusp on a degeneracy
      xlib: map-to-image requires an extents
      xcb: _cairo_scaled_font_reset_cache does it own locking
      xlib: Only fallback through the mask intermediate if we can composite the mask
      xlib: Handle lack of XRenderFillRectangles
      compositor: Convert image surface into backend source
      compositor: Pass back the internal failure
      xlib: Initialise Pixmap for proxy sources
      script: Set decompression length prior to calling decompressors
      1.12.10 release
      version: Post-release bump to 1.12.11

Chuanbo Weng (2):
      gl/msaa: Use GL_IMG_multisampled_render_to_texture when available
      gl: Support the GL_IMG_texture_npot extension

David Maxwell (1):
      type1-subset: always include subroutine 4 (hint replacement idiom)

Henry Song (6):
      gl: Flush context upon evicting a gradient
      gl/msaa: Also setmsaa_active to true for non-texture surfaces
      gl: Properly disable ctx->spans when necessary
      gl/msaa: Add full support for masking with the SOURCE operator
      gl/msaa: Only clear parts of the stencil buffer we will use
      gl: Support for non-texture sources and masks

Kouhei Sutou (2):
      gobject: Add the correct macro name for the hint-metrics type
      gobject: Fix "text_cluster_flags_get_type" typo

Martin Robinson (11):
      gl/msaa: Share the depth/stencil buffer among all surfaces
      gl: Add BGRA download support for GLES2
      Revert "gl/msaa: Share the depth/stencil buffer among all surfaces"
      gl: Cleanup selection of multisampling mode
      boilerplate/glx: Add a target with multisampling and stencil support
      gl: Better handling of clear surfaces
      gl: Do less work when acquiring and releasing devices
      gl/msaa: No need to set the clip when masking
      gl/msaa: Rely on the stencil buffer to cache the clip
      gl/msaa: Check for more extensions before using MSAA
      gl: Follow up fix for the previous commit

Uli Schlachter (4):
      boilerplate-xcb: Ignore MappingNotify events
      context: Use recording surfaces for unbounded groups
      test: Add xcb-huge-subimage
      xcb: Fix xcb-huge-subimage

Zozó Teki (1):
      recording: Append new elements to the end of the bbtree chain

Comment 12

[Boaz Dodin]

(In reply to NVD from comment #11)
> Cairo 1.12.10 released, please update the bug title.
> 
> http://www.cairographics.org/news/cairo-1.12.10/
> 
> http://lists.cairographics.org/archives/cairo-announce/2013-January/000209.
> html

NVD - please post only the link to, or the short version of the release notes and the changelog.

Comment 13

[Dennis Jakobsen]

With WinXP's dwindling market share, i doubt this is a high priority, there are so many other important issues Firefox needs fixed first, before Danish Governments, Chinese & Indian software pirates are attended to.

Comment 14

[Benoit Girard (:BenWa)]

Thanks for the update. We have a few dependencies at the moment that are blocking us but once they are solved we will update cairo.

AFAIK We're looking to switching to using Azure Quartz before updating. Cairo Quartz has historically given us the most headache when upgrading cairo.

Jeff would know better.

Comment 15

[NVD]

Cairo 1.12.12 released.

http://www.cairographics.org/news/cairo-1.12.12/

Comment 16

[NVD]

Cairo 1.12.14 released.

http://www.cairographics.org/news/cairo-1.12.14/

Comment 17

[NVD]

Cairo 1.12.16 released.

http://www.cairographics.org/news/cairo-1.12.16/

Comment 18

[Landry Breuil (:gaston)]

BenWa: can you list the dependencies preventing this bug from being resolved ? Updating the bundled cairo would help reducing the differences with third-parties *nix distributors which atm build with systemwide cairo (usually 1.12.x) and face strange issues (at least an issue with fonts rendering weird sometimes, cant find the bug # right now)

Comment 19

[Landry Breuil (:gaston)]

I was referring to https://bugzilla.mozilla.org/show_bug.cgi?id=856419

Comment 20

[xunxun]

And someone know which edition of cairo mozilla tree has at present.

Comment 21

[Ryan VanderMeulen [:RyanVM]]

12d521df8acc483b2daa844d4f05dc2fe2765ba6, whatever that amounts to.
http://mxr.mozilla.org/mozilla-central/source/gfx/cairo/README?raw=1

Comment 22

[NVD]

https://bugzilla.mozilla.org/show_bug.cgi?id=562746

I think this is the last time cairo was updated in Firefox, could be wrong though.

Comment 23

[Benoit Girard (:BenWa)]

(In reply to Landry Breuil (:gaston) from comment #18)
> BenWa: can you list the dependencies preventing this bug from being resolved
> ? Updating the bundled cairo would help reducing the differences with
> third-parties *nix distributors which atm build with systemwide cairo
> (usually 1.12.x) and face strange issues (at least an issue with fonts
> rendering weird sometimes, cant find the bug # right now)

Right now we have so many patches against Cairo quartz that aren't upstream that the process is very tedious. Once we switch to Azure Quartz this might happen but right now were priotizing performance improvements.

Comment 24

[xunxun]

Since cairo update has some trouble, can we build system cario directly using --enable-system-cairo --enable-system-pixman even on Windows.

Comment 25

[Benoit Girard (:BenWa)]

We'd likely accept a patch (I can check if someone is wanting/needs this) to do this but we wouldn't ship it with an official builds to use the system libraries.

Comment 27

[Ahmad Saleem]

Cairo 1.14.0 released.

http://www.cairographics.org/news/cairo-1.14.0/

Comment 28

[Ahmad Saleem]

Cairo 1.14.2 released.

http://www.cairographics.org/news/cairo-1.14.2/

Comment 29

[Thomas Mayer]

Updating Cairo might also resolve printing problems I experienced in bug 1204551, which deals with PDF output. In respect to PDF generation, there have been a lot of improvements in the meantime: http://cgit.freedesktop.org/cairo/log/?qt=grep&q=pdf

Comment 30

[NVD]

http://cairographics.org/news/cairo-1.14.4/

Cairo 1.14.4 released.

Comment 31

[Zéfling]

http://cairographics.org/news/cairo-1.15.2/

Cairo 1.15.2 released.

Comment 32

[Milan Sreckovic [:milan] (needinfo for best results)]

We're most likely not going to update the "in tree" version of Cairo, but instead move to using system Cairo on some platforms.

Comment 33

[Hussam Al-Tayeb]

(In reply to Milan Sreckovic [:milan] from comment #32)
> We're most likely not going to update the "in tree" version of Cairo, but
> instead move to using system Cairo on some platforms.

You may want to do test runs first with system cairo. It was causing massive amounts of crashes in firefox under 1.14.xx.

Comment 34

[Boaz Dodin]

(In reply to Hussam Al-Tayeb from comment #33)
> (In reply to Milan Sreckovic [:milan] from comment #32)
> > We're most likely not going to update the "in tree" version of Cairo, but
> > instead move to using system Cairo on some platforms.
> 
> You may want to do test runs first with system cairo. It was causing massive
> amounts of crashes in firefox under 1.14.xx.

I am some confused - how Firefox crashes with 1.14.xx cairo if the "in tree" version not been updated for years and using system version just considered now?

BTW - how I can check what version of cairo used by particular build of Firefox?

Comment 35

[Hussam Al-Tayeb]

(In reply to Boaz Dodin from comment #34)
> I am some confused - how Firefox crashes with 1.14.xx cairo if the "in tree"
> version not been updated for years and using system version just considered
> now?
It is because the intree version has not been updated in years so firefox is not tested properly to work with newer cairo versions (the ones your linux distribution uses).
go to about:buildconfig and look for a --with-system-cairo. If you find it there, firefox is using the system version of cairo (which may lead to crashes). In that cause, use your distribution's package manager to find the cairo version.
if no "--with-system-cairo" is found under about:buildconfig, then firefox is using in-tree one which is like 1.9.5 or something.

Comment 36

[Hussam Al-Tayeb]

Note that Fedora, Archlinux, etc... use the intree cairo to avoid crashes.

Comment 37

[Boaz Dodin]

So it looks like a lot of Linux users are with "in tree" cairo from several years ago.

May I ask what was the rational to not update it, while cairo is an active project?

Comment 38

[Jeff Muizelaar [:jrmuizel]]

(In reply to Boaz Dodin from comment #37)
> So it looks like a lot of Linux users are with "in tree" cairo from several
> years ago.
> 
> May I ask what was the rational to not update it, while cairo is an active
> project?

Our version of the win32 backend has diverged significantly from the upstream one for performance reasons. We plan on dropping our dependency on this backend and after we do that it should be much easier for us to update cairo.

Comment 39

[Hussam Al-Tayeb]

I just finished a firefox trunk build with --enable-system-cairo. I'll test how stable it is. The system cairo version is 1.15.2

Comment 40

[Thomas Mayer]

Using ubuntu 16.04, I tried --enable-system-cairo as described in bug #1204551. Now that ubuntu 16.04 comes along with libcairo version 1.14.6-1, --enable-system-cairo fixed the whole issue #1204551 (Kyocera fs-1020d can't print complex pdf output sometimes), which is a problem I struggled with for so many years.

Comment 41

[Takanori MATSUURA]

1.14.8 and 1.15.4 is available.
https://www.cairographics.org/news/cairo-1.14.8/
https://www.cairographics.org/news/cairo-1.15.4/

BTW 1.15.4 is a snapshot version. Which version is our target in this bug?

Comment 42

[Masatoshi Kimura [:emk]]

I'm not sure this bug is still relevant. Our in-tree cairo has many local changes and too old. We should simply migrate to Skia and drop the cairo backend.

Comment 44

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Remove support for CGLayer-backed cairo quartz surface. r=jrmuizel

This was a performance optimization that we no longer care about.
Removing it will avoid the need to merge the additions into the new
cairo-quartz-surface code from upstream.

Comment 45

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Remove cairo_quartz_{get,finish}_cg_context_with_clip. r=jrmuizel

No longer needed in Gecko.

Depends on D112556

Comment 46

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Update cairo to commit 277a1daec80cb6cf7bfb0e200cf78e7842cb2f82 (release 1.17.4 + post-release fixes on trunk) from https://gitlab.freedesktop.org/cairo/cairo/-/tree/master. r=jrmuizel

Depends on D112557

Comment 47

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Update moz.build for new cairo sources. r=jrmuizel

Depends on D112558

Comment 48

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/b5e84a97833e8e1d082f4409383b09f9827ada09

Quartz: Ensure that image data and surface are available to draw.

Snapshot the cairo_surface_t and copy the image data to provide to
the CGDataProvider so that it is independent of the cairo_surface_t's
lifetime.

Depends on D112559

Comment 49

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/1ddfccca31e23820a30e8f618e216fe2931b49b2

Quartz image drawing: Remove containers for cairo_surface_t.

Since we now copy the data that CGImage needs we don't need to
keep the surface around anymore, nor release it or the image in the
DataProviderReleaseCallback.

Depends on D112560

Comment 50

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/0bf2c140d2dc6521c6f925b1757f2d4809eabf18

Fix two uninitialised variables in _cairo_recording_surface_snapshot()

Depends on D112561

Comment 51

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/7ac8220c1fd4afb259e53b7a5cd93a23859bdbda

Fix copying tags in a recording surface

Depends on D112562

Comment 52

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/cb3618f76d10c2e0cd1e6196ed79d4af4d7d5e44

Added checks for failed strdups in cairo-recording-surface.c

Depends on D112563

Comment 53

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/e45875142fc5d0bb7968b150f379172130415342

Slightly improve dealing with error snapshots

Depends on D112564

Comment 54

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/b29d0f3d582bbafdbac1bb0b21de3994296aedf6

recording-surface: Fix offset error

Depends on D112565

Comment 55

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/48194cf093d6eadf1a64f45d76b1fe001f467a05

Avoid use after free in cairo_fill

Depends on D112566

Comment 56

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/44f808fce9f437e14f2b0ef4e1583def8ab578ae

Avoid a use-after-scope

Depends on D112567

Comment 57

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/ee90ce594678ff227a28cdf1d9b55b7c3494f49c

Plug a memory leak in an error case

Depends on D112568

Comment 58

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/f5d32c071f25604f60b38ba5fdc91ce226f9efc9

Add missing call to cairo_surface_reference

Depends on D112569

Comment 59

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/941cddfa39584e812d41b16fc80c67a2f9007658

'Fix' a NULL pointer 'dereference' in cairo-pdf-surface

Depends on D112570

Comment 60

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/891468f191cccc80b8e491318e0646a7fffa22bf

Add a bounds check to cairo_cff_font_read_fdselect()

Depends on D112571

Comment 61

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/974791b4eede7f2ff774b56dd90234ed2cd70311

Fix _cairo_surface_paint not setting is_clear

Depends on D112572

Comment 62

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/1904d7bf35099f775d68cb7203a34ac7a4e72514

Rename cairo_lines_compare_at_y into _cairo_lines_compare_at_y and fix syntax

Depends on D112573

Comment 63

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/26663cf3be1d7f67b49aa02428d7e303b7286f98

Fix undefined left-shifts

Depends on D112574

Comment 64

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/277a1daec80cb6cf7bfb0e200cf78e7842cb2f82

Fix out of bounds access in cairo_type1_font_subset_find_segments

Depends on D112575

Comment 65

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/f4e43b6822ec73f2fd8046859315e5d2a5709bf5

pdf font subset: Generate valid font names

Depends on D112576

Comment 66

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/c62c43a8c653f5e4d20ef868065d4f882df9768e

Fix a memory leak with cairo_tag_begin() + pdf

Depends on D112577

Comment 67

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/d49d8558199fb594a57bd84d386928aa8ebcfb7e

cairo_win32_surface_get_dc: Return NULL for error surfaces

Depends on D112578

Comment 68

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/38e486b34d435130f2fb38c429e6016c3c82cd53

cairo-quartz-image-surface.c: Fix types

Depends on D112579

Comment 69

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Cherry-pick https://gitlab.freedesktop.org/cairo/cairo/-/commit/1569dcd4d8a2235052aa227dc9f5708cccaea23e

UBSan fixes in cairo-truetype-subset.c and cairo-array.c

Depends on D112580

Comment 70

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Remove obsolete CAIRO_VERSION_* macros. r=jrmuizel

Depends on D112581

Comment 71

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Patch cairo-xlib-xrender-private.h to work around lack of full configure support. r=jrmuizel

Depends on D112582

Comment 72

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Don't declare local var with 'register' in C++. r=jrmuizel

Depends on D112583

Comment 73

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Expose cairo_font_options_{get,set}_lcd_filter as public API. r=jrmuizel

Depends on D112584

Comment 74

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Add cairo_surface_{get,set}_subpixel_antialiasing. r=jrmuizel

Depends on D112585

Comment 75

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Add a synth_flags parameter to cairo_ft_font_face_create_for_ft_face. r=jrmuizel

Depends on D112586

Comment 76

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Hook up support for SharedFTFace in the cairo-ft-font code. r=jrmuizel

Depends on D112587

Comment 77

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Look up new FreeType variation APIs at runtime, in case we were built against an older installed freetype. r=jrmuizel

Depends on D112588

Comment 78

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Merge some Mozilla additions for DirectWrite support. r=jrmuizel

Depends on D112589

Comment 79

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Merge some Mozilla extensions to cairo-quartz-surface. r=jrmuizel

Depends on D112590

Comment 80

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Adjust wpt fuzzy annotations for cairo update r=jrmuizel

Depends on D112591

Comment 81

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/dbb07ebff6b6
Remove support for CGLayer-backed cairo quartz surface. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/d058e7c043e6
Remove cairo_quartz_{get,finish}_cg_context_with_clip. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/e80fbceb9023
Update cairo to commit 277a1daec80cb6cf7bfb0e200cf78e7842cb2f82 (release 1.17.4 + post-release fixes on trunk) from https://gitlab.freedesktop.org/cairo/cairo/-/tree/master. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/2623ebd24944
Update moz.build for new cairo sources. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/1dfadbc743b9
Remove obsolete CAIRO_VERSION_* macros. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/df4fd9df9d67
Patch cairo-xlib-xrender-private.h to work around lack of full configure support. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/1675a980f8f3
Don't declare local var with 'register' in C++. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/67d7ab4394ab
Expose cairo_font_options_{get,set}_lcd_filter as public API. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/566742bad252
Add cairo_surface_{get,set}_subpixel_antialiasing. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/6c67d3e98e1e
Add a synth_flags parameter to cairo_ft_font_face_create_for_ft_face. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/08c8584cadbb
Hook up support for SharedFTFace in the cairo-ft-font code. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/46598f67b43a
Look up new FreeType variation APIs at runtime, in case we were built against an older installed freetype. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/d4ffa7c2db22
Merge some Mozilla additions for DirectWrite support. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/fb2b5f607015
Merge some Mozilla extensions to cairo-quartz-surface. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/75e6eca97890
Adjust wpt fuzzy annotations for cairo update r=jrmuizel

Comment 82

[Cristian Brindusan [:cbrindusan]]

Backed out 15 changesets (bug 739096) for causing valgrind bustages.
https://hg.mozilla.org/integration/autoland/rev/3162683f93cec511a2ff95d3d1d90d37370ba125

Push with failures:
https://treeherder.mozilla.org/jobs?repo=autoland&revision=75e6eca97890cf3b5e4ee753daca5abae683397d&selectedTaskRun=UF1HfebaRfOAz5kdmEKMwQ.0

Failure log:
https://treeherder.mozilla.org/logviewer?job_id=338068945&repo=autoland&lineNumber=61405

Comment 83

[Jonathan Kew [:jfkthame]]

Attachment: Bug 739096 - Don't leave the shm field uninitialized in cairo_xlib_display. r?jrmuizel

This resulted in a valgrind failure on autoland.

Submitted upstream at https://gitlab.freedesktop.org/cairo/cairo/-/merge_requests/170.

Comment 84

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/698ee02c6416
Remove support for CGLayer-backed cairo quartz surface. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/b7ea562a84c7
Remove cairo_quartz_{get,finish}_cg_context_with_clip. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/d41a2bcb4ff2
Update cairo to commit 277a1daec80cb6cf7bfb0e200cf78e7842cb2f82 (release 1.17.4 + post-release fixes on trunk) from https://gitlab.freedesktop.org/cairo/cairo/-/tree/master. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/7769e86f921a
Update moz.build for new cairo sources. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/da66cd84f284
Remove obsolete CAIRO_VERSION_* macros. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/7821deb4d77b
Patch cairo-xlib-xrender-private.h to work around lack of full configure support. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/e37242f1457e
Don't declare local var with 'register' in C++. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/c26360666d39
Expose cairo_font_options_{get,set}_lcd_filter as public API. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/b891ee09f246
Add cairo_surface_{get,set}_subpixel_antialiasing. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/80b1d39dfebd
Add a synth_flags parameter to cairo_ft_font_face_create_for_ft_face. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/4ffec5746bf7
Hook up support for SharedFTFace in the cairo-ft-font code. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/3dc1308521e9
Look up new FreeType variation APIs at runtime, in case we were built against an older installed freetype. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/e9524df5bcb4
Merge some Mozilla additions for DirectWrite support. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/b506da63486a
Merge some Mozilla extensions to cairo-quartz-surface. r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/543f30ace90e
Adjust wpt fuzzy annotations for cairo update r=jrmuizel
https://hg.mozilla.org/integration/autoland/rev/7387be4b195f
Don't leave the shm field uninitialized in cairo_xlib_display. r=jrmuizel

Comment 85

[Narcis Beleuzu [:NarcisB]]

https://hg.mozilla.org/mozilla-central/rev/698ee02c6416
https://hg.mozilla.org/mozilla-central/rev/b7ea562a84c7
https://hg.mozilla.org/mozilla-central/rev/d41a2bcb4ff2
https://hg.mozilla.org/mozilla-central/rev/7769e86f921a
https://hg.mozilla.org/mozilla-central/rev/da66cd84f284
https://hg.mozilla.org/mozilla-central/rev/7821deb4d77b
https://hg.mozilla.org/mozilla-central/rev/e37242f1457e
https://hg.mozilla.org/mozilla-central/rev/c26360666d39
https://hg.mozilla.org/mozilla-central/rev/b891ee09f246
https://hg.mozilla.org/mozilla-central/rev/80b1d39dfebd
https://hg.mozilla.org/mozilla-central/rev/4ffec5746bf7
https://hg.mozilla.org/mozilla-central/rev/3dc1308521e9
https://hg.mozilla.org/mozilla-central/rev/e9524df5bcb4
https://hg.mozilla.org/mozilla-central/rev/b506da63486a
https://hg.mozilla.org/mozilla-central/rev/543f30ace90e
https://hg.mozilla.org/mozilla-central/rev/7387be4b195f

Comment 86

[Alexandru Ionescu (needinfo me) [:alexandrui]]

(In reply to Pulsebot from comment #84)

https://hg.mozilla.org/integration/autoland/rev/7387be4b195f
Don't leave the shm field uninitialized in cairo_xlib_display. r=jrmuizel

The regression below was fixed for some reason in Bug 739096.
== Change summary for alert #29954 (as of Tue, 04 May 2021 05:24:22 GMT) ==

Regressions:

Ratio	Suite	Test	Platform	Options	Absolute values (old vs new)
0.10%	installer size		osx-shippable	instrumented	115,119,222.58 -> 115,232,483.83

For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=29954

Comment 87

[Marco Castelluccio [:marco]]

This fixed https://github.com/mozilla/pdf.js/issues/12282.