Reland: Compile FreeType with HarfBuzz support

Reland: Compile FreeType with HarfBuzz support

FreeType's autohinter uses HarfBuzz API to collect additional GSUB and
GPOS mappings to detect ligatures that should be aligned by the
autohinter. Previously we were not able to build FreeType with HarfBuzz
support because of the cyclic dependency. This CL resolves the cyclic
dependency by building a bootstrap FreeType in order to build HarfBuzz'
hb_ft_* functions as a static library called harfbuzz-ng-ft. Then we
build harfbuzz-ng separately (which does not depend on FreeType), then
we build FreeType depending on harfbuzz-ng and harfbuzz-ng-ft.

This CL also removes the previous pangoft2 link hack since the
:harfbuzz-ng target now brings all symbols required by pangoft2.

This resolves issues with fi and ffi ligatures in Roboto looking like
they were shifted to a different baseline.

I tried developing a pixel test for this, which works if I force usage
of the FreeType autohinter through SkPaint::kSlight_Hinting, however we
are currently unable to automatically test this since our Linux layout
tests do not exercise the autohinting code and do not set this hinting
mode, probably due to the special fontconfig settings that we are using
for the layout tests. Manually verifying the Roboto ligatures however
confirms that this works.

Reland after revert in https://codereview.chromium.org/2879843003/

BUG=617168
Review-Url: https://codereview.chromium.org/2880223002
Cr-Commit-Position: refs/heads/master@{#472413}
Committed: https://chromium.googlesource.com/chromium/src/+/fbec38011870753cffb3cb2ff98aa0c00c63b25c

[drott, 2017-05-15 09:24:06]

The CQ bit was checked by drott@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-15 09:24:26]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-05-15 09:34:24]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-15 09:34:25]

Dry run: Try jobs failed on following builders:
  chromeos_amd64-generic_chromium_compile_only_ng on
master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromeos_amd64-...)
  chromeos_daisy_chromium_compile_only_ng on master.tryserver.chromium.linux
(JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromeos_daisy_...)

[drott, 2017-05-15 11:17:57]

The CQ bit was checked by drott@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-15 11:18:13]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-05-15 11:30:53]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-15 11:30:54]

Dry run: Try jobs failed on following builders:
  linux_chromium_chromeos_ozone_rel_ng on master.tryserver.chromium.linux
(JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[drott, 2017-05-15 12:20:32]

The CQ bit was checked by drott@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-15 12:20:54]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-05-15 12:34:13]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-15 12:34:14]

Dry run: Try jobs failed on following builders:
  chromeos_amd64-generic_chromium_compile_only_ng on
master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromeos_amd64-...)

[drott, 2017-05-15 12:37:25]

The CQ bit was checked by drott@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-15 12:37:40]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-05-15 12:52:06]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-15 12:52:07]

Dry run: Try jobs failed on following builders:
  linux_chromium_compile_dbg_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[drott, 2017-05-15 15:38:02]

The CQ bit was checked by drott@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-15 15:38:28]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[drott, 2017-05-15 15:39:56]

The CQ bit was checked by drott@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-15 15:40:42]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[drott, 2017-05-15 15:47:09]

drott@chromium.org changed reviewers:
+ nicholss@chromium.org, zijiehe@chromium.org

[drott, 2017-05-15 15:50:32]

drott@chromium.org changed reviewers:
+ bungeman@chromium.org, dpranke@chromium.org, erg@chromium.org,
estade@chromium.org

[drott, 2017-05-15 15:50:32]

zijiehe@, nicholss@: Please review the change to the remoting BUILD.gn file.


> # You can only be a GTK owner if your name starts with "e".

erg@, estade@, requesting review for the change to the libgtkui BUILD.gn file
from someone who's username starts with d, not with e - which is close, but
maybe I get there someday.

[nicholss, 2017-05-15 15:52:41]

lgtm for remoting/host/it2me/BUILD.gn

[commit-bot: I haz the power, 2017-05-15 16:34:41]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-15 16:34:42]

Dry run: Try jobs failed on following builders:
  linux_chromium_asan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[Evan Stade, 2017-05-15 16:40:06]

estade@chromium.org changed reviewers:
+ bungeman-chromium. thomasanderson@chromium.org
- bungeman@chromium.org

[Evan Stade, 2017-05-15 16:40:07]

https://codereview.chromium.org/2880223002/diff/100001/chrome/browser/ui/libg...
File chrome/browser/ui/libgtkui/BUILD.gn (right):

https://codereview.chromium.org/2880223002/diff/100001/chrome/browser/ui/libg...
chrome/browser/ui/libgtkui/BUILD.gn:120: "//third_party/harfbuzz-ng",
this looks like something thomasanderson@ knows about, so I defer to him.

[drott, 2017-05-15 17:32:32]

The CQ bit was checked by drott@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-15 17:33:50]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Hzj_jie, 2017-05-15 17:53:21]

On 2017/05/15 16:40:07, Evan Stade wrote:
>
https://codereview.chromium.org/2880223002/diff/100001/chrome/browser/ui/libg...
> File chrome/browser/ui/libgtkui/BUILD.gn (right):
> 
>
https://codereview.chromium.org/2880223002/diff/100001/chrome/browser/ui/libg...
> chrome/browser/ui/libgtkui/BUILD.gn:120: "//third_party/harfbuzz-ng",
> this looks like something thomasanderson@ knows about, so I defer to him.

+1 lgtm for remoting/host/it2me/BUILD.gn

[commit-bot: I haz the power, 2017-05-15 19:04:41]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-15 19:04:42]

Dry run: This issue passed the CQ dry run.

[Tom Anderson, 2017-05-16 18:00:01]

libgtkui lgtm

[Elliot Glaysher, 2017-05-16 18:08:34]

On 2017/05/16 18:00:01, Thomas Anderson wrote:
> libgtkui lgtm

owners lgtm since tom is happy with it

[Dirk Pranke, 2017-05-16 22:53:32]

lgtm

[drott, 2017-05-17 06:18:46]

The CQ bit was checked by drott@chromium.org

[commit-bot: I haz the power, 2017-05-17 06:19:38]

CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-05-17 10:49:08]

CQ is committing da patch.

Bot data: {"patchset_id": 100001, "attempt_start_ts": 1495001926534210,
"parent_rev": "1d357b6a46d70097aa28428b1f106284f7fd9990", "commit_rev":
"fbec38011870753cffb3cb2ff98aa0c00c63b25c"}

[commit-bot: I haz the power, 2017-05-17 10:49:19]

Description was changed from

==========
Reland: Compile FreeType with HarfBuzz support

FreeType's autohinter uses HarfBuzz API to collect additional GSUB and
GPOS mappings to detect ligatures that should be aligned by the
autohinter. Previously we were not able to build FreeType with HarfBuzz
support because of the cyclic dependency. This CL resolves the cyclic
dependency by building a bootstrap FreeType in order to build HarfBuzz'
hb_ft_* functions as a static library called harfbuzz-ng-ft. Then we
build harfbuzz-ng separately (which does not depend on FreeType), then
we build FreeType depending on harfbuzz-ng and harfbuzz-ng-ft.

This CL also removes the previous pangoft2 link hack since the
:harfbuzz-ng target now brings all symbols required by pangoft2.

This resolves issues with fi and ffi ligatures in Roboto looking like
they were shifted to a different baseline.

I tried developing a pixel test for this, which works if I force usage
of the FreeType autohinter through SkPaint::kSlight_Hinting, however we
are currently unable to automatically test this since our Linux layout
tests do not exercise the autohinting code and do not set this hinting
mode, probably due to the special fontconfig settings that we are using
for the layout tests. Manually verifying the Roboto ligatures however
confirms that this works.

Reland after revert in https://codereview.chromium.org/2879843003/

BUG=617168
==========

to

==========
Reland: Compile FreeType with HarfBuzz support

FreeType's autohinter uses HarfBuzz API to collect additional GSUB and
GPOS mappings to detect ligatures that should be aligned by the
autohinter. Previously we were not able to build FreeType with HarfBuzz
support because of the cyclic dependency. This CL resolves the cyclic
dependency by building a bootstrap FreeType in order to build HarfBuzz'
hb_ft_* functions as a static library called harfbuzz-ng-ft. Then we
build harfbuzz-ng separately (which does not depend on FreeType), then
we build FreeType depending on harfbuzz-ng and harfbuzz-ng-ft.

This CL also removes the previous pangoft2 link hack since the
:harfbuzz-ng target now brings all symbols required by pangoft2.

This resolves issues with fi and ffi ligatures in Roboto looking like
they were shifted to a different baseline.

I tried developing a pixel test for this, which works if I force usage
of the FreeType autohinter through SkPaint::kSlight_Hinting, however we
are currently unable to automatically test this since our Linux layout
tests do not exercise the autohinting code and do not set this hinting
mode, probably due to the special fontconfig settings that we are using
for the layout tests. Manually verifying the Roboto ligatures however
confirms that this works.

Reland after revert in https://codereview.chromium.org/2879843003/

BUG=617168

Review-Url: https://codereview.chromium.org/2880223002
Cr-Commit-Position: refs/heads/master@{#472413}
Committed:
https://chromium.googlesource.com/chromium/src/+/fbec38011870753cffb3cb2ff98a...
==========

[commit-bot: I haz the power, 2017-05-17 10:49:22]

Committed patchset #6 (id:100001) as
https://chromium.googlesource.com/chromium/src/+/fbec38011870753cffb3cb2ff98a...