Closed Bug 1443095 Opened 6 years ago Closed 6 years ago

Update graphite2 to release 1.3.11.

Categories

(Core :: Graphics: Text, defect, P3)

defect

Tracking

()

RESOLVED FIXED
mozilla60
Tracking Status
firefox60 --- fixed

People

(Reporter: jfkthame, Assigned: jfkthame)

References

Details

Attachments

(1 file, 1 obsolete file)

This is a minor update that includes fixes for issues identified during code auditing and fuzzing. Would be good to take this now so that it goes into the next ESR, as well as regular firefox.
Attachment #8956003 - Flags: review?(jmuizelaar)
Blocks: 1442830
Attachment #8956003 - Flags: review?(jmuizelaar) → review+
Pushed by jkew@mozilla.com:
https://hg.mozilla.org/integration/mozilla-inbound/rev/f24d341229db
Update graphite2 library to release 1.3.11. r=jrmuizel
Ugh, that's frustrating. :( Let's see if there's an obvious fix...
Flags: needinfo?(jfkthame)
Attached patch followup, fix leak in graphite2 (obsolete) — Splinter Review
I believe this will fix the problem; I've pushed a try run to double-check, and will also raise it upstream.
I've pushed https://github.com/silnrsi/graphite/commit/6e24eb7edbd0872b46441d7397e8b87cccfede73 as a fix that reinstates the use of new [] and delete [] operators as the commit that removed them shouldn't have.
Pushed by jkew@mozilla.com:
https://hg.mozilla.org/integration/mozilla-inbound/rev/e805bff87d63
Update graphite2 library to release 1.3.11 (plus upstream commit 6e24eb7e to fix leak). r=jrmuizel
Comment on attachment 8956143 [details] [diff] [review]
followup, fix leak in graphite2

Obsoleting this, as a different fix for the leak has been landed upstream; and re-landing the update with the upstream fix cherry-picked.
Attachment #8956143 - Attachment is obsolete: true
https://hg.mozilla.org/mozilla-central/rev/e805bff87d63
Status: ASSIGNED → RESOLVED
Closed: 6 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla60
This reduced our build times on Windows! \o/

== Change summary for alert #12055 (as of Sat, 10 Mar 2018 08:45:40 GMT) ==

Improvements:

 26%  build times windows2012-32 opt rusttests taskcluster-c4.4xlarge     2,263.97 -> 1,684.86
 18%  build times windows2012-32 opt taskcluster-c4.4xlarge               2,511.51 -> 2,069.52
  4%  build times windows2012-64 pgo taskcluster-c4.4xlarge               4,684.88 -> 4,491.54

For up to date results, see: https://treeherder.mozilla.org/perf.html#/alerts?id=12055
That seems quite surprising to me; this was quite a minor update.

In particular, a big gain on "rusttests" doesn't seem to make any sense. This library is C++, and doesn't have any direct connection to any of our rust code, AFAIK.

I wonder if something else changed (e.g. in the configuration of the build machines) around the same time?
Blocks: 1436775
Blocks: 1442831
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: