2016471 - 0.22% installer size (OSX) regression on Thu January 29 2026

Status: NEW

(Core :: Graphics: ImageLib, defect)

Description

[Treeherder Bug Filer]

Perfherder has detected a build_metrics performance regression from push 7f78b26ed53da6433636b6a834d296cf527fa223. As author of one of the patches included in that push, we need your help to address this regression.

Please acknowledge, and begin investigating this alert within 3 business days, or the patch(es) may be backed out in accordance with our regression policy.

Regression	Test	Platform	Options	Absolute values [old vs new]
0.22%	installer size	osx-shippable	nightly	117,891,062.08 -> 118,150,396.83

Support & Actionable Steps

For questions, please contact [e-mail]. You can also find help on Slack (#perf-help) or Matrix (#perftest).

Action guide details:

• See the guide to handling regression bugs for investigation steps.
• Review the alert summary for affected tests, graphs, and comparisons.
• Check the PerfCompare results.
• Tests on Try: Run all tests using ./mach try perf --alert 48365 (See documentation link for details).
• Profiling jobs: Trigger jobs from Treeherder job view or request them from [e-mail].

Comment 1

[Timothy Nikkel (:tnikkel)]

Attachment: Screenshot 2026-02-12 at 10.15.45 AM.png

Comment 2

[Timothy Nikkel (:tnikkel)]

Looking at the screenshot, the vertical line indicates bug 2013142. Enabling simd code makes sense to increase binary size I think.

Looks like we have an decrease just before this, I think this likely corresponds to https://hg-edge.mozilla.org/integration/autoland/rev/45aab1808108 and https://hg-edge.mozilla.org/mozilla-central/rev/7f78b26ed53d which removed the old C++ jpegxl code.

And even after bug 2013142 we are still below the previous baseline. But then at some point after bug 2013142 we increase. If we care about binary size increase then that is the increase that we should bisect down.

Comment 3

[Kagami Rosylight [:saschanaz] (they/them)]

Meaning the bot may be wrong here? 👀

Comment 4

[Timothy Nikkel (:tnikkel)]

(In reply to Kagami Rosylight [:saschanaz] (they/them) from comment #3)

Meaning the bot may be wrong here? 👀

No, the bot is not wrong here. If you zoom in more you can see there is a jump between bug 2013142 and the previous. What happens in that small hour-long window is distinct from what looks like is happening on a week-long window. I'm saying that there is another increase here that happens after bug 2013142, or it's just noise.

So this specific bug I think should be resolved with no action, and if someone wants to look into that other increase they can.

Comment 5

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 2013142

Comment 6

[Kagami Rosylight [:saschanaz] (they/them)]

What happens in that small hour-long window is distinct from what looks like is happening on a week-long window.

Hmm. Hi Andra, any idea how to let the bot bisect that? Maybe just wait and the bot will just do it? 👀

Comment 7

[Andra Esanu (needinfo me)]

Hi :saschanaz, there is no bot. Backfills and retriggers are done manually. Do you need more data on the graph or more retriggers for some specific revisions?

Comment 8

[Timothy Nikkel (:tnikkel)]

Take a look at the attached screenshot. Doesn't it look like there is a different regression that happens after the one that this bug is filed for?

Comment 9

[Kagami Rosylight [:saschanaz] (they/them)]

(In reply to Andra Esanu (needinfo me) from comment #7)

Hi :saschanaz, there is no bot. Backfills and retriggers are done manually. Do you need more data on the graph or more retriggers for some specific revisions?

Yeah, specifically:

1. Comparing https://hg-edge.mozilla.org/integration/autoland/rev/f07383f1170d6f02c21d587f814d72be5c4473cc and https://hg-edge.mozilla.org/integration/autoland/rev/765f2d91912204d920fe0d30b60f60aeffb97d52, is there a significant size regression?
2. Comparing https://hg-edge.mozilla.org/integration/autoland/rev/765f2d91912204d920fe0d30b60f60aeffb97d52 and https://hg-edge.mozilla.org/integration/autoland/rev/ec0f40c330681223008c51175c810336e4445e64, is there a significant size regression?

1 is about this bug and 2 is about some significant size regression after this that seemingly happened around Feb 4, per the graph.

Comment 10

[Kagami Rosylight [:saschanaz] (they/them)]

Attachment: size graph on macos between january and february. At some point around Feb 04 there was a size increase.

Comment 11

[Timothy Nikkel (:tnikkel)]

If you zoom in there's another regression in the day after the one identified in this bug too.