1294738 - 0.54 - 1.05% installer size (linux32, windows8-64, windowsxp) regression on push bfb4f7467dd4b39005d61fbcb2872da14a29d66b (Fri Aug 5 2016)

Status: RESOLVED FIXED

(DevTools :: Debugger, defect, P2)

Description

[William Lachance (:wlach)]

It looks like bug 1288511 increases the size of the installer rather significantly. I figured I should file this bug to track this. James, is there a plan to mitigate this increase?

==

We have detected a build metrics regression from push bfb4f7467dd4b39005d61fbcb2872da14a29d66b. As author of one of the patches included in that push, we need your help to address this regression.

Summary of tests that regressed:

  installer size summary linux32 pgo: 59756192.25 -> 60078766.25 (0.54% worse)
  installer size summary windowsxp pgo: 53975031.42 -> 54539300.25 (1.05% worse)
  installer size summary windows8-64 pgo: 56946469.17 -> 57528616.25 (1.02% worse)


You can find links to graphs and comparison views for each of the above tests at: https://treeherder.mozilla.org/perf.html#/alerts?id=2315

On the page above you can see an alert for each affected platform as well as a link to a graph showing the history of scores for this test. There is also a link to a treeherder page showing the jobs in a pushlog format.

To learn more about the regressing test(s), please see: https://developer.mozilla.org/en-US/docs/Mozilla/Performance/Automated_Performance_Testing_and_Sheriffing/Build_Metrics

Comment 1

[James Long (:jlongster)]

Yes, we are going to follow-up and remove about half of the size of that JS by sharing modules with the rest of devtools. We will do this soon, within a week possibly. Is that ok?

Comment 2

[William Lachance (:wlach)]

(In reply to James Long (:jlongster) from comment #1)
> Yes, we are going to follow-up and remove about half of the size of that JS
> by sharing modules with the rest of devtools. We will do this soon, within a
> week possibly. Is that ok?

Sure, that's fine. Shall we use this bug to track that?

Comment 3

[James Long (:jlongster)]

I just filed bug 1294749 for us, so we can close this bug unless it's helpful for you.

Comment 4

[William Lachance (:wlach)]

The sheriffing policy for installer size regressions is kind of adhoc, but let's just leave this open I guess. I'll mark bug 1294749 as a dependency.

Comment 5

[Eddy Bruel [:ejpbruel]]

James said he will get to this soon (see comment 1), so I'm triaging this as P2.

Comment 6

[James Long (:jlongster)]

This will be fixed with bug 1301487 when we push the new debugger bundle. The new debugger size has gone from 2.2MB to ~1.3MG because we're sharing more code with devtools.

Comment 7

[Gerry Chang [:gchang]]

Hi :wlach,
Would you mind checking if the issue still exists since bug 1301487 is fixed?

Comment 8

[William Lachance (:wlach)]

(In reply to Gerry Chang [:gchang] from comment #7)
> Hi :wlach,
> Would you mind checking if the issue still exists since bug 1301487 is fixed?

Yeah, the numbers went back down again! Feel free to resolve this if you're satisfied with the results.

== Change summary for alert #3203 (as of September 12 2016 12:18 UTC) ==

Improvements:

  1%  installer size summary windowsxp pgo       55086537.42 -> 54732163.83
  1%  installer size summary windows8-64 pgo     58102283.46 -> 57759777.42
  0%  installer size summary linux32 pgo         59854719.08 -> 59680678

For up to date results, see: https://treeherder.mozilla.org/perf.html#/alerts?id=3203

Comment 9

[Gerry Chang [:gchang]]

Mark 51 as fixed since bug 1301487 is fixed and is verified in comment #8.