On any tree where we run non-PGO Linux64 builds (and where we have codesighs enabled, which will not include try after bug 798225 and before bug 798229), take a look at the build log or the line in the bottom right part of tbpl when the B is selected - for the current codesize, we report "Z: 0B" which is surprisingly lean and mean. For Linux32 non-PGO, we report a number (~45MB), for Linux32 PGO we report a number (~53MB), for Linux64 PGO we report a number (~143MB), but whether or not any of those number are reasonable and have actual meaning, they look much more meaningful than Linux64 non-PGO, 0B.
The values codesighs return have been useless for a very long time. It should be replaced with something else.
confusion alert - checked something into inbound with this bug number when the correct bug number was 798423. Sorry!
(In reply to Mike Hommey [:glandium] from comment #1) > The values codesighs return have been useless for a very long time. It > should be replaced with something else. Conveniently, Ted already wrote that something else, in bug 659950, it's just taken so long to deploy that I'd completely forgetten it, despite having seen the code while spending two hours filing multiple bugs about the brokenness of codesighs.
Status: NEW → RESOLVED
Last Resolved: 6 years ago
Resolution: --- → WONTFIX
Assignee: nobody → mcmanus
(In reply to Patrick McManus [:mcmanus] from comment #2) > confusion alert - checked something into inbound with this bug number when > the correct bug number was 798423. Sorry! Without a matching backout in the same merge (to fix the bug number), the script we use will just mark bugs regardless; philor luckily happened to spot this.
Assignee: mcmanus → nobody
You need to log in before you can comment on or make changes to this bug.