743606 - "missing talos data" script needs to be told that little trace_malloc_*s come from debug builds

Status: RESOLVED FIXED

(Release Engineering :: General, defect, P4)

Description

[Phil Ringnalda (:philor)]

"Firefox MacOSX 10.6.2 (rev3) hasn't reported trace_malloc_maxheap since 2012-04-07 10:37:54 (last build on 2012-04-08 12:00:03)"
"Firefox MacOSX 10.6.2 (rev3) hasn't reported trace_malloc_allocs since 2012-04-07 10:37:54 (last build on 2012-04-08 12:00:03)"
"Firefox MacOSX 10.6.2 (rev3) hasn't reported trace_malloc_leaks since 2012-04-07 10:37:54 (last build on 2012-04-08 12:00:03)"

Nope, last 10.6.2 debug build was on 2012-04-07 10:37:54. Opt doesn't count, nightlies don't count, on platforms that do them periodic PGO doesn't count.

"Firefox MacOSX 10.6.2 (rev3) hasn't reported codesighs since 2012-04-07 10:37:54 (last build on 2012-04-08 12:00:03)"

Nope, last non-nightly, because we don't do codesighs on nightlies, was on 2012-04-07 10:37:54

That "12:00:03" is a bit suspicious, too, since no Mac builds started then - is it looking at the last build of any sort on any platform as a guide to when it should have last gotten platform-specific and opt/debug-specific data?

Comment 2

[Chris Cooper [:coop] (he/him)]

Is this an issue for rev5 machines too, i.e. do we even track this data for rev5 minis yet?

Comment 3

[Phil Ringnalda (:philor)]

Dunno, but "treats any build on any platform rather than debug build on the same platform as last build" is not Mac-specific, I just picked Mac from the top of the message instead of

Firefox CentOS (x86_64) release 5 (Final) hasn't reported trace_malloc_leaks since 2012-04-07 10:37:54 (last build on 2012-04-08 12:00:03)
Firefox CentOS release 5 (Final) hasn't reported trace_malloc_maxheap since 2012-04-07 10:37:54 (last build on 2012-04-08 12:00:03)
Firefox CentOS release 5 (Final) hasn't reported trace_malloc_leaks since 2012-04-07 10:37:54 (last build on 2012-04-08 12:00:03)

from the bottom.

Comment 4

[Chris AtLee [:catlee]]

I hate this stupid script.

Anyway, I've made it ignore trace_malloc_* and codesighs.