1297198 - platform.ini incorrectly points at c-c revisions not m-c revisions

Status: RESOLVED DUPLICATE of bug 1298719

(Thunderbird :: Build Config, defect)

Description

[aleth [:aleth]]

Fallout from Bug 1247162. See bug 1267270 comment 24 and following.

While this is obviously wrong, it's currently unclear if it actually breaks anything.

Comment 1

[Tony Mechelynck [:tonymec]]

(In reply to aleth [:aleth] from comment #0)
> Fallout from Bug 1247162. See bug 1267270 comment 24 and following.
> 
> While this is obviously wrong, it's currently unclear if it actually breaks
> anything.

IIUC, the app works just the same with or without this bug, _except_ that if a bug is found that comes from Core or Toolkit the user or even the developer, no matter how experienced, will be unable to find out the m-c CSID range from a regression window, which will be expressed exclusively as Build IDs and/or c-c CSIDs — unless maybe the build logs for the "last good" and  "first bad" can be found, and I don't expect build logs to be kept long enough for that.

Comment 3

[aleth [:aleth]]

I suspect it also breaks source packaging.

Comment 4

[Edmund Wong (:ewong)]

(In reply to aleth [:aleth] from comment #0)
> Fallout from Bug 1247162. See bug 1267270 comment 24 and following.
> 
> While this is obviously wrong, it's currently unclear if it actually breaks
> anything.

Sure, it breaks the l10n nightly builds since it now expects to update
the mozilla-central source to the comm-central cset.  (re: bug 1304285
for SeaMonkey)

Comment 5

[Edmund Wong (:ewong)]

sorry for the confirmation of what you know already... my head's just
spinning on the number of bustages for l10n alone..

Comment 6

[aleth [:aleth]]

(In reply to Edmund Wong (:ewong) from comment #5)
> sorry for the confirmation of what you know already... my head's just
> spinning on the number of bustages for l10n alone..

Thanks, it's good to collect all known instances where this makes problems here (that's what this bug is for), since fixing it will probably require making an argument for m-c changes that are not required for Fx.

Re l10n, definitely... c-c and c-a are still running into the tooltool issue in bug 1267523 on Linux/Win, despite the disable-compile-environment.

Comment 7

[Edmund Wong (:ewong)]

would anyone object if I duped this bug against bug 1298719 and
changed that bug to MailNews::BuildConfig?