728930 - Move SeaMonkey mozconfigs into source tree

Status: RESOLVED FIXED

(SeaMonkey :: Release Engineering, defect)

Description

[Justin Wood (:Callek)]

We should move our mozconfigs into the source tree like Firefox has done.

Will make it much clearer on when something changes, and when it has not.

Comment 1

[Justin Wood (:Callek)]

Attachment: [buildbot-configs] Part 1, v1 [Checked in: Comment 18]

the /dep/ changes are just to match the /nightly/ changes, and should be fine.

I *believe* (but will verify) that the absense of the src configs will not break any builders though.

Comment 2

[Justin Wood (:Callek)]

Attachment: dump-master output from buildbot-configs part 1

This is just a helper to understand *what* changes by doing this. It is a diff of the dump-master output both before and after the change.

Comment 3

[Justin Wood (:Callek)]

Attachment: [buildbotcustom] Part 2, v1 [Checked in: Comment 19]

This officially switches all dep builds to use the same mozconfig as "nightly"

Comment 4

[Justin Wood (:Callek)]

Attachment: [buildbot-configs] Part 3, v1

This patch drops the _dep mozconfigs which are no longer used (see Part 1 for what changed in mozconfigs by doing so)

Comment 5

[Justin Wood (:Callek)]

Attachment: [buildbot-configs] Part 4, v1 [Checked in: Comment 22]

This makes use of update_channel and adjust release Configs to use srcmozconfig where applicable. Currently the l10n repack code that we use does not support srcMozconfig, so that will be a separate bug though

Comment 6

[Justin Wood (:Callek)]

Attachment: [buildbot-configs] Part 3, v1.5 [Checked in: Comment 21]

Slight changes here based on checkconfig catching!

Comment 7

[Justin Wood (:Callek)]

Attachment: [comm-central] Part 5, copy mozconfigs from buildbot-configs directly. (Not Release Configs yet) [Checked in: Comment 16 & 32]

This is just a straight copy of the various mozconfigs from our real builds, I did not include release mozconfigs in this batch, I'll do those last (since they are not required for much of this bug)

Comment 8

[Justin Wood (:Callek)]

Attachment: [comm-central] Part 6, make use of MOZ_UPDATE_CHANNEL so that other trains work right [Checked in: Comment 17 & 32]

Comment 9

[Robert Kaiser]

Comment on attachment 598930 [details] [diff] [review]
[buildbot-configs] Part 1, v1
[Checked in: Comment 18]

Not sure if it makes too much sense to patch comm-1.9.1 and comm-2.0 configs inb this patch, could just delete them...

Comment 10

[Robert Kaiser]

Comment on attachment 598936 [details] [diff] [review]
dump-master output from buildbot-configs part 1

*Looks* good though I'm not really sure what dump-master is...

Comment 11

[Robert Kaiser]

Comment on attachment 599023 [details] [diff] [review]
[buildbot-configs] Part 3, v1.5
[Checked in: Comment 21]

And here you are deleting the mozconfigs anyhow, so I wonder if patching them first makes too much sense, it's not like we can track them across repos anyhow.
Also, can we do this across branches yet, don't we need to check in the mozconfigs on all branches before?

Comment 12

[Robert Kaiser]

Comment on attachment 599295 [details] [diff] [review]
[comm-central] Part 6, make use of MOZ_UPDATE_CHANNEL so that other trains work right
[Checked in: Comment 17 & 32]

Could do this right away, but might make sense.

Comment 13

[Justin Wood (:Callek)]

(In reply to Robert Kaiser (:[e-mail]) from comment #10)
> Comment on attachment 598936 [details] [diff] [review]
> dump-master output from buildbot-configs part 1
> 
> *Looks* good though I'm not really sure what dump-master is...

dump-master is a script in the braindump repo that prints out info from all the Buidlers, steps, and schedulers etc. So that you can better assess what a config/buildbotcustom change is actually changing.

By using it you can see that in this case we switch from always wget'ing the configs from buildbot-configs we test if they are present in our source tree...

(In reply to Robert Kaiser (:[e-mail]) from comment #11)
> Comment on attachment 599023 [details] [diff] [review]
> [buildbot-configs] Part 3, v1.5
> 
> And here you are deleting the mozconfigs anyhow, so I wonder if patching
> them first makes too much sense, it's not like we can track them across
> repos anyhow.

Sure, the patching first was `mostly` just a "lets be sure we're not changing stuff that we don't expect when we switch.

> Also, can we do this across branches yet, don't we need to check in the
> mozconfigs on all branches before?

Yea, the buildbot-configs configs need to be on the |default| branch, but I merged both ways this past week, so will land on default and just merge to seamonkey-production.

Comment 14

[Philip Chee]

Comment on attachment 599292 [details] [diff] [review]
[comm-central] Part 5, copy mozconfigs from buildbot-configs directly. (Not Release Configs yet)
[Checked in: Comment 16 & 32]

a=me for comm-aurora.

Callek says over IRC "Because they enable srcdir mozconfigs properly, such that we can enable msvc2010 soonish much much more easily"

Comment 15

[Philip Chee]

Comment on attachment 599295 [details] [diff] [review]
[comm-central] Part 6, make use of MOZ_UPDATE_CHANNEL so that other trains work right
[Checked in: Comment 17 & 32]

a=me for comm-aurora.

Callek says over IRC "Because they enable srcdir mozconfigs properly, such that we can enable msvc2010 soonish much much more easily"

Comment 16

[Serge Gautherie (:sgautherie)]

Comment on attachment 599292 [details] [diff] [review]
[comm-central] Part 5, copy mozconfigs from buildbot-configs directly. (Not Release Configs yet)
[Checked in: Comment 16 & 32]

http://hg.mozilla.org/comm-central/rev/6e7659bcca95

Comment 17

[Serge Gautherie (:sgautherie)]

Comment on attachment 599295 [details] [diff] [review]
[comm-central] Part 6, make use of MOZ_UPDATE_CHANNEL so that other trains work right
[Checked in: Comment 17 & 32]

http://hg.mozilla.org/comm-central/rev/f3aee29e61df

Comment 18

[Serge Gautherie (:sgautherie)]

Comment on attachment 598930 [details] [diff] [review]
[buildbot-configs] Part 1, v1
[Checked in: Comment 18]

http://hg.mozilla.org/build/buildbot-configs/rev/9c7c078d58b5

Comment 19

[Serge Gautherie (:sgautherie)]

Comment on attachment 598945 [details] [diff] [review]
[buildbotcustom] Part 2, v1
[Checked in: Comment 19]

http://hg.mozilla.org/build/buildbotcustom/rev/7641faab0337

Comment 20

[Serge Gautherie (:sgautherie)]

Comment on attachment 599023 [details] [diff] [review]
[buildbot-configs] Part 3, v1.5
[Checked in: Comment 21]

http://hg.mozilla.org/build/buildbotcustom/rev/12c68b23e307

Comment 21

[Serge Gautherie (:sgautherie)]

Comment on attachment 599023 [details] [diff] [review]
[buildbot-configs] Part 3, v1.5
[Checked in: Comment 21]

(In reply to Serge Gautherie (:sgautherie) from comment #20)
> http://hg.mozilla.org/build/buildbotcustom/rev/12c68b23e307

http://hg.mozilla.org/build/buildbot-configs/rev/12c68b23e307 !

Comment 22

[Serge Gautherie (:sgautherie)]

Comment on attachment 599016 [details] [diff] [review]
[buildbot-configs] Part 4, v1
[Checked in: Comment 22]

http://hg.mozilla.org/build/buildbot-configs/rev/fcd16841e442

Comment 23

[Serge Gautherie (:sgautherie)]

Attachment: (AAv1) Remove overridden --enable-jemalloc, in-tree [Checked in: Comment 25 & 32]

Part of bug 603153 fix.

Comment 24

[Alex Keybl [:akeybl]]

Comment on attachment 611410 [details] [diff] [review]
(AAv1) Remove overridden --enable-jemalloc, in-tree
[Checked in: Comment 25 & 32]

Please re-nominate once r+'d and ready.

Comment 25

[Serge Gautherie (:sgautherie)]

Comment on attachment 611410 [details] [diff] [review]
(AAv1) Remove overridden --enable-jemalloc, in-tree
[Checked in: Comment 25 & 32]

http://hg.mozilla.org/comm-central/rev/6f65dcd99c1a

Comment 26

[Serge Gautherie (:sgautherie)]

Attachment: (ABv1) Remove overridden --enable-jemalloc, in buildbot-configs tree (except for SeaMonkey 2.0) [Checked in: Comment 28]

http://mxr.mozilla.org/build/search?string=jemalloc&case=1&find=%2Fseamonkey%2F.*%2Fdebug%2F
"Found 5 matching lines in 5 files"

http://mxr.mozilla.org/build/source/buildbot-configs/seamonkey/win32/comm-2.0/debug/mozconfig
is left alone as it is obsolete.

Comment 27

[Justin Wood (:Callek)]

Comment on attachment 613995 [details] [diff] [review]
(ABv1) Remove overridden --enable-jemalloc, in buildbot-configs tree (except for SeaMonkey 2.0)
[Checked in: Comment 28]

Review of attachment 613995 [details] [diff] [review]:
-----------------------------------------------------------------

Please land on _default_ branch

Comment 28

[Serge Gautherie (:sgautherie)]

Comment on attachment 613995 [details] [diff] [review]
(ABv1) Remove overridden --enable-jemalloc, in buildbot-configs tree (except for SeaMonkey 2.0)
[Checked in: Comment 28]

http://hg.mozilla.org/build/buildbot-configs/rev/7acf96676562

Comment 29

[Jens Hatlak (:InvisibleSmiley)]

Serge, changeset 6f65dcd99c1a (cf. whiteboard) alone cannot be checked into c-a because the file it modifies does not exist yet. Please clarify.

Comment 30

[Serge Gautherie (:sgautherie)]

(In reply to Jens Hatlak (:InvisibleSmiley) from comment #29)
> Serge, changeset 6f65dcd99c1a (cf. whiteboard) alone cannot be checked into
> c-a because the file it modifies does not exist yet. Please clarify.

Oh, right:
Callek, do you want to land your parts 5 and 6 on aurora/2.10 now? Or what's the plan?

Comment 31

[Justin Wood (:Callek)]

Yes 5 and 6 should land on aurora, I actually *thought* I already did... but unless Jens beats me to it I'll land these 3 patches myself within the next day or two

Comment 32

[Justin Wood (:Callek)]

http://hg.mozilla.org/releases/comm-aurora/rev/d3d5803a732f
http://hg.mozilla.org/releases/comm-aurora/rev/f128b236d7b0
http://hg.mozilla.org/releases/comm-aurora/rev/604396ff867c

We'll need to do work for the release configs in-tree for 2.10 beta, but I'll mark this fixed for now anyway.