492026 - Port |Bug 480069 - Recursive submakes without shell loops| to comm-central

Status: RESOLVED FIXED

(MailNews Core :: Build Config, defect)

Description

[Serge Gautherie (:sgautherie)]

Bug 480069 has 'wanted1.9.1-'.

I guess comm-central will want to (im)port that bug,
but I don't know if we should '#ifdef 1.9.1' yet or wait until comm-central branches?

Comment 1

[Mark Banner (:standard8)]

If its not hurting us in terms of being able to get things done and keep trunk running, then we can wait.

Comment 2

[Serge Gautherie (:sgautherie)]

Attachment: (Av1) Just copy it

Comment 3

[Justin Wood (:Callek)]

Comment on attachment 420036 [details] [diff] [review]
(Av1) Just copy it

r+ assuming this works on at least trunk ++ 1.9.2 [on at least 1 tested platform] still.

Did not test locally. Does look correct based on ported bug's changeset.

Comment 4

[Serge Gautherie (:sgautherie)]

http://build.mozillamessaging.com/buildbot/try/changes/294
{
http://tinderbox.mozilla.org/showlog.cgi?log=ThunderbirdTry/1262686294.1262690699.28127.gz&fulltext=1
Try server comm-central mac hg builder on 2010/01/05 02:11:34

Undefined symbols:
  "CrashReporter::SetServerURL(nsACString_internal const&)", referenced from:
[...]
ld: symbol(s) not found
collect2: ld returned 1 exit status
}

Trying a 2nd time...

Comment 5

[Serge Gautherie (:sgautherie)]

(In reply to comment #4)

http://build.mozillamessaging.com/buildbot/try/changes/295:
same error.
Mark, would you have a clue? (some wrong/missing dependency or something?)

Comment 6

[Justin Wood (:Callek)]

Comment on attachment 420036 [details] [diff] [review]
(Av1) Just copy it

Serge, given your issues with try, lets have Mark sr+ this before you check in.

Comment 7

[Serge Gautherie (:sgautherie)]

(In reply to comment #4)
> http://tinderbox.mozilla.org/showlog.cgi?log=ThunderbirdTry/1262686294.1262690699.28127.gz&fulltext=1
> Try server comm-central mac hg builder on 2010/01/05 02:11:34

This build had bug 538090 !
This patch should pass now: let's try...

Comment 8

[Mark Banner (:standard8)]

Why should this block 3.1? There's nothing here that is affecting functions is there?

Comment 9

[Serge Gautherie (:sgautherie)]

(In reply to comment #8)
> Why should this block 3.1?

Sorry, I didn't mean to set that flag.

Comment 10

[Serge Gautherie (:sgautherie)]

(In reply to comment #7)
> This build had bug 538090 !
> This patch should pass now: let's try...

http://tinderbox.mozilla.org/showlog.cgi?log=ThunderbirdTry/1262770830.1262777187.5399.gz&fulltext=1
Try server comm-central mac hg builder on 2010/01/06 01:40:30
Bug 538090 is fixed, but this patch still fails :-(

To be explicit, the only hint here is that it's MacOSX specific...

Comment 11

[Serge Gautherie (:sgautherie)]

Attachment: (Av1a) Just copy it [Checkin: Comment 21]

Av1, (context) un-bitrotted.

I'm submitting it to Try again, in hope something else fixed the issue in the meantime...

Comment 12

[Serge Gautherie (:sgautherie)]

(In reply to comment #11)

Passed for TB:
http://build.mozillamessaging.com/buildbot/try/builders/Try%20server%20comm-central%20linux%20hg%20builder/builds/349
http://build.mozillamessaging.com/buildbot/try/builders/Try%20server%20comm-central%20mac%20hg%20builder/builds/351
http://build.mozillamessaging.com/buildbot/try/builders/Try%20server%20comm-central%20win32%20hg%20builder/builds/348

Failed for SM/Mac, but different error (which may be random):
http://build.mozillamessaging.com/buildbot/try/builders/Try%20server%20comm-central%20linux%20hg%20builder/builds/348
http://build.mozillamessaging.com/buildbot/try/builders/Try%20server%20comm-central%20mac%20hg%20builder/builds/350
{
command timed out: 3600 seconds without output, killing pid 38517

instead of

ranlib libgklayout.a
}
http://build.mozillamessaging.com/buildbot/try/builders/Try%20server%20comm-central%20win32%20hg%20builder/builds/347

Trying again...

Comment 13

[Serge Gautherie (:sgautherie)]

(In reply to comment #12)

This one proceeded much further:
{
http://build.mozillamessaging.com/buildbot/try/builders/Try%20server%20comm-central%20mac%20hg%20builder/builds/352

Undefined symbols:
  "nsAddbookUrl::nsAddbookUrl()", referenced from:
      nsAddbookUrlConstructor(nsISupports*, nsID const&, void**)in libmail.a(nsMailModule.o)
ld: symbol(s) not found
}

Am I out of luck?
stefanh, standard8, could you try it locally? Thanks.

Comment 14

[Serge Gautherie (:sgautherie)]

{
make[2]: MAKE@: Command not found
make[2]: *** [default] Error 127
}

http://hg.mozilla.org/comm-central/rev/b80103b363ea
(Bv1) configure.in: restore 'AC_PATH_PROGS(MAKE, ...' removed too soon by bug 485405.

Comment 15

[Justin Wood (:Callek)]

(In reply to comment #14)

replied on Bug 485405. bustage fix worked, but there is no need for a block on that fix for this bug.

Comment 16

[Serge Gautherie (:sgautherie)]

(In reply to comment #15)

> replied on Bug 485405. bustage fix worked,

Well, I commented here because I didn't want to bother that bug more than required:
the port happened there, but I think c-c issues should be handled separately.

> but there is no need for a block on that fix for this bug.

In reply to your comment there:
I figured out |MAKE = @MAKE@| was involved, which lead me here, where that line will be removed;
I didn't want to mix/mess with bug fixing and tracking so a partial backout looked much easier and surer (to fix red asap).
I might have looked into this matter at review time, but now, unless the restored block causes an issue, I'm in favor of just letting it be until this very bug lands.

Comment 17

[Justin Wood (:Callek)]

Serge, since Mark is certainly swamped can you do a quick double-check on this, that its both not bitrotted and would pass a try run still? (if its bitrotted please attach a new patch) Thanks

Comment 18

[Mark Banner (:standard8)]

Comment on attachment 426452 [details] [diff] [review]
(Av1a) Just copy it
[Checkin: Comment 21]

I've just got around to trying this. There's one minor bit of bitrot, but that's about it.

A clobbered debug build worked fine on my machine.

Therefore I propose that we land this and keep an eye out for unexpected build failures - if we're getting them then we can assess if we need to backout and rework or otherwise.

Comment 19

[Justin Wood (:Callek)]

serge, I'd love if you could land this; the fact that we have MAKE = @MAKE@ here prevents me from trying pymake on a c-c build (recurses using GNUMake)

Comment 20

[Serge Gautherie (:sgautherie)]

I'm sending it to Try first...

Comment 21

[Serge Gautherie (:sgautherie)]

Comment on attachment 426452 [details] [diff] [review]
(Av1a) Just copy it
[Checkin: Comment 21]


(In reply to comment #20)
> I'm sending it to Try first...

Now, passes for SM and TB :-)

http://hg.mozilla.org/comm-central/rev/895b22a261fe

Comment 22

[Serge Gautherie (:sgautherie)]

(In reply to comment #14)
> http://hg.mozilla.org/comm-central/rev/b80103b363ea
> (Bv1) configure.in: restore 'AC_PATH_PROGS(MAKE, ...' removed too soon by bug
> 485405.

http://hg.mozilla.org/comm-central/rev/dc9bbe908fae
(Cv1) configure.in: remove now unneeded 'AC_PATH_PROGS(MAKE, ...'.