494367 - Maemo l10n repacks busted with echo None

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Axel Hecht [:Pike]]

The l10n maemo builds are currently hung with a SetProperties echo None, which waits for stdin.

That's the step right after "clean old builds", I think it's the echo tree one, and the tree property isn't set on the maemo builds.

Coop?

Comment 1

[Axel Hecht [:Pike]]

KaiRo mentions on IRC that 
self.tinderboxPrint('toolkit_revision',WithProperties('%(toolkit_revision)s'))
in the fennec repack is likely going to break, too. That should be a
self.tinderboxPrint('gecko_revision',WithProperties('%(gecko_revision)s'))

Sorry, that one is a bug in my initial comment in bug 488099.

Comment 2

[Axel Hecht [:Pike]]

Attachment: map what we did for fx to fennec to have the tree in mobile_config.py

Coop, I just hacked up a fix for mobile2, the same patch should apply to staging more or less, I guess.

Comment 3

[Axel Hecht [:Pike]]

Attachment: print gecko instead of toolkit

This changes the print to gecko_revision, which matches what we hand out in http://hg.mozilla.org/mobile-browser/file/default/locales/Makefile.in#l93

Comment 4

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

Comment on attachment 379128 [details] [diff] [review]
print gecko instead of toolkit

changeset:   311:dcc31f2ad425

Comment 5

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

Comment on attachment 379127 [details] [diff] [review]
map what we did for fx to fennec to have the tree in mobile_config.py

changeset:   1163:9edebdbbb8ee

Comment 6

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

FYI, I'm holding off on a reconfig until later today post-freeze (hopefully).

Comment 7

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

I lied, reconfig just happened so we don't lose any more slaves.

Comment 8

[Chris AtLee [:catlee]]

This is upstream bug http://buildbot.net/trac/ticket/569, and only happens when the slaves are using twisted 2.4.0.

Comment 9

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

This change seems to have taken.

Do we want to change the default from None to "notset" or an empty string to be more robust here?

Comment 10

[Axel Hecht [:Pike]]

(In reply to comment #9)
> Do we want to change the default from None to "notset" or an empty string to be
> more robust here?

Yeah, I think so.

Comment 11

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

Fixed by attachment 379571 [details] [diff] [review] in bug 488099.