538021 - clean up bogus entries in our confvars.sh

Status: RESOLVED DUPLICATE of bug 608742

(Firefox for Android Graveyard :: General, defect)

Description

[:Gavin Sharp [email: [e-mail]]]

Our confvars.sh has entries that are copied from the default xulrunner confvars.sh (http://mxr.mozilla.org/mozilla-central/source/xulrunner/confvars.sh ), and most are bogus as far as I can tell. Patch in a sec.

Comment 1

[:Gavin Sharp [email: [e-mail]]]

>-MOZ_XUL_APP=1

Killed in https://hg.mozilla.org/mozilla-central/rev/9b921a3fdb5e , but even before that I suspect it had no effect.

>-MOZ_UPDATER=0

Has no effect, as bug 530393 discovered. I'm not sure why exactly... I suspect it's because we're doing a two-step build process, and this only gets set in the mobile phase (while the updater is built in the xulrunner phase).

>-MOZ_ENABLE_LIBXUL=1

Same deal.

>-MOZ_NO_XPCOM_OBSOLETE=1

Removed in http://hg.mozilla.org/mozilla-central/rev/d578481cbc37 .

>-MOZ_XULRUNNER=1

This makes no sense, it's supposed to signify that we're building xulrunner, but since it's only checked in m-c/xulrunner, it has no effect.

Comment 2

[:Gavin Sharp [email: [e-mail]]]

Attachment: patch

Comment 3

[Mark Finkle (:mfinkle) (use needinfo?)]

Comment on attachment 420181 [details] [diff] [review]
patch

Oh the joy of copy paste project creation. Thanks for the cleanup.

Comment 4

[:Gavin Sharp [email: [e-mail]]]

Hmm, looks like this was forgotten! I'd check it in but I'm not entirely sure what the state of m-b is at the moment...

Comment 5

[:Gavin Sharp [email: [e-mail]]]

Though http://hg.mozilla.org/mobile-browser/rev/7227b5b43ebe complicates this a bit... I think it means that the entries in our confvars.sh can now actually influence the "xulrunner" portion of the single-build, when we build that way.

Comment 6

[:Gavin Sharp [email: [e-mail]]]

Attachment: updated patch

I'm reasonably confident that these changes have no effect. I'm less sure about removing MOZ_MORK=, since its default is enabled, so it may actually have an effect once bug 557027 is landed and we switch over.

That might also mean that we can disable the updater for Maemo in confvars.sh, too, rather than relying on the entry in releng's mozconfig.

Comment 7

[Mark Finkle (:mfinkle) (use needinfo?)]

patch in bug 608742 does some cleanup, but i don't think we want to remove MOZ_UPDATER and I notice the Firefox still uses MOZ_ENABLE_LIBXUL.

MOZ_XUL_APP and MOZ_NO_XPCOM_OBSOLETE are removed.

Duping to bug 608742