1130812 - AppConstants.jsm is not substituting correctly

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: General, defect)

Description

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

Attachment: fix-appconstants v0.1

We have a spurious ';' which stops the text from being substituted. We also have a duplicate MOZ_DEVICES property.

Comment 1

[Richard Newman [:rnewman]]

Comment on attachment 8560951 [details] [diff] [review]
fix-appconstants v0.1

Review of attachment 8560951 [details] [diff] [review]:
-----------------------------------------------------------------

FFFFFFFF

Comment 2

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

https://hg.mozilla.org/integration/fx-team/rev/04faeec9c9f0

Comment 3

[Richard Newman [:rnewman]]

Nick, can we get a tiny smoke test for this? E.g., checking that build dates are within sane ranges, that sort of thing?

Comment 4

[Nick Alexander :nalexander [he/him]]

(In reply to Richard Newman [:rnewman] from comment #3)
> Nick, can we get a tiny smoke test for this? E.g., checking that build dates
> are within sane ranges, that sort of thing?

We certainly can: I've filed Bug 1130872 as a mentor bug for this.  I have often observed JS errors in browser.js, or suspected them.  Is there a reasonable way to communicate to Java that browser.js has finished loading?  It's hard to "message from the end of the script" because one can never be sure that the script has really finished parsing and executing...

Comment 5

[Richard Newman [:rnewman]]

(In reply to Nick Alexander :nalexander from comment #4)

> We certainly can: I've filed Bug 1130872 as a mentor bug for this.  I have
> often observed JS errors in browser.js, or suspected them.  Is there a
> reasonable way to communicate to Java that browser.js has finished loading? 
> It's hard to "message from the end of the script" because one can never be
> sure that the script has really finished parsing and executing...

Gecko:Ready might be close enough; depends on your definition of "loaded".

Certainly that event would indicate that BrowserApp, BrowserEventHandler, all the helper classes listed around line 469 of browser.js, etc. are loaded and working to a point.

Comment 6

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/04faeec9c9f0

Comment 7

[Nick Alexander :nalexander [he/him]]

(In reply to Richard Newman [:rnewman] from comment #5)
> (In reply to Nick Alexander :nalexander from comment #4)
> 
> > We certainly can: I've filed Bug 1130872 as a mentor bug for this.  I have
> > often observed JS errors in browser.js, or suspected them.  Is there a
> > reasonable way to communicate to Java that browser.js has finished loading? 
> > It's hard to "message from the end of the script" because one can never be
> > sure that the script has really finished parsing and executing...
> 
> Gecko:Ready might be close enough; depends on your definition of "loaded".
> 
> Certainly that event would indicate that BrowserApp, BrowserEventHandler,
> all the helper classes listed around line 469 of browser.js, etc. are loaded
> and working to a point.

We already block on Gecko:Ready in pretty much every test, so no sense piling on.  I just feel like browser.js evaluation can fail in ways that look like the evaluation has stopped.  Anyway, ticket filed for AppConstants.jsm test.