Closed Bug 580360 Opened 14 years ago Closed 14 years ago

aboutRedirector broken - about pages like privatebrowsing, sessionrestore, robots all busted

Categories

(Firefox :: General, defect)

x86
macOS
defect
Not set
normal

Tracking

()

VERIFIED FIXED
Firefox 4.0b2
Tracking Status
blocking2.0 --- beta2+

People

(Reporter: aaronmt, Unassigned)

References

Details

(Keywords: regression, Whiteboard: [4b2])

Currently on the latest trunk nightly, about:privatebrowsing has been removed and thus when entering PB mode, an alert pops up with 'The URL is not valid and cannot be loaded'.

STR: Enter PB mode

Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:2.0b2pre) Gecko/20100720 Minefield/4.0b2pre
Confirmed for me on a nightly build, though gavin isn't seeing it on his own (up to date) build. Packaging issue?
chrome://browser/content/aboutPrivateBrowsing.xhtml still works, so it looks like the component is missing?
about:sessionrestore and about:robots also broken - looks like aboutRedirector is unhappy with us.  :(

Updating summary.
Summary: Private browsing entry/welcome page removed in trunk? → aboutRedirector broken - about pages like privatebrowsing, sessionrestore, robots all busted
blocking2.0: --- → ?
Whiteboard: [4b2]
That means safebrowsing is busted, too. :(

On the other hand, can't reproduce this on Vista - might be mac specific?
This was caused by landing bug 578751, without review even! components/*.dylib refuse to load because @loader_path/libxpcom.dylib isn't present, when it should really be @loader_path/../libxpcom.dylib

We'll need to at least back that out, and perhaps bug 557225.
Yep, bisecting on tinderbox builds gives us this range:

http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=274a672f9acf&tochange=0c116ba35956

Who's on the backout?
I pushed a backout http://hg.mozilla.org/mozilla-central/rev/2a32e764c0ca and a candidate fix for the prior build bustage http://hg.mozilla.org/mozilla-central/rev/f4e545ef44dc

If this fails for any reason, we'll back out bug 557225, I left instructions in a release-drivers thread.
Component: Private Browsing → General
QA Contact: private.browsing → general
Don't have we tests to make sure that those pages are getting loaded correctly?
Flags: in-testsuite?
We do, but I think the test suite is setting DYLD_LIBRARY_PATH in the environment, which isn't how it runs out of the bundle, so the testsuite was passing incorrectly.
Status: NEW → RESOLVED
Closed: 14 years ago
Resolution: --- → FIXED
(In reply to comment #11)
> We do, but I think the test suite is setting DYLD_LIBRARY_PATH in the
> environment, which isn't how it runs out of the bundle, so the testsuite was
> passing incorrectly.

Worth fixing/changing the test so that we catch this sort of thing in the future?
Verified fixed with Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:2.0b3pre) Gecko/20100721 Minefield/4.0b3pre
Status: RESOLVED → VERIFIED
Target Milestone: --- → Firefox 4.0b2
blocking2.0: ? → beta2+
You need to log in before you can comment on or make changes to this bug.