Closed
Bug 916604
Opened 11 years ago
Closed 8 years ago
DocShell assertion when testing app redirects
Categories
(DevTools Graveyard :: WebIDE, defect, P3)
DevTools Graveyard
WebIDE
Tracking
(Not tracked)
RESOLVED
INVALID
People
(Reporter: ochameau, Unassigned)
References
Details
(Whiteboard: [btpp-backlog])
Attachments
(1 file, 1 obsolete file)
2.34 KB,
patch
|
Details | Diff | Splinter Review |
I've been writing a mochitest to cover app redirect for multiple weeks now, and this test used to pass but has started triggering an assertion recently:
mLoadType != LOAD_ERROR_PAGE && mLoadType != LOAD_BYPASS_HISTORY (Do not add error or bypass pages to global history)
http://mxr.mozilla.org/mozilla-central/source/docshell/base/nsDocShell.cpp#11461
When creating an mozapp iframe whose app has a special redirects rule in its manifest, it triggers this exception when loading a document that is redirected and whose redirection is catched by the app redirect rule and is supposed to be redirected to a local app document.
See test script in attachment.
Reporter | ||
Comment 1•11 years ago
|
||
Reporter | ||
Comment 2•11 years ago
|
||
Fabrice, you wrote the original code in docshell.cpp for app redirects, does that assertion make any sense to you? Otherwise, could we find some docshell/history experts to help us figure this out?
Also, do we have any unit test for app redirects? If not, it may have been broken recently...
Flags: needinfo?(fabrice)
Comment 3•11 years ago
|
||
I can reproduce the failure with a test app here also, and will start bisecting. I'm gonna check with QA also since the facebook import relies on this and that was not flagged as broken recently.
Flags: needinfo?(fabrice)
Comment 4•11 years ago
|
||
Fabrice mentions this affects fb import, which means this isn't a devtools specific bug - it's a core bug.
blocking-b2g: --- → koi?
Component: Developer Tools: App Manager → DOM
Keywords: regression,
regressionwindow-wanted
Product: Firefox → Core
Version: unspecified → Trunk
Comment 5•11 years ago
|
||
I checked with tchung and on my own build, but FB import is not affected. Still investigating though.
Comment 6•11 years ago
|
||
And I can't reproduce either with another oauth api (dropbox).
Comment 7•11 years ago
|
||
(In reply to Fabrice Desré [:fabrice] from comment #3)
> I can reproduce the failure with a test app here also, and will start
> bisecting. I'm gonna check with QA also since the facebook import relies on
> this and that was not flagged as broken recently.
Can you attach the test app? I can get someone to do a regression window here if I can define clear STR with the test app to do the regression window against.
Comment 8•11 years ago
|
||
(In reply to Jason Smith [:jsmith] from comment #7)
> (In reply to Fabrice Desré [:fabrice] from comment #3)
> > I can reproduce the failure with a test app here also, and will start
> > bisecting. I'm gonna check with QA also since the facebook import relies on
> > this and that was not flagged as broken recently.
>
> Can you attach the test app? I can get someone to do a regression window
> here if I can define clear STR with the test app to do the regression window
> against.
That was actually a network error on the service side. I can't repro for now, so I'll check with ochameau if we can bisect the test failure instead.
Reporter | ||
Updated•11 years ago
|
blocking-b2g: koi? → ---
Component: DOM → Developer Tools: App Manager
Product: Core → Firefox
Reporter | ||
Comment 9•11 years ago
|
||
I reclassified this bug and removed koi? as it looks like a test specific issue that isn't landed yet because of this assertion. So that this bug doesn't seem that serious at the end. The redirects features has been confirmed to work fine.
I'll try to dig into this assertion when I get some free cycle.
Updated•11 years ago
|
Priority: -- → P2
Updated•11 years ago
|
Priority: P2 → P1
Comment 10•9 years ago
|
||
Is there still an issue here or can we close this out?
Flags: needinfo?(poirot.alex)
Reporter | ||
Comment 11•9 years ago
|
||
Reporter | ||
Comment 12•9 years ago
|
||
Rebased, with a new try, to see if the assertion still happens.
Attachment #805049 -
Attachment is obsolete: true
Reporter | ||
Comment 13•9 years ago
|
||
Reporter | ||
Comment 14•9 years ago
|
||
There is no more assertion with this test, but it doesn't work anymore and I don't know why.
It would still be great to cover this feature with a test... especially since the test already somewhat exists.
Flags: needinfo?(poirot.alex)
Comment 15•9 years ago
|
||
I'll leave it to you to mutate the bug accordingly in that respect, thanks for the update.
Keywords: regression,
regressionwindow-wanted
Reporter | ||
Comment 16•9 years ago
|
||
Reporter | ||
Comment 17•9 years ago
|
||
Reporter | ||
Comment 18•9 years ago
|
||
Priority: P1 → P3
Whiteboard: [btpp-backlog]
Apps are gone now.
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → INVALID
Updated•6 years ago
|
Product: Firefox → DevTools
Updated•5 years ago
|
Product: DevTools → DevTools Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•