339087 - Sync trunk code with 1.8.0 and 1.8 branches

Status: RESOLVED FIXED

(Core Graveyard :: XForms, defect)

Description

[Allan Beaufour]

Same trick as for 0.5, bascially just removing SOAP support.

(hmm, I've just done this bug once. Iid I type this into a wrong window the first time, or did bugzilla just swallow my bug....)

Comment 1

[Allan Beaufour]

Attachment: Patch for trunk code

Apply this to extensions/xforms (HEAD), and compile it on 1_8_0.

Comment 2

[Allan Beaufour]

Running through my test suite, crashes ff. Not on the same test case. Just suddenly.

nsProfileLock::FatalSignalHandler(int)+0x0000013B [dist/bin/firefox-bin +0x00024FFF]
__kernel_sigreturn+0x00000000 [ +0x00000420]
nsCOMPtr<nsIContent>::assign_from_qi(nsQueryInterface, nsID const&)+0x00000021 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x002CB129]
nsCOMPtr<nsIContent>::nsCOMPtr(nsQueryInterface)+0x00000031 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x002CB187]
nsCOMPtr<nsIContent>::Assert_NoQueryNeeded()+0x00000046 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x002CB1D6]
nsCOMPtr<nsIContent>::operator=(nsIContent*)+0x00000030 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x002F33D0]
nsDOMEventRTTearoff::LastRelease()+0x00000050 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x005D8218]
UNKNOWN [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x005D835D]
nsCOMPtr<nsIDOMEventTarget>::~nsCOMPtr()+0x00000050 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x003C1456]
nsXMLEventsListener::Unregister()+0x000000D0 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x006470CE]
UNKNOWN [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x0064711B]
nsBaseHashtable<nsISupportsHashKey, nsCOMPtr<nsXMLEventsListener>, nsXMLEventsListener*>::s_EnumStub(PLDHashTable*, PLDHashEntryHdr*, unsigned int, void*)+0x0000004D [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x0064900F]
PL_DHashTableEnumerate+0x00000086 [dist/bin/libxpcom_core.so +0x00076726]
nsBaseHashtable<nsISupportsHashKey, nsCOMPtr<nsXMLEventsListener>, nsXMLEventsListener*>::Enumerate(PLDHashOperator (*)(nsISupports*, nsCOMPtr<nsXMLEventsListener>&, void*), void*)+0x00000070 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x00648C52]
nsXMLEventsManager::DocumentWillBeDestroyed(nsIDocument*)+0x0000003E [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x00646FC8]
nsDocument::~nsDocument()+0x00000228 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x005AD5E8]
nsHTMLDocument::~nsHTMLDocument()+0x000002D0 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x006D7994]
UNKNOWN [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x005A7D54]
UNKNOWN [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x006D4E10]
nsCOMPtr<nsIDOMEventTarget>::~nsCOMPtr()+0x00000050 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x003C1456]
nsDOMEvent::~nsDOMEvent()+0x00000132 [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x0063AC0E]
UNKNOWN [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x00639500]
XPCJSRuntime::GCCallback(JSContext*, JSGCStatus)+0x0000094B [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libxpconnect.so +0x00068821]
UNKNOWN [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x007806F0]
js_GC+0x000013FA [dist/bin/libmozjs.so +0x00052C52]
js_ForceGC+0x0000005D [dist/bin/libmozjs.so +0x0005184B]
JS_GC+0x00000064 [dist/bin/libmozjs.so +0x00017E82]
UNKNOWN [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libgklayout.so +0x00780684]
nsTimerImpl::Fire()+0x00000291 [dist/bin/libxpcom_core.so +0x000FE1FD]
handleTimerEvent(TimerEventType*)+0x0000012E [dist/bin/libxpcom_core.so +0x000FE46A]
PL_HandleEvent+0x00000055 [dist/bin/libxpcom_core.so +0x000F5892]
PL_ProcessPendingEvents+0x000000D3 [dist/bin/libxpcom_core.so +0x000F571E]
UNKNOWN [dist/bin/libxpcom_core.so +0x000F89E1]
UNKNOWN [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libwidget_gtk2.so +0x0003503C]
UNKNOWN [/opt/gnome/lib/libglib-2.0.so.0 +0x000521DF]
g_main_context_dispatch+0x000001FC [/opt/gnome/lib/libglib-2.0.so.0 +0x0002834C]
UNKNOWN [/opt/gnome/lib/libglib-2.0.so.0 +0x0002B7BB]
g_main_loop_run+0x000001B7 [/opt/gnome/lib/libglib-2.0.so.0 +0x0002BAD7]
gtk_main+0x000000C1 [/opt/gnome/lib/libgtk-x11-2.0.so.0 +0x0012E861]
UNKNOWN [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libwidget_gtk2.so +0x00035AD8]
UNKNOWN [/store/mozilla/source/firefox15/mozilla/dist/bin/components/libtoolkitcomps.so +0x0004105D]
UNKNOWN [dist/bin/firefox-bin +0x0000FC77]
UNKNOWN [dist/bin/firefox-bin +0x000084A6]
__libc_start_main+0x000000D0 [/lib/tls/libc.so.6 +0x00014EA0]

Comment 3

[Allan Beaufour]

*** Bug 339084 has been marked as a duplicate of this bug. ***

Comment 4

[Allan Beaufour]

Attachment: Testcase (crasher)

(In reply to comment #2)
> Running through my test suite, crashes ff. Not on the same test case. Just
> suddenly.

I think I found the testcase. The problem was that it happens on document destruction so there were some delay... load this, exit the page, and wait.

Comment 5

[Allan Beaufour]

Attachment: Testcase (crasher)

Another crasher, same stack.

Repeat is involved again.

Comment 6

[Allan Beaufour]

There is some weirdness involved in this. I cannot get the testcases to crash when I load then from bugzilla, but the same testcases loaded through my testsuite (localhost) crashes.

Comment 7

[Allan Beaufour]

Attachment: Testcase (crasher)

Same stack. And repeat again.

Comment 8

[Allan Beaufour]

On the bright side: Except for these crashers (which all seems to be related with repeats somehow), we seem to run just fine on 1.8.0

Comment 9

[Allan Beaufour]

Any event listener on an element (ev:event="...") inside repeat content possibly combined with some JS addEventListener() calls seems to trigger this.

Comment 10

[aaronr]

don't forget that we need schema-validation changes, too.  I think there are at least one or two that aren't on the branches.  And didn't you have an xpath change?  I guess that probably can't go in without approval, though, since it isn't our component.

Comment 11

[Allan Beaufour]

(In reply to comment #10)
> don't forget that we need schema-validation changes, too.  I think there are at
> least one or two that aren't on the branches.

I trust that Doron are on top of that.

> And didn't you have an xpath change?  I guess that probably can't go in without 
> approval, though, since it isn't our component.

Yes, the fix to instance(), but we need approval for that one (bug 338263)

Comment 12

[Allan Beaufour]

Bug 339104 fixed the crashers, and we seem to be running smooth on 1_8_0 now. Only  problem is bug 332803, which seems to still exist on 1_8_0.

Comment 13

[Doron Rosenberg (IBM)]

(In reply to comment #10)
> don't forget that we need schema-validation changes, too.  I think there are at
> least one or two that aren't on the branches.  And didn't you have an xpath
> change?  I guess that probably can't go in without approval, though, since it
> isn't our component.
> 

which schema changes?

Comment 14

[aaronr]

(In reply to comment #13)
> (In reply to comment #10)
> > don't forget that we need schema-validation changes, too.  I think there are at
> > least one or two that aren't on the branches.  And didn't you have an xpath
> > change?  I guess that probably can't go in without approval, though, since it
> > isn't our component.
> > 
> 
> which schema changes?
> 

bug 338005, bug 326530, and bug 331897 aren't in the branches yet, it doesn't look like.

Comment 15

[Allan Beaufour]

Attachment: Patch for 1_8_0

Here's a complete patch for 1_8_0 extensions/xforms. It includes the patch in bug 338451. So applying this makes xforms fly in 1_8_0 land :)

So basically we just need to land bug 338451, and fix the permission dialog (bug 339284).

Well, and the tree needs to open...

Comment 16

[Allan Beaufour]

Attachment: Patch for 1_8

Comment 17

[Allan Beaufour]

(In reply to comment #16)
> Created an attachment (id=223682) [edit]
> Patch for 1_8

Check in on MOZILLA_1_8_BRANCH.

Comment 18

[Allan Beaufour]

Attachment: Patch for 1.8.0

Here's the branch that I've just synced 1.8.0 with.

Comment 19

[Allan Beaufour]

Both 1_8_0 and 1_8 are up-to-date with trunk now.

Comment 20

[aaronr]

can we close this now, or is it waiting on bug 338451 to go to 1.8.1 and 1.8.0 branches?