668944 - URLs don't trigger the actions when clicked on

Status: RESOLVED INCOMPLETE

(Thunderbird :: General, defect)

Description

[Yuri]

User Agent: Mozilla/5.0 (X11; FreeBSD amd64; rv:5.0) Gecko/20100101 Firefox/5.0
Build ID: 20110701024755

Steps to reproduce:

I have the e-mail with https:// link in it.
Preferences/Attachments specify that action for https is "Use firefox (default)"
But when I click on the link nothing happens.

Comment 1

[Yuri]

Same behavior with http:// links.

Interesting that after I set an option for particular URL type to "Always Ask", dialog box "Launch Application" pops up when I click on the URL link. The first option is 'firefox'. But nothing happens when I select 'firefox' and click Ok. Dialog box just stays open.

So actions feature is broken. If application can't be launched for some reason, user should see the message explaining what ism the problem.

Setting importance to 'major' since this is a major usability issue and annoyance.

Comment 2

[Ludovic Hirlimann [:Usul]]

What version of glibc is installed on your system ?

Comment 3

[Yuri]

FreeBSD doesn't use glibc (GNU C library).

Comment 4

[Ludovic Hirlimann [:Usul]]

(In reply to comment #3)
> FreeBSD doesn't use glibc (GNU C library).

That's how we register what uri doe and what they launch.

Comment 5

[Yuri]

Could you elaborate? How does TB register this through glibc?

Comment 6

[Mike Conley (:mconley) (:⚙️)]

Yuri:

Is it possible that Firefox is opening the link in the background, and just not focusing?

-Mike

Comment 7

[Yuri]

No, this is not the case.

Comment 8

[Mike Conley (:mconley) (:⚙️)]

Yuri:

Can you open up the Error Console (Tools > Error Console), click the "CLEAR" button, try to click a link, and see if anything pops up in the dialog?

Let us know what happens - and if a message pops up, please right-click, copy, and paste into this bug.

Thanks,

-Mike

Comment 9

[Ludovic Hirlimann [:Usul]]

(In reply to comment #5)
> Could you elaborate? How does TB register this through glibc?

see bug 624341

Comment 10

[Joshua Cranmer [:jcranmer]]

Let me clarify: Ludovic probably meant `glib', not `glibc' (the two are very differnet libraries).

Comment 11

[Yuri]

Actually the error console gets the error:
Error: uncaught exception: [Exception... "Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIExternalProtocolService.loadUrl]"  nsresult: "0x80004005 (NS_ERROR_FAILURE)"  location: "JS frame :: chrome://communicator/content/contentAreaClick.js :: openLinkExternally :: line 188"  data: no]

Comment 12

[Yuri]

glib-2.26.1

Comment 13

[Mike Conley (:mconley) (:⚙️)]

Hm...this is sounding very similar to bug 668327.

Could it be that you too have a corrupted mimeTypes.rdf in your profile directory?

Very mysterious...

-Mike

Comment 14

[Yuri]

Attachment: /home/yuri/.thunderbird/qysr9gn2.default/mimeTypes.rdf

Comment 15

[Yuri]

Attachment: /home/yuri/.thunderbird/qysr9gn2.default/US/mimeTypes.rdf

Comment 16

[Yuri]

There are two mimeTypes.rdf fils in my TB profile:
/home/yuri/.thunderbird/qysr9gn2.default/mimeTypes.rdf
/home/yuri/.thunderbird/qysr9gn2.default/US/mimeTypes.rdf

Comment 17

[Ludovic Hirlimann [:Usul]]

does it work if you remove them ?

Comment 18

[Yuri]

Same thing after removing the mimeTypes.rdf
New file got created at /home/yuri/.thunderbird/qysr9gn2.default/mimeTypes.rdf after TB relaunch.

Comment 19

[Mike Conley (:mconley) (:⚙️)]

Hrm.  Frustrating.

Are you familiar with debugging Thunderbird?  If so, could you set a breakpoint at nsExternalHelperAppService::LoadUrl, step through, and see where exactly it's throwing the NS_ERROR_FAILURE?

Comment 20

[Yuri]

Will try this.

Comment 21

[Yuri]

I think this is also the case of poor error reporting.
Exception is thrown, caught and printed and yet it doesn't carry enough information about the cause.

Comment 22

[Yuri]

Here is where an exception is thrown:

CallMethodHelper::Call (this=0x7fffffffad90) at xpcwrappednative.cpp:2407
2407        mCallContext.GetXPCContext()->SetLastResult(invokeResult);
(gdb) n
2409        if(NS_FAILED(invokeResult))
(gdb) n
2411            ThrowBadResult(invokeResult, mCallContext);

Comment 23

[Mike Conley (:mconley) (:⚙️)]

Hm - that's going a bit deep.  Are you able to get a backtrace from that point?

Comment 24

[Yuri]

DEBUG version of thunderbird randomly crashes on FreeBSD by itself for some other reasons. Very hard to debug.

Comment 25

[Yuri]

=en-US&version=5.0&os=FreeBSD&buildid=20110706135403]
--DOMWINDOW == 19 (0x8130e2d78) [serial = 11] [outer = 0x811e34100] [url = about:blank]
--DOMWINDOW == 18 (0x8131b8278) [serial = 14] [outer = 0x8111d9000] [url = about:blank]
--DOMWINDOW == 17 (0x81ca2cb78) [serial = 21] [outer = 0x8111d9000] [url = imap://[e-mail]:143/fetch%3EUID%3E/INBOX%3E138342]
--DOMWINDOW == 16 (0x8142af878) [serial = 12] [outer = 0x8111da200] [url = about:blank]
###!!! ASSERTION: Not a UTF-8 string. This code should only be used for converting from known UTF-8 strings.: 'Error', file ../../../dist/include/nsUTF8Utils.h, line 453
###!!! ASSERTION: length mismatch: 'calculator.Length() == converter.Length()', file nsReadableUtils.cpp, line 402
Assertion failure: chars[length] == 0, at jsstrinlines.h:405
Abort trap: 6
===> crashes here <====

Comment 26

[Mike Conley (:mconley) (:⚙️)]

I think we're losing focus here - now we're debugging an assertion failure, which doesn't sound like the subject of this bug.  For the assertion failure, could you please file a new bug?

Were you able to get a backtrace for the point where LoadUrl fails for you?

Comment 27

[Yuri]

I wasn't able to get the stack since debug version crashes before it reaches LoadUrl.

Comment 28

[Joshua Cranmer [:jcranmer]]

Build an optimized version but don't strip the binaries...

Comment 29

[Kent James (:rkent)]

I have a similar issue, but I caution that I am running on a debug build with a few days old aurora, with my ExQuilla extension loaded, so I don't even know if my issue is valid. However the problem is appearing on IMAP trying to download the attachment of a spam message.

I'm getting a crash with the following stack:

 	mozjs.dll!JS_Assert(const char * s=0x5e5a21b4, const char * file=0x5e5a2200, int ln=171)  Line 79	C++
 	mozjs.dll!JSExternalString::new_(JSContext * cx=0x040ccbd0, const wchar_t * chars=0x05382a10, unsigned int length=101, int type=0, void * closure=0x00000000)  Line 171 + 0x24 bytes	C++
 	mozjs.dll!JS_NewExternalString(JSContext * cx=0x040ccbd0, const wchar_t * chars=0x05382a10, unsigned int length=101, int type=0)  Line 2719 + 0x17 bytes	C++
 	xul.dll!XPCConvert::NativeData2JS(XPCLazyCallContext & lccx={...}, jsval_layout * d=0x002bcccc, const void * s=0x002bcf98, const nsXPTType & type={...}, const nsID * iid=0x002bcd90, unsigned int * pErr=0x00000000)  Line 423 + 0x19 bytes	C++
 	xul.dll!XPCConvert::NativeData2JS(XPCCallContext & ccx={...}, jsval_layout * d=0x002bcccc, const void * s=0x002bcf98, const nsXPTType & type={...}, const nsID * iid=0x002bcd90, unsigned int * pErr=0x00000000)  Line 3173 + 0x20 bytes	C++
 	xul.dll!nsXPCWrappedJSClass::CallMethod(nsXPCWrappedJS * wrapper=0x09c99e78, unsigned short methodIndex=14, const XPTMethodDescriptor * info=0x03570528, nsXPTCMiniVariant * nativeParams=0x002bcf98)  Line 1551 + 0x2a bytes	C++
 	xul.dll!nsXPCWrappedJS::CallMethod(unsigned short methodIndex=14, const XPTMethodDescriptor * info=0x03570528, nsXPTCMiniVariant * params=0x002bcf98)  Line 586	C++
 	xul.dll!PrepareAndDispatch(nsXPTCStubBase * self=0x0f1d52c0, unsigned int methodIndex=14, unsigned int * args=0x002bd058, unsigned int * stackBytesToPop=0x002bd048)  Line 114 + 0x21 bytes	C++
 	xul.dll!SharedStub()  Line 142	C++
>	xul.dll!mimeEmitterStartAttachment(MimeDisplayOptions * opt=0x091b1730, const char * name=0x09b595f8, const char * contentType=0x0c627660, const char * url=0x09b596b0, int aIsExternalAttachment=0)  Line 1867 + 0x27 bytes	C++
 	xul.dll!NotifyEmittersOfAttachmentList(MimeDisplayOptions * opt=0x091b1730, nsMsgAttachmentData * data=0x042a9ba0)  Line 684 + 0x2a bytes	C++
 	xul.dll!mime_display_stream_complete(_nsMIMESession * stream=0x0c64c948)  Line 1040 + 0x10 bytes	C++
 	xul.dll!nsStreamConverter::OnStopRequest(nsIRequest * request=0x0bc2fbd0, nsISupports * ctxt=0x0428d648, unsigned int status=0)  Line 1090 + 0xf bytes	C++
 	xul.dll!nsImapCacheStreamListener::OnStopRequest(nsIRequest * request=0x0bfee208, nsISupports * aCtxt=0x0428d648, unsigned int aStatus=0)  Line 8652 + 0x30 bytes	C++
 	xul.dll!nsInputStreamPump::OnStateStop()  Line 579	C++
 	xul.dll!nsInputStreamPump::OnInputStreamReady(nsIAsyncInputStream * stream=0x09b01a40)  Line 403 + 0xb bytes	C++
 	xul.dll!nsInputStreamReadyEvent::Run()  Line 115	C++
 	xul.dll!nsThread::ProcessNextEvent(int mayWait=0, int * result=0x002bd340)  Line 617 + 0x19 bytes	C++

In mimeEmitterStartAttachment the url is:

"imap://[e-mail]:143/fetch%3EUID%3E.INBOX.spam%3E62525?header=filter&emitter=js&part=1.2"

Comment 30

[Kent James (:rkent)]

See my bug 677360 for additional details.

Comment 31

[Wayne Mery (:wsmwk)]

(In reply to Kent James (:rkent) from comment #30)
> See my bug 677360 for additional details.

dependent?  dupeu?

Comment 32

[Kent James (:rkent)]

(In reply to Wayne Mery (:wsmwk) from comment #31)
> (In reply to Kent James (:rkent) from comment #30)
> > See my bug 677360 for additional details.
> 
> dependent?  dupeu?

I think that Jonathan Protzenko [:protz] might be better able to address that. I was mainly posting because I had a lot of detail, since I was running debug builds when I saw the issue.

What I am not sure of is whether a core bug exists, as Jonathan and I said in bug 677360, or whether mailnews is simply misusing a core function.

Comment 33

[Phoenix]

Still here
Mozilla/5.0 (X11; FreeBSD i386; rv:10.0.2) Gecko/20120229 Thunderbird/10.0.2

Error: uncaught exception: [Exception... "Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIExternalProtocolService.loadUrl]"  nsresult: "0x80004005 (NS_ERROR_FAILURE)"  location: "JS frame :: chrome://communicator/content/contentAreaClick.js :: openLinkExternally :: line 188"  data: no]

Comment 34

[Phoenix]

Is this still present in latest Thunderbird version? Don't have FreeBSD installation with GUI on hardware to retest

Comment 35

[Phoenix]

Resolved per whiteboard