http://bugs.winehq.org/show_bug.cgi?id=17188
Summary: Oracle Web Conferencing Console installer resources cannot be properly loaded Product: Wine Version: 1.1.13 Platform: PC URL: https://conference.oracle.com/imtapp/app/nuf_sys.uix OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: kernel32 AssignedTo: wine-bugs@winehq.org ReportedBy: arethusa26@gmail.com
Created an attachment (id=19083) --> (http://bugs.winehq.org/attachment.cgi?id=19083) Oracle Web Conferencing Console installer +resource trace
With today's Git (wine-1.1.13-560-g4b27dfe), while attempting to test bug 16684 by running the application installer, I noticed that the installer evidently had issues loading its executable resources, as these fixmes were printed:
fixme:propsheet:PROPSHEET_CollectPageInfo Could not load resource #0000? fixme:propsheet:PROPSHEET_CollectPageInfo Could not load resource #0000? fixme:propsheet:PROPSHEET_CollectPageInfo Could not load resource #0000? fixme:propsheet:PROPSHEET_CollectPageInfo Could not load resource #0000?
and the installer window seemed to have placeholders for the presumed text resources. After hitting Next a few times the installer appeared to be in extraction phase, but promptly crashed with a backtrace in FormatMessageW.
http://bugs.winehq.org/show_bug.cgi?id=17188
Andrew Nguyen arethusa26@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |download, Installer
http://bugs.winehq.org/show_bug.cgi?id=17188
Louis Lenders xerox_xerox2000@yahoo.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW CC| |xerox_xerox2000@yahoo.co.uk Component|kernel32 |msxml3 Ever Confirmed|0 |1
--- Comment #1 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2009-08-16 18:52:33 --- This looks more like some bug in msxml3. Try 'winetricks msxml3' and the installer finishes fine
http://bugs.winehq.org/show_bug.cgi?id=17188
Andrew Nguyen arethusa26@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #19083|0 |1 is obsolete| |
--- Comment #2 from Andrew Nguyen arethusa26@gmail.com 2010-02-21 20:00:37 --- Created an attachment (id=26397) --> (http://bugs.winehq.org/attachment.cgi?id=26397) Oracle Web Conferencing Console installer +msxml trace
Indeed, this seems to be a msxml3 issue. I'm attaching an updated log.
http://bugs.winehq.org/show_bug.cgi?id=17188
--- Comment #3 from Nikolay Sivov bunglehead@gmail.com 2010-06-20 17:24:57 --- Created an attachment (id=29042) --> (http://bugs.winehq.org/attachment.cgi?id=29042) patch
After this patch installer is able to get an xml:lang value and starts to load strings:
--- 0009:trace:msxml:xmlelem_getAttribute (0x127138, L"xml:lang", 0x33722c) 0009:trace:msxml:xmlelem_getAttribute returning L"en" ---
but next I see some strange text corruptions in a loaded strings:
--- 0009:trace:msxml:xmlelem_get_text (0x127120, 0x337260) 0009:trace:msxml:xmlelem_get_text returning L"\0003|ect" 0009:trace:msxml:xmlelem_getAttribute (0x127120, L"id", 0x337124) 0009:trace:msxml:xmlelem_getAttribute returning L"3" 0009:trace:msxml:xmlelem_getAttribute (0x127120, L"name", 0x337124) 0009:trace:msxml:xmlelem_getAttribute returning (null) ---
text should be "Connect" in this case. File has a utf-8 encoding attribute, and seems to be fine - no non-ascii characters used in it.
Anyway after that there's a minor problem with enumerator that I need to fix.
http://bugs.winehq.org/show_bug.cgi?id=17188
--- Comment #4 from Nikolay Sivov bunglehead@gmail.com 2010-06-21 13:14:08 --- Commits
b00046ccc44ad3ed3368f1b62d8d1ad2bd7eeade 473c5c4cc02b376ae561368cd5623372fe4223b5
fixes a problem with loading strings from xml.
Now it crashes because of broken strings returned that I mentioned in comment 3. All strings have first four chars broken, don't know why this happens. Native msxml3 still helps.
http://bugs.winehq.org/show_bug.cgi?id=17188
--- Comment #5 from Andrew Nguyen arethusa26@gmail.com 2010-09-19 03:42:18 CDT --- Created an attachment (id=30862) --> (http://bugs.winehq.org/attachment.cgi?id=30862) +msxml Wine trace and Valgrind output excerpt
I think this is some sort of reference counting problem. My attachment of an excerpt of a Valgrind run on the installer shows that, for some reason, the reference count of the XML document object goes to zero when IPersistStreamInit::Release is called, causing the libxml2 resources to be freed and creating many problems later on.
If I hack xmldoc_IPersistStreamInit_Release to always return 1, the installer seems to work as expected.
http://bugs.winehq.org/show_bug.cgi?id=17188
--- Comment #6 from Nikolay Sivov bunglehead@gmail.com 2010-09-19 04:38:05 CDT --- Created an attachment (id=30863) --> (http://bugs.winehq.org/attachment.cgi?id=30863) test
Simple test shows that ::QueryInterface() on document for IPersistStreamInit doesn't touch document reference, and even more - ::AddRef() on IPersistStreamInit instance doesn't touch document references.
So we need to make IPersistStreamInit a separate object internally linked to document data.
Please let me know if you're ready to fix that.
http://bugs.winehq.org/show_bug.cgi?id=17188
--- Comment #7 from Andrew Nguyen arethusa26@gmail.com 2010-09-19 05:10:38 CDT --- (In reply to comment #6)
Simple test shows that ::QueryInterface() on document for IPersistStreamInit doesn't touch document reference, and even more - ::AddRef() on IPersistStreamInit instance doesn't touch document references.
Thanks for the fast follow-up with the test.
So we need to make IPersistStreamInit a separate object internally linked to document data.
Please let me know if you're ready to fix that.
Unfortunately, I don't have very much time for Wine, so I probably would be slow to do the IPersistStreamInit separation job.
http://bugs.winehq.org/show_bug.cgi?id=17188
Nikolay Sivov bunglehead@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #30863|0 |1 is obsolete| |
--- Comment #8 from Nikolay Sivov bunglehead@gmail.com 2010-09-22 17:27:21 CDT --- Created an attachment (id=30918) --> (http://bugs.winehq.org/attachment.cgi?id=30918) patch
Please try this one to see if it fixes reference count problem.
http://bugs.winehq.org/show_bug.cgi?id=17188
--- Comment #9 from Andrew Nguyen arethusa26@gmail.com 2010-09-22 21:30:30 CDT --- Created an attachment (id=30919) --> (http://bugs.winehq.org/attachment.cgi?id=30919) Oracle Web Conferencing Console installer +msxml and Valgrind trace with patch applied
(In reply to comment #8)
Created an attachment (id=30918)
--> (http://bugs.winehq.org/attachment.cgi?id=30918) [details]
patch
Please try this one to see if it fixes reference count problem.
Your patch certainly satisfies the tests you wrote, but it doesn't satisfy the installer, as it still shows the incorrect behavior. Judging from:
fixme:msxml:streaminit_IPersistStreamInit_QueryInterface interface {f52e2b61-18a1-11d1-b105-00805f49916b} not implemented
the installer seems to be asking for IXMLDocument through the IPersistStreamInit instance. When Valgrind is run on the installer with the patch applied, I notice some invalid reads in streaminit_IPersistStreamInit_Load also. I'm attaching a complete +msxml and Valgrind trace.
http://bugs.winehq.org/show_bug.cgi?id=17188
Nikolay Sivov bunglehead@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #30918|0 |1 is obsolete| |
--- Comment #10 from Nikolay Sivov bunglehead@gmail.com 2010-09-23 02:30:51 CDT --- Created an attachment (id=30926) --> (http://bugs.winehq.org/attachment.cgi?id=30926) patch2
Ok, here is next iteration.
http://bugs.winehq.org/show_bug.cgi?id=17188
--- Comment #11 from Andrew Nguyen arethusa26@gmail.com 2010-09-23 02:47:24 CDT --- Created an attachment (id=30927) --> (http://bugs.winehq.org/attachment.cgi?id=30927) Oracle Web Conferencing Console installer crash backtrace
(In reply to comment #10)
Created an attachment (id=30926)
--> (http://bugs.winehq.org/attachment.cgi?id=30926) [details]
patch2
Ok, here is next iteration.
Now, strangely, the installer just crashes. The backtrace and +msxml output is attached.
http://bugs.winehq.org/show_bug.cgi?id=17188
--- Comment #12 from Nikolay Sivov bunglehead@gmail.com 2010-09-23 02:57:41 CDT --- Probably it does:
- create doc; - get IPersistStreamInit - release doc; - ::Load() from streaminit; - get a doc pointer back.
This isn't handled, probably we need to create a doc again on ::Load() if it was freed (we could see if it's freed zeroing streaminit->doc on document release).
So I hope I'll get into it again later today.
http://bugs.winehq.org/show_bug.cgi?id=17188
--- Comment #13 from Austin English austinenglish@gmail.com 2012-04-02 00:09:02 CDT --- https://oraprdcs01.imminet.com/imtapp/app/nuf_sys.uix
gives me: 7de04cc226f1a397c9cf07b876a0039d37979b72 cnsconf_en.exe
which installs fine in 1.5.1. The resource placeholder problem is still there, however.
Do you still have your original installer Andrew? Can you please retest?
https://bugs.winehq.org/show_bug.cgi?id=17188
Frédéric Delanoy frederic.delanoy@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |frederic.delanoy@gmail.com Resolution|--- |ABANDONED
--- Comment #14 from Frédéric Delanoy frederic.delanoy@gmail.com --- (In reply to Austin English from comment #13)
https://oraprdcs01.imminet.com/imtapp/app/nuf_sys.uix
gives me: 7de04cc226f1a397c9cf07b876a0039d37979b72 cnsconf_en.exe
which installs fine in 1.5.1. The resource placeholder problem is still there, however.
Do you still have your original installer Andrew? Can you please retest?
No answer from OP. Marking abandoned.
Feel free to reopen if problem persists with the latest wine version (1.7.40 or later).
https://bugs.winehq.org/show_bug.cgi?id=17188
Frédéric Delanoy frederic.delanoy@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #15 from Frédéric Delanoy frederic.delanoy@gmail.com --- Closing