Closed Bug 39170 Opened 24 years ago Closed 24 years ago

win32 build fails to start up

Categories

(SeaMonkey :: General, defect, P3)

x86
Windows NT
defect

Tracking

(Not tracked)

VERIFIED FIXED

People

(Reporter: sspitzer, Assigned: asa)

References

Details

(Keywords: platform-parity, regression)

I hope this is just a fluke. linux starts up fine. here's where it is hung. tried -nosplash, to see if that would help, but no luck. USER32! 77eb08f5() USER32! 77e73122() USER32! 77ea10ae() USER32! 77e87251() USER32! 77ea0857() nsNativeAppSupportWin::Start(nsNativeAppSupportWin * const 0x00c449e0, int * 0x0012ff50) line 565 + 27 bytes main(int 1, char * * 0x00c44a20) line 1013 + 16 bytes mainCRTStartup() line 338 + 17 bytes KERNEL32! 77f1ba06()
2000051220 seemed to load fine for me on win98
how about the 2000051308 build?
will try on that when it becomes available
You're right - I can't start 2000051308 on win98. Get a GPF in XPCOM.dll, cc'ing dp. Talkback incident ID: TB10440393W Raising severity
Severity: normal → blocker
Keywords: crash, regression
Summary: win32 build fails to start up → win32 build fails to start up - crashes
the 5-13-9am build from ftp://ftp.mozilla.org (mozilla-win32.zip) doesn't crash for me. hmm. maybe I'm crying wolf here? let me wait until my build is done.
I made a mistake in my last comment: I wasn't on win98 when it didn't work, I was on Windows ME (Millennium) RC0. (I've upgraded from 98 so I can't test and see if 51308 also crashes under that for me). I sure hope the problem isn't just that Mozilla doesn't work under Millennium, period, cause then I gotta revert back to win98. Furthermore, apparently ME (or this release candidate, anyways) no longer spews back anything in GPF's...the Details button is totally gone. Can someone get the talkback data please (I posted the incident ID earlier)? Seth, lemme know how your build goes. You wouldn't happen to have a copy of ME RC0 lying around to get me a stack trace, would you? Perhaps you could try it in one of the nscp labs?
I can run 2000-05-13-08 without problem on win2k. Will try win95 in a moment.
Here's the stack trace from the talkback incident. PL_strcmp() nsStdURL::GetFileName [d:\builds\seamonkey\mozilla\netwerk\base\src\nsStdURL.cpp, line 980] NKFILE.DLL + 0x1534 (0x60691534) NKFILE.DLL + 0x22ee (0x606922ee) nsIOService::QueryInterface [d:\builds\seamonkey\mozilla\netwerk\base\src\nsIOService.cpp, line 106] NS_OpenURI [..\..\..\dist\include\nsNetUtil.h, line 188] NS_OpenURI [..\..\..\dist\include\nsNetUtil.h, line 124] RDFXMLDataSourceImpl::Refresh [d:\builds\seamonkey\mozilla\rdf\base\src\nsRDFXMLDataSource.cpp, line 911] nsChromeRegistry::LoadDataSource [d:\builds\seamonkey\mozilla\rdf\chrome\src\nsChromeRegistry.cpp, line 612] nsChromeRegistry::AddToCompositeDataSource [d:\builds\seamonkey\mozilla\rdf\chrome\src\nsChromeRegistry.cpp, line 1604] nsChromeRegistry::ConvertChromeURL [d:\builds\seamonkey\mozilla\rdf\chrome\src\nsChromeRegistry.cpp, line 390] nsChromeProtocolHandler::NewChannel [d:\builds\seamonkey\mozilla\rdf\chrome\src\nsChromeProtocolHandler.cpp, line 688] nsIOService::QueryInterface [d:\builds\seamonkey\mozilla\netwerk\base\src\nsIOService.cpp, line 106] NS_OpenURI [..\..\dist\include\nsNetUtil.h, line 104] nsDocShell::DoURILoad [d:\builds\seamonkey\mozilla\docshell\base\nsDocShell.cpp, line 2531] nsDocShell::InternalLoad [d:\builds\seamonkey\mozilla\docshell\base\nsDocShell.cpp, line 2297] nsDocShell::LoadURI [d:\builds\seamonkey\mozilla\docshell\base\nsDocShell.cpp, line 215] GlobalWindowImpl::OpenInternal [d:\builds\seamonkey\mozilla\dom\src\base\nsGlobalWindow.cpp, line 2644] GlobalWindowImpl::OpenDialog [d:\builds\seamonkey\mozilla\dom\src\base\nsGlobalWindow.cpp, line 1630] NS_NewURI [..\..\dist\include\nsNetUtil.h, line 52] Ensure1Window [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp, line 589] LaunchApplication [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp, line 324] nsPref::EnumerateChildren [d:\builds\seamonkey\mozilla\modules\libpref\src\nsPref.cpp, line 1000] LaunchApplicationWithArgs [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp, line 340] HandleArbitraryStartup [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp, line 510] DoCommandLines [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp, line 547] main [d:\builds\seamonkey\mozilla\xpfe\bootstrap\nsAppRunner.cpp, line 1021] NS_SetupRegistry [d:\builds\seamonkey\mozilla\xpfe\bootstrap\../../webshell/tests/viewer/nsSetupR egistry.cpp, line 325] MOZILLA.EXE + 0x397a (0x0040397a) KERNEL32.DLL + 0x1b9e4 (0xbff7b9e4) KERNEL32.DLL + 0x1b896 (0xbff7b896) KERNEL32.DLL + 0x1a24f (0xbff7a24f)
Thanks, John. Any ideas from the stack, anyone?
Lowering severity to critical; I now fear this might just be a problem with Windows Millennium, and since that's only a RC, we can't be expected to support it. I don't want to hold the tree/anything up with this blocker.
Severity: blocker → critical
marking invalid, my last build works fine. false alarm. I'll put down the crack pipe now.
Status: NEW → RESOLVED
Closed: 24 years ago
Resolution: --- → INVALID
Any ideas about my situation (see the stack trace)?
Status: RESOLVED → VERIFIED
Blake -- can you try a clean install (just to be certain) -- delete mozregistry.dat, your .\users50 directory, and install to a clean/empty directory (or move these items to c:\foobar to get them out of the way). If you still get the crash, then file a new bug. Thanks.
This will happen to a lot of people, but only once. I sent a note to the mozilla newsgroup and to seamonkey-internal about this, but doubtless a lot of people will miss the message. They'll all have problems. I wasn't expecting a crash... The problem is, there's obsolete information in your dist/bin/chrome directory that confuses things. The bad things should be cleared out after you run once, but you really want to delete dist before building after pulling today. By the way, the Netscape commercial build is quite broken on Windows and Unix after this checkin. I'm resurrecting that right now.
On Windows Millennium RC0, I cannot start 051316. (John, I'm not gonna open a new bug since this seems to be the same problem as initially reported it - it's no longer crashing for me, just hanging as seth reported). This appears in the console upon starting (while it's just hanging at the splash screen): stdout directed to dynamic console stderr directed to dynamic console Profile Manager : Profile Wizard and Manager activites : Begin Profile Manager : Command Line Options : Begin Profile Manager : Command Line Options : End ProfileManager : GetProfileDir ProfileManager : GetProfileDir Profile Manager : Profile Wizard and Manager activites : End has multiple monitor apis is 1 WEBSHELL+ = 1 WEBSHELL+ = 2 Java HotSpot(TM) Client VM warning: Setting of property "java.compiler" is ignor ed initialization error: Can't load class netscape/javascript/JSUtil This is after wiping everything out and using the mozilla-win32-installer.exe build to do a completely fresh install into a new directory.
Status: VERIFIED → REOPENED
Keywords: crash
Resolution: INVALID → ---
Summary: win32 build fails to start up - crashes → win32 build fails to start up
As danm noted above, he's been whacking the chrome directory pretty hard. It looks possible that this has affected the verification build for win32 -- I can run viewer from the -talkback.zip but not mozilla; my linux clobber build just finished and runs fine. I expect that for tomorrow's a.m. builds, everything should be fine, but this particular ftp.mozilla.org build is hosed (oh, well). (reducing the cc: list)
Severity: critical → normal
QA Contact: jelwell → jrgm
*** Bug 39169 has been marked as a duplicate of this bug. ***
Raising severity back to blocker. I've heard from multiple people that the 051409 builds aren't starting on win98. Mine still isn't getting past the splash screen, console reads: stdout directed to dynamic console stderr directed to dynamic console Profile Manager : Profile Wizard and Manager activites : Begin Profile Manager : Command Line Options : Begin Profile Manager : Command Line Options : End ProfileManager : GetProfileDir ProfileManager : GetProfileDir Profile Manager : Profile Wizard and Manager activites : End has multiple monitor apis is 1 WEBSHELL+ = 1 I am inside the initialize Hey : You are in QFA Startup (QFA)Talkback loaded Ok. WEBSHELL+ = 2
Severity: normal → blocker
Keywords: pp
I dunno if that's for all builds, it seems to be true for any "pre-compilied" nightlies from mozilla.org since 2000-05-13-16. And for all platforms. The builds hang while splash screen is showing. And it's not only happening once and running fine at the second try to start. It's hanging at every start of mozilla. (used 2000-05-14-09, win98)
there was a big chrome re-org this weekend. could this be a packaging problem? my builds are working ok. dan, did the packager-* files (ns and mozilla tree) get updated?
This is a different problem from this report's original. We've rearranged the chrome into a hierarchy in which the application can't find files without hints from the installer. The build system fakes some hints, but the installer folks haven't had a chance to catch up. So builds will work fine, but pre-packaged installer builds will choke on their own chrome. Relenting from our rudeness, we've since patched the default chrome location code so it won't require installer hints. The pressure's off the installer guys. Tonight's nightly builds should be fixed.
Status: REOPENED → RESOLVED
Closed: 24 years ago24 years ago
Resolution: --- → FIXED
Target Milestone: --- → M16
*** Bug 39249 has been marked as a duplicate of this bug. ***
*** Bug 39272 has been marked as a duplicate of this bug. ***
vrfy fixed (60609 win98/2k)
Status: RESOLVED → VERIFIED
Product: Browser → Seamonkey
You need to log in before you can comment on or make changes to this bug.