Closed Bug 53639 Opened 24 years ago Closed 24 years ago

Mozilla fails to start, infinite loop

Categories

(Core :: Networking, defect, P3)

x86
Linux
defect

Tracking

()

VERIFIED DUPLICATE of bug 42184

People

(Reporter: hal, Assigned: warrensomebody)

Details

Starting maybe a week ago, I am unable to start any of the nightly builds for Linux. The program starts to load, then appears to hang. Strace shows an infinite loop with gettimeofday(). poll([{fd=12, events=POLLIN}, {fd=13, events=POLLIN}, {fd=5, events=POLLIN}], 3, 9) = 0 gettimeofday({969552414, 880317}, NULL) = 0 gettimeofday({969552414, 880453}, NULL) = 0 ioctl(12, FIONREAD, [0]) = 0 poll([{fd=12, events=POLLIN}, {fd=13, events=POLLIN}, {fd=5, events=POLLIN}], 3, 0) = 0 gettimeofday({969552414, 880908}, NULL) = 0 ioctl(12, FIONREAD, [0]) = 0 poll([{fd=12, events=POLLIN}, {fd=13, events=POLLIN}, {fd=5, events=POLLIN}], 3, 9) = 0 gettimeofday({969552414, 900542}, NULL) = 0 gettimeofday({969552414, 900689}, NULL) = 0 ioctl(12, FIONREAD, [0]) = 0 poll([{fd=12, events=POLLIN}, {fd=13, events=POLLIN}, {fd=5, events=POLLIN}], 3, 0) = 0 gettimeofday({969552414, 901126}, NULL) = 0 ioctl(12, FIONREAD, [0]) = 0 poll([{fd=12, events=POLLIN}, {fd=13, events=POLLIN}, {fd=5, events=POLLIN}], 3, 9) = 0 gettimeofday({969552414, 920313}, NULL) = 0 gettimeofday({969552414, 920462}, NULL) = 0 ioctl(12, FIONREAD, [0]) = 0 ad infinitum. Mozilla-bin has to be manually killed. I start with a clean tree each day, and tried removing ~/.mozilla. The system is RH6.2 Intel (updated) SMP (wild hunch this is a factor). Until this point, the nightlies were doing just fine. I have tried nightlies pretty much everyday during this period, including today (9-21). hal@feenix hal]$ procinfo Linux 2.2.16pre3 (root@feenix) (gcc egcs-2.91.66) #1 2CPU [feenix.] Memory: Total Used Free Shared Buffers Cached Mem: 128176 115084 13092 51384 2884 29484 Swap: 190592 71372 119220 Bootup: Wed May 17 19:40:49 2000 Load average: 0.09 0.06 0.02 7/95 7733 user : 5d 6:15:14.02 2.0% page in : 52948884 disk 1: 457134r 96699w nice : 0:00:10.86 0.0% page out: 25763749 disk 2: 531138r 1833967w system: 7d 7:58:17.35 2.9% swap in : 5205856 disk 3: 8342603r 7463311w idle : 241d 8:45:07.01 95.0% swap out: 1264101 uptime: 126d 23:29:24.60 context :3969318336 irq 0:1097096462 timer irq 9: 11403832 eth0 irq 1: 13961249 keyboard irq 10: 93048 eth1 irq 2: 0 cascade [4] irq 11: 33165 ide2 irq 4: 238230 irq 12: 22918783 PS/2 Mouse irq 5: 342887 soundblaster irq 13: 1 fpu irq 6: 28501 irq 14: 2869773 ide0 irq 8: 515 rtc irq 15: 15687021 ide1
maybe bug 52111. over to waren.
Assignee: asa → warren
Status: UNCONFIRMED → NEW
Component: Browser-General → Networking
Ever confirmed: true
After looking at the possibly related bug report, I'll add that my glibc version is currently glibc-2.1.3-21, though had the same situation with whatever the shipped version is with RH6.2. I would also note that this happens with *every* attempt to start Mozilla. HTH
What's the failure you get in the log file as you start up? (Just include the entire long). Also, this isn't an infinite loop -- your chrome just failed to load leaving you with no window.
Sorry to be clueless, for cannot find a log file for the life of me. Nor any references to one in FAQ or elsewhere. The best I can get is program output: wait4(-1, MOZILLA_FIVE_HOME=/usr/lib/mozilla LD_LIBRARY_PATH=/usr/lib/mozilla LIBRARY_PATH=/usr/lib/mozilla:/usr/lib/mozilla/components SHLIB_PATH=/usr/lib/mozilla LIBPATH=/usr/lib/mozilla ADDON_PATH=/usr/lib/mozilla MOZ_PROGRAM=/usr/lib/mozilla/mozilla-bin MOZ_TOOLKIT= moz_debug=0 moz_debugger= And hangs here. Glady give more info if you can point me a bit in the right direction. TIA
I'm talking about the program output when it starts up. Usually you see a bunch of messages about registering components, then stuff about registering chrome, etc. Are you running a debug build? Try touching your dist/bin/installed-chrome.txt file first and re-launching.
This is the grand total of output I get: [hal@feenix mozilla]$ ./mozilla ./run-mozilla.sh ./mozilla-bin MOZILLA_FIVE_HOME=. LD_LIBRARY_PATH=. LIBRARY_PATH=.:./components SHLIB_PATH=. LIBPATH=. ADDON_PATH=. MOZ_PROGRAM=./mozilla-bin MOZ_TOOLKIT= moz_debug=0 moz_debugger= Stops right there. Tried touching installed-chrome.txt. I am running mozilla-i686-pc-linux-gnu.tar.gz. Latest was yesterday's.
I don't know -- is that a debug build? I can only suggest you try running it in the debugger and seeing where it hangs since we don't have that problem here.
Well I hope I am not wasting anybody's time with this. I keep looking over my should to see if I am doing something really stupid. Anyway AFAICT the nightly binaries have no debugging support, so I dl'd the nightly source and built that yesterday. Did not run either at first, then deleted ~/.mozilla, and bingo! it ran. Then tried the nightly binary (from 2 days ago) which had been a no go, and it ran too. Used it yesterday and today no problem. Really coming along great BTW. Presumably .mozilla was the problem, though I know I had tried deleting this before. So anyway today, I decide to try the latest nightly binary, and back to square one. Will not run. Hangs just after starting with I guess no useful output. Deleting ~/.mozilla is not doing any good. I redownloaded thinking maybe I got a corrupted something or other on the way. But still no go. The built source from yesterday still does run however. So I am lost I guess at this point since I can't get any useful output from the nightly binaries. Is there something besides ~/.mozilla that might be munging things? I've looked and don't see anything. Sorry to be a bother.
if you install... you have to run moz once as root before running as a normal user. A handful files and directories are created. Can this be the problem?
Yes, Kristen, this seems to do the trick, thanks. Is this something new, or have I just been lucky in the past by not doing this?
*** This bug has been marked as a duplicate of 42184 ***
Status: NEW → RESOLVED
Closed: 24 years ago
Resolution: --- → DUPLICATE
verified
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.