Closed Bug 59480 Opened 25 years ago Closed 24 years ago

Modern skin crashes mozilla when ftp browsing mozilla site

Categories

(SeaMonkey :: Themes, defect, P5)

x86
Linux
defect

Tracking

(Not tracked)

VERIFIED WORKSFORME

People

(Reporter: augustmiles, Assigned: hewitt)

References

()

Details

(Keywords: crash, modern)

Suse linux 6.3 with helixgnome installed.Gecko/20001107 downloaded from mozilla.org If I do a clean install of mozilla after \rm -rf ./.mozilla change skin to modern open the above url using keyboard naviagation Mozilla crashes. Happened several times. Talkback from clean install: TB20623132W I had no problems with the classic skin.
Does not seem to be the case with 2000110604 on Win2k
works for me on 2000110706/Mandrake 7.2
Status: UNCONFIRMED → RESOLVED
Closed: 25 years ago
Resolution: --- → WORKSFORME
Does also work for me with latest nightly build on Debian woody. Since there are two other worksforme reports, resolving bug accordingly. augustmiles, if the problem persists please provide additional information.
I tried again with Gecko/20001108 Suse 6.3 HelixGnome, Biprocessor SMP machine. It crashes every time: Two talkbacks: TB20688346M (clean install) TB20688463X (with a pre-existing .mozilla) I can crash it as often as you wish...
Status: RESOLVED → UNCONFIRMED
Resolution: WORKSFORME → ---
I have other talkbacks from another FTP site: TB20691933M for instance at ftp.espci.fr I have got a little further in finding the exact conditions needed: It seems to be linked with the memory of a vist to the ftp site. Ie Choose modern skin visit the site and open a subdirectory by clicking on the little triangle. Quit mozilla. restart mozilla revisit the top level of the site crash This does not happen with the classic skin.
Did you start with a fresh mozilla directory (not ~/.mozilla, the actual install directory), or overwrite old builds?
This was a download using the linux installer which removes the previous install directory. I then rm my .mozilla by hand, it should thus be a 100% clean setup. I do one non-standard thing in the install: I change the directory from /usr/local since I keep a M18 there.
i'm tempted to mark invalid. I suspect you did not run ./regxpcom && ./regchrome if you install into a readonly directory and run as mortal then you will crash. please run the two mentioned programs (as root or whatever owns mozilla)
I don't install mozilla as root. I just run the linux installer as a normal user. Then I lauch mozilla from a terminal. I change the install directory in order install in my home directory, I don't need root privelages for this. I have never run ./regxpcom && ./regchrome that sure! I shall do so at once and generate a talkback Ok I ran regxpcom (I had to set LD_LIBRARY_PATH and MOZILLA_FIVE_HOME). regchrome does not exist on my machine ( I checked the locate datebase), Nor does it exist in netscape PR3. Lets go... Launch mozilla, open ftp.mozilla.org using keyboard navigation. Talkback: TB20759882H I can get it NOT to crash in the following way: If I configure ftp to go through a squid ftp proxy then mozilla does not crash. However Ftp "rewrites" the ftp display so that its not a very good test case. Otherwise mozilla does run, It does NOT crash with the classic skin. It is only with the modern skin that this happens.
please try running as root once and see if that fixes this
Well I ran it as root, it crashed Talkback: TB20772447X I tried installing some other skins: ThinIce and Alfred. They work perfectly, only modern leads to the crash
when I run under gdb I generate the following Program received signal SIGSEGV, Segmentation fault. 0x40857628 in ?? () (gdb) where #0 0x40857628 in ?? () #1 0x408571d0 in ?? () #2 0x400c39b3 in PL_HandleEvent () from /home2/tony/mozilla/libxpcom.so #3 0x400c38d6 in PL_ProcessPendingEvents () from /home2/tony/mozilla/libxpcom.so #4 0x400c462d in nsEventQueueImpl::ProcessPendingEvents () from /home2/tony/mozilla/libxpcom.so #5 0x404ade0f in NSGetModule () from /home2/tony/mozilla/components/libwidget_gtk.so #6 0x404adbcd in NSGetModule () from /home2/tony/mozilla/components/libwidget_gtk.so #7 0x4066d24a in g_io_unix_dispatch (source_data=0x41c6c6c8, current_time=0xbffff4e0, user_data=0x41c6c6a0) at giounix.c:135 #8 0x4066ea83 in g_main_dispatch (dispatch_time=0xbffff4e0) at gmain.c:656 #9 0x4066f0cb in g_main_iterate (block=1, dispatch=1) at gmain.c:877 #10 0x4066f281 in g_main_run (loop=0x825add0) at gmain.c:935 #11 0x40589deb in gtk_main () at gtkmain.c:476 #12 0x404ae2fc in NSGetModule () from /home2/tony/mozilla/components/libwidget_gtk.so #13 0x4045ac0a in nsJSUtils::mCachedSecurityManager () from /home2/tony/mozilla/components/libnsappshell.so #14 0x804e275 in JS_PushArguments () #15 0x804e845 in JS_PushArguments () #16 0x4026d313 in __libc_start_main (main=0x804e718 <JS_PushArguments+12180>, argc=1, argv=0xbffff6e4, init=0x804b2d4 <_init>, fini=0x8054c48 <_fini>, rtld_fini=0x4000ac70 <_dl_fini>, stack_end=0xbffff6dc) at ../sysdeps/generic/libc-start.c:90
hrm, Asa could you see if the other talkback reports are different? let's try xpcom registration. Hewitt: could your theme be missing something required for ftp view?
Assignee: asa → rayw
Status: UNCONFIRMED → NEW
Component: Browser-General → XPCOM Registry
Ever confirmed: true
Keywords: newmod
QA Contact: doronr → rayw
OK I found one way of not getting it to crash. If I type the url in the toolbar mozilla stays stable. If I do ^L to open a dialogue and type the url there then return it crashes. All this under modern. Other skins are stable.
No, Modern isn't missing anything. I can't get this to happen on windows either, btw. If Modern is at fault, it may have something to do with the large number of images modern uses. That's all I can think of.
Keywords: crash
There is no clue in this document why this was assigned to XPCOM Registry. I am no longer the owner of XPCOM nor have I worked on XPCOM Registry since Warren took it, and it is not clear why this got assigned to me. It is far from clear who the problem does belong to. I have seldom been successful doing anything with skins in Mozilla. The inability to change skins in a developer build from the themes list (Redhat 7) prevents me from testing this. I have seldom ever had themes work properly from a clean developer build. Someone who knows themes and how to make them work should take ownership of this bug or mark it void.
Assignee: rayw → hewitt
Component: XPCOM Registry → Themes
QA Contact: rayw → pmac
workforme on linux (2001-01-12-08-Mtrunk)
AM, can you still make this happen with recent nightly builds? Myself and pmac have not been able to duplicate it.
Status: NEW → ASSIGNED
Recent build now do work for me
marking worksforme based on comments from Hewitt, pmac, AM.
Status: ASSIGNED → RESOLVED
Closed: 25 years ago24 years ago
Priority: P3 → P5
Resolution: --- → WORKSFORME
Verified "workforme" on Linux (2001-03-08-21-Mtrunk).
Status: RESOLVED → VERIFIED
Product: Core → SeaMonkey
You need to log in before you can comment on or make changes to this bug.