74115 - very unstable build, crash in many sites and components

Status: VERIFIED FIXED

(SeaMonkey :: General, defect)

Description

[Tracy Walker [:tracy]]

seen on linux commercial build 2001-03-30-05-mtrunk

this build is very unstable.  It crashes in nearly every site from the smoketest 
suite.  It also crashes attempting to open the mail app and crashes attempting 
to open a chat room in IM.  There should be plenty of talk back reports 
available.  I sent one each time I crashed.  I eventually gave up testing by the 
time I got to international browser testing {it crashed there too).  This could 
be more than one bug, but since I can't tell the difference and writing up a 
slew of bugs seems silly, I'm just calling this build unstable.  I am going to 
write another bug on not being able to launch from a new profile, because I do 
believe that may be unrelated to this.

note:  it's only on linux...mac is okay and I haven't seen windows build for 
this morning.

Comment 1

[Simon Fraser [no longer active]]

Can you paste one or two stack traces here?

Comment 2

[Tracy Walker [:tracy]]

stack trace for mail crash

libXIE.so.6 + 0x3a3b (0x40948a3b) 
libgfx_gtk.so + 0x96cb (0x409296cb) 
libgfx_gtk.so + 0x986b (0x4092986b) 
libgfx_gtk.so + 0x12137 (0x40932137) 
libgfx_gtk.so + 0x121ea (0x409321ea) 
libgkgfx.so + 0x1a73d (0x4002d73d) 
libgklayout.so + 0x4a080 (0x40d35080) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x33734 (0x40d1e734) 
libgklayout.so + 0x3359a (0x40d1e59a) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x371ac (0x40d221ac) 
libgklayout.so + 0xcc5c7 (0x40db75c7) 
libgklayout.so + 0xdb0f7 (0x40dc60f7) 
libgklayout.so + 0xdafc2 (0x40dc5fc2) 
libgklayout.so + 0xdcf57 (0x40dc7f57) 
libgklayout.so + 0xdce45 (0x40dc7e45) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x371ac (0x40d221ac) 
libgklayout.so + 0xd1c3c (0x40dbcc3c) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0xd761e (0x40dc261e) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x33734 (0x40d1e734) 
libgklayout.so + 0x3359a (0x40d1e59a) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x33734 (0x40d1e734) 
libgklayout.so + 0x3359a (0x40d1e59a) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x33734 (0x40d1e734) 
libgklayout.so + 0x3359a (0x40d1e59a) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x371ac (0x40d221ac) 
libgklayout.so + 0x42c09 (0x40d2dc09) 
libgklayout.so + 0x63b0f (0x40d4eb0f) 
libgkview.so + 0x37b7 (0x40f4b7b7) 
libgkview.so + 0xb9b6 (0x40f539b6) 
libgkview.so + 0xb79f (0x40f5379f) 
libgkview.so + 0xaa85 (0x40f52a85) 
libgkview.so + 0xcaf0 (0x40f54af0) 
libgkview.so + 0x334d (0x40f4b34d) 
libwidget_gtk.so + 0x1a60a (0x404c160a) 
libwidget_gtk.so + 0x1a535 (0x404c1535) 
libwidget_gtk.so + 0x1deb5 (0x404c4eb5) 
libwidget_gtk.so + 0x1dfb9 (0x404c4fb9) 
libwidget_gtk.so + 0x1dd79 (0x404c4d79) 
libglib-1.2.so.0 + 0x10ba9 (0x40655ba9) 
libglib-1.2.so.0 + 0xfbe6 (0x40654be6) 
libglib-1.2.so.0 + 0x101a1 (0x406551a1) 
libglib-1.2.so.0 + 0x10341 (0x40655341) 

Comment 3

[Tracy Walker [:tracy]]

stack trace for crash in my.netscape.com

libXIE.so.6 + 0x3a3b (0x40948a3b) 
libgfx_gtk.so + 0x96cb (0x409296cb) 
libgfx_gtk.so + 0x986b (0x4092986b) 
libgfx_gtk.so + 0x12137 (0x40932137) 
libgfx_gtk.so + 0x121ea (0x409321ea) 
libgkgfx.so + 0x1a73d (0x4002d73d) 
libgklayout.so + 0x4a080 (0x40d35080) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x33734 (0x40d1e734) 
libgklayout.so + 0x3359a (0x40d1e59a) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x371ac (0x40d221ac) 
libgklayout.so + 0xcc5c7 (0x40db75c7) 
libgklayout.so + 0xdb0f7 (0x40dc60f7) 
libgklayout.so + 0xdafc2 (0x40dc5fc2) 
libgklayout.so + 0xdcf57 (0x40dc7f57) 
libgklayout.so + 0xdce45 (0x40dc7e45) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x371ac (0x40d221ac) 
libgklayout.so + 0xd1c3c (0x40dbcc3c) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0xd761e (0x40dc261e) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x33734 (0x40d1e734) 
libgklayout.so + 0x3359a (0x40d1e59a) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x33734 (0x40d1e734) 
libgklayout.so + 0x3359a (0x40d1e59a) 
libgklayout.so + 0x372d1 (0x40d222d1) 
libgklayout.so + 0x371ac (0x40d221ac) 
libgklayout.so + 0x42c09 (0x40d2dc09) 
libgklayout.so + 0x63b0f (0x40d4eb0f) 
libgkview.so + 0x37b7 (0x40f4b7b7) 
libgkview.so + 0xb9b6 (0x40f539b6) 
libgkview.so + 0xb79f (0x40f5379f) 
libgkview.so + 0xaa85 (0x40f52a85) 
libgkview.so + 0xcaf0 (0x40f54af0) 
libgkview.so + 0x334d (0x40f4b34d) 
libwidget_gtk.so + 0x1a60a (0x404c160a) 
libwidget_gtk.so + 0x1a535 (0x404c1535) 
libwidget_gtk.so + 0x1deb5 (0x404c4eb5) 
libwidget_gtk.so + 0x1e035 (0x404c5035) 
libwidget_gtk.so + 0x1e11c (0x404c511c) 
libwidget_gtk.so + 0x1dd79 (0x404c4d79) 
libglib-1.2.so.0 + 0x10ba9 (0x40655ba9) 
libglib-1.2.so.0 + 0xfbe6 (0x40654be6) 
libglib-1.2.so.0 + 0x101a1 (0x406551a1) 
libglib-1.2.so.0 + 0x10341 (0x40655341) 
libgtk-1.2.so.0 + 0x8c209 (0x4057f209) 
libwidget_gtk.so + 0xd5ec (0x404b45ec) 
libnsappshell.so + 0xd76a (0x4045976a) 
mozilla-bin + 0x60b4 (0x0804e0b4) 
mozilla-bin + 0x6945 (0x0804e945) 
libc.so.6 + 0x181eb (0x4023f1eb) 

Comment 4

[Tracy Walker [:tracy]]

the stack for the IM chat room crash (bug 4502)  looks identical to the above 
trace.

Comment 5

[Tracy Walker [:tracy]]

ooops I cut and paste poorly...the first trace is missing these lines at the 
end:

libgtk-1.2.so.0 + 0x8c209 (0x4057f209) 
libwidget_gtk.so + 0xd5ec (0x404b45ec) 
libnsappshell.so + 0xd76a (0x4045976a) 
mozilla-bin + 0x60b4 (0x0804e0b4) 
mozilla-bin + 0x6945 (0x0804e945) 
libc.so.6 + 0x181eb (0x4023f1eb) 

Comment 6

[tor]

XIE is part of the pr0n landing for linux.  Reassigning to pavlov, CCing syd.

Comment 7

[Jens-Uwe]

*** Bug 74110 has been marked as a duplicate of this bug. ***

Comment 8

[Stuart Parmenter]

Attachment: avoid drawing if we have a negative width or height

Comment 9

[Stuart Parmenter]

i'm working on figuring out why we have a negative width/height in the first 
place

Comment 10

[Syd Logan]

Comments don't match the code ( < 1 != is negative ). 

Comment 11

[Stuart Parmenter]

whutever

Comment 12

[saari (gone)]

change that to check to <= 0 and r=saari.
0x0 drawing will crash GDI on windows (I know this is linux code, but best to be
safe)

Comment 13

[Stuart Parmenter]

Attachment: better patch

Comment 14

[Daniel (Leaf) Nunes]

can we get more r=, sr=?

Comment 15

[saari (gone)]

r=saari on new patch

ever figure out why we were getting negative sizes?

Comment 16

[Stuart Parmenter]

the previous conditioning code was fucking up the coords when things were 
scaled.

Comment 17

[Scott Collins]

sr=scc

Comment 18

[Stuart Parmenter]

fixed

Comment 19

[Syd Logan]

Attachment: patch to fix the previously described problem

Comment 20

[Syd Logan]

Above patch is for the following:

Assume you don't have gdkpixbuf, and compiled with HAS_XIE but the XIE extension
is not supported by the server (or the extension module not loaded). Then, when
we init XIE scaled image code, we set inited to PR_TRUE and return PR_FALSE to
fail because XieInitialize() fails to query the extension. However, setting
inited to true is bad because later we call the routine to scale the image and
we try packing a photoflo element into a photoflo graph that was never
allocated, and core dump. Pavlov, might need to check to ensure we are not
trying to free the photoflo graph later, or unallocate a photospace that was
never allocated in the first place.

Comment 21

[Syd Logan]

Attachment: ignore the previous patch

Comment 22

[Daniel Veditz [:dveditz]]

r=dveditz for the 3/31 00:34  patch

Comment 23

[tor]

sr=tor

Comment 24

[Tracy Walker [:tracy]]

verified with commercial builds from 04/02