bug#25282: emacs 25,1 crash w/bt

bug#25282: emacs 25,1 crash w/bt

[Hlöðver Sigurðsson]

[-- Attachment #1: Type: text/plain, Size: 2054 bytes --]

My emacs 25.1 is crashing these days every 3 hours, irrelevant if Im
on my laptop or work computer, and if Im programming in C or
Clojure. Im always using Fedora, and I install emacs either by
compiling from sources or installed from dnf. Just a paste of my last
gdb crash backtrace: https://www.refheap.com/124399

Crashes are on my computer more likely when saveing or simply using the
cursor around a document. I have aggressive indent mode enabled, haven't
ruled that out yet. Looks to me still that something is not right in
intervals.c?

Thanks for all feedback, hope I don't have to downgrade my emacs.

Best,
Hlödver

(the beginning of the trace)
Thread 1 "emacs" received signal SIGSEGV, Segmentation fault.
0x00000000005a0937 in balance_an_interval (i=0x685d630) at intervals.c:443
443    + RIGHT_TOTAL_LENGTH (i->left) - LEFT_TOTAL_LENGTH (i->left);
(gdb) bt
#0  0x00000000005a0937 in balance_an_interval (i=0x685d630) at
intervals.c:443
#1  0x00000000005a0ba5 in balance_possible_root_interval
(interval=<optimized out>) at intervals.c:482
#2  0x00000000005a1749 in find_interval (tree=<optimized out>, position=0)
at intervals.c:673
#3  0x00000000005a4d8e in validate_interval_range (object=<optimized out>,
begin=begin@entry=0x7fffffff0028, end=end@entry=0x7fffffff0028,
force=force@entry=false) at textprop.c:192
#4  0x00000000005a4ffb in Ftext_properties_at (position=<optimized out>,
position@entry=2, object=<optimized out>, object@entry=34216612) at
textprop.c:568
#5  0x00000000005a506c in Fget_text_property (position=position@entry=2,
prop=prop@entry=20256, object=object@entry=34216612) at textprop.c:588
#6  0x00000000004b750c in face_at_string_position (w=<optimized out>,
string=34216612, pos=<optimized out>, bufpos=bufpos@entry=11967,
endptr=endptr@entry=0x7fffffff0158, base_face_id=22, mouse_p=false) at
xfaces.c:6081
#7  0x000000000043e33e in handle_face_prop (it=0x7fffffff61e0) at
xdisp.c:3974
#8  0x000000000043b8da in handle_stop (it=it@entry=0x7fffffff61e0) at
xdisp.c:3371

[-- Attachment #2: Type: text/html, Size: 2797 bytes --]

bug#25282: emacs 25,1 crash w/bt

[Eli Zaretskii]

> From: Hlöðver Sigurðsson
> 	<hlolli@gmail.com>
> Date: Tue, 27 Dec 2016 12:21:32 +0100
> 
> My emacs 25.1 is crashing these days every 3 hours, irrelevant if Im
> on my laptop or work computer, and if Im programming in C or
> Clojure. Im always using Fedora, and I install emacs either by
> compiling from sources or installed from dnf. Just a paste of my last
> gdb crash backtrace: https://www.refheap.com/124399
> 
> Crashes are on my computer more likely when saveing or simply using the cursor around a document. I have
> aggressive indent mode enabled, haven't ruled that out yet. Looks to me still that something is not right in
> intervals.c?
> 
> Thanks for all feedback, hope I don't have to downgrade my emacs.
> 
> Best,
> Hlödver
> 
> (the beginning of the trace)
> Thread 1 "emacs" received signal SIGSEGV, Segmentation fault.
> 0x00000000005a0937 in balance_an_interval (i=0x685d630) at intervals.c:443
> 443 + RIGHT_TOTAL_LENGTH (i->left) - LEFT_TOTAL_LENGTH (i->left);
> (gdb) bt
> #0 0x00000000005a0937 in balance_an_interval (i=0x685d630) at intervals.c:443
> #1 0x00000000005a0ba5 in balance_possible_root_interval (interval=<optimized out>) at intervals.c:482
> #2 0x00000000005a1749 in find_interval (tree=<optimized out>, position=0) at intervals.c:673
> #3 0x00000000005a4d8e in validate_interval_range (object=<optimized out>,
> begin=begin@entry=0x7fffffff0028, end=end@entry=0x7fffffff0028, force=force@entry=false) at textprop.c:192

Please try the latest pretest of Emacs 25.2, I think this is due to
known bugs that were already fixed since v25.1 was released.

Thanks.

bug#25282: emacs 25,1 crash w/bt

[Hlöðver Sigurðsson]

[-- Attachment #1: Type: text/plain, Size: 1929 bytes --]

I compiled from master branch on Savannah (hope I spelled it right),
running on v 26.0.50.1 does not crash at all.

Thanks for that!

2016-12-27 23:52 GMT+01:00 Eli Zaretskii <eliz@gnu.org>:

> > From: Hlöðver Sigurðsson
> >       <hlolli@gmail.com>
> > Date: Tue, 27 Dec 2016 12:21:32 +0100
> >
> > My emacs 25.1 is crashing these days every 3 hours, irrelevant if Im
> > on my laptop or work computer, and if Im programming in C or
> > Clojure. Im always using Fedora, and I install emacs either by
> > compiling from sources or installed from dnf. Just a paste of my last
> > gdb crash backtrace: https://www.refheap.com/124399
> >
> > Crashes are on my computer more likely when saveing or simply using the
> cursor around a document. I have
> > aggressive indent mode enabled, haven't ruled that out yet. Looks to me
> still that something is not right in
> > intervals.c?
> >
> > Thanks for all feedback, hope I don't have to downgrade my emacs.
> >
> > Best,
> > Hlödver
> >
> > (the beginning of the trace)
> > Thread 1 "emacs" received signal SIGSEGV, Segmentation fault.
> > 0x00000000005a0937 in balance_an_interval (i=0x685d630) at
> intervals.c:443
> > 443 + RIGHT_TOTAL_LENGTH (i->left) - LEFT_TOTAL_LENGTH (i->left);
> > (gdb) bt
> > #0 0x00000000005a0937 in balance_an_interval (i=0x685d630) at
> intervals.c:443
> > #1 0x00000000005a0ba5 in balance_possible_root_interval
> (interval=<optimized out>) at intervals.c:482
> > #2 0x00000000005a1749 in find_interval (tree=<optimized out>,
> position=0) at intervals.c:673
> > #3 0x00000000005a4d8e in validate_interval_range (object=<optimized out>,
> > begin=begin@entry=0x7fffffff0028, end=end@entry=0x7fffffff0028,
> force=force@entry=false) at textprop.c:192
>
> Please try the latest pretest of Emacs 25.2, I think this is due to
> known bugs that were already fixed since v25.1 was released.
>
> Thanks.
>

[-- Attachment #2: Type: text/html, Size: 2539 bytes --]

bug#25282: emacs 25,1 crash w/bt

[Eli Zaretskii]

> From: Hlöðver Sigurðsson <hlolli@gmail.com>
> Date: Wed, 4 Jan 2017 14:47:19 +0100
> Cc: 25282@debbugs.gnu.org
> 
> I compiled from master branch on Savannah (hope I spelled it right), running on v 26.0.50.1 does not crash at
> all.
> 
> Thanks for that!

Thanks for testing, I'm therefore closing this bug.