Bugs item #1107922, was opened at 2005-01-23 11:51
Message generated for change (Settings changed) made by keescook
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=604306&aid=1107922&group_id=93438
Category: Node Editing
Group: None
Status: Closed
Resolution: Fixed
Priority: 9
Submitted By: Alexander Clausen (sk1p)
>Assigned to: Kees Cook (keescook)
Summary: all drawing tools crash
Initial Comment:
to reproduce:
- press shift+f6 to activate straight lines tool.
- make a line by clicking two times somewhere on the
canvas (diff. places)
- press and hold ctrl and click on the initial node
=> crash
version: latest cvs
bt:
Inkscape::XML::SimpleNode::synthesizeEvents
(this=0x1060a110, vector=0x0, data=0x10878310) at
xml/simple-node.cpp:398
398 if (vector->attr_changed) {
(gdb) bt
#0 Inkscape::XML::SimpleNode::synthesizeEvents
(this=0x1060a110, vector=0x0, data=0x10878310) at
xml/simple-node.cpp:398
#1 0x10107008 in sp_desktop_get_style
(desktop=0x10878310, with_text=false) at
desktop-style.cpp:166
#2 0x10107134 in sp_desktop_apply_style_tool
(desktop=0x1060a090, repr=0x11341168, tool=0x102a6568
"tools.freehand.pen",
with_text=false) at desktop-style.cpp:198
#3 0x1014b754 in spdc_flush_white (dc=0x113420f8,
gc=0x7fffed50) at draw-context.cpp:547
#4 0x1014b1c4 in spdc_concat_colors_and_flush
(dc=0x113420f8, forceclosed=277316368) at
draw-context.cpp:489
#5 0x10129f84 in spdc_pen_finish (pc=0x113420f8,
closed=288625000) at pen-context.cpp:767
#6 0x10129340 in pen_handle_button_release
(pc=0x113420f8, revent=@0x4036db6d) at pen-context.cpp:539
#7 0x10128a54 in sp_pen_context_root_handler
(ec=0x113420f8, event=0x11317a48) at pen-context.cpp:235
#8 0x1011566c in sp_event_context_root_handler
(event_context=0x7fffee30, event=0x7fffee50) at
event-context.cpp:687
#9 0x10104470 in sp_desktop_root_handler
(item=0x1060a090, event=0x1, desktop=0x0) at
desktop-events.cpp:56
#10 0x10206aec in sp_marshal_BOOLEAN__POINTER
(closure=0x0, return_value=0x7ffff200,
n_param_values=0, param_values=0x112e6ff8,
invocation_hint=0x0, marshal_data=0x0) at
helper/sp-marshal.cpp:350
#11 0x0f1c86a8 in g_closure_invoke () from
/usr/lib/libgobject-2.0.so.0
#12 0x0f1c86a8 in g_closure_invoke () from
/usr/lib/libgobject-2.0.so.0
[...]
#36 0x0f1c86a8 in g_closure_invoke () from
/usr/lib/libgobject-2.0.so.0
#37 0x0f1c86a8 in g_closure_invoke () from
/usr/lib/libgobject-2.0.so.0
Previous frame inner to this frame (corrupt stack?)
----------------------------------------------------------------------
Comment By: Alexander Clausen (sk1p)
Date: 2005-02-02 08:05
Message:
Logged In: YES
user_id=934032
this one is fixed for me on both amd64 and ppc. good job!
----------------------------------------------------------------------
Comment By: bulia byak (buliabyak)
Date: 2005-02-02 00:17
Message:
Logged In: YES
user_id=741217
Ah, sorry, I see you (maybe) fixed this one, not the crash
on load. Now sp_desktop_get_style works OK, and I hope it
also fixes the crash.
----------------------------------------------------------------------
Comment By: bulia byak (buliabyak)
Date: 2005-02-01 23:58
Message:
Logged In: YES
user_id=741217
Kees, I'm confused. Are you commenting on the crash-on-load
bug? If so then please comment in that bug, not here:
https://sourceforge.net/tracker/index.php?func=detail&aid=1113187
&group_id=93438&atid=604306
----------------------------------------------------------------------
Comment By: Kees Cook (nemies)
Date: 2005-02-01 23:53
Message:
Logged In: YES
user_id=26678
This bug is related to the missing default attrs, actually.
Can you try this against current CVS? Attrs are working
for me again now after some fixups.
----------------------------------------------------------------------
Comment By: bulia byak (buliabyak)
Date: 2005-02-01 07:54
Message:
Logged In: YES
user_id=741217
Kees: I found that one yesterday and submitted it
separately, and even wrote about it in the list:
https://sourceforge.net/tracker/index.php?func=detail&aid=1113187&group_id=93438&atid=604306
It's also assigned to Mental but I think should be simpler
to fix...
----------------------------------------------------------------------
Comment By: Kees Cook (nemies)
Date: 2005-02-01 07:51
Message:
Logged In: YES
user_id=26678
Jimmac found another bug, opening this file:
http://jimmac.musichall.cz/stuff/fspot-logo.svg
crashes in a similar place.
** ERROR **: file xml/simple-node.h: line 38 (virtual void
Inkscape::XML::SimpleNode::setCodeUnsafe(int)): assertion
failed: (!_logger && !_listeners)
aborting...
----------------------------------------------------------------------
Comment By: Alexander Clausen (sk1p)
Date: 2005-01-31 09:46
Message:
Logged In: YES
user_id=934032
seems to be caused by this commit by mental:
[... lots of files ...]
got rid of xml/repr-private.h, removed
sp_repr_document_createTextNode(), made SPRepr and
SPReprDoc abstract
interfaces, and moved the SPRepr implementations into
separate files
in the Inkscape::XML namespace
----------------------------------------------------------------------
Comment By: Alexander Clausen (sk1p)
Date: 2005-01-31 08:59
Message:
Logged In: YES
user_id=934032
ok, tracked it down to between "2005-01-23 20:00" and
"2005-01-23 21:00"
----------------------------------------------------------------------
Comment By: Alexander Clausen (sk1p)
Date: 2005-01-31 06:07
Message:
Logged In: YES
user_id=934032
ok, i'll try.
----------------------------------------------------------------------
Comment By: bulia byak (buliabyak)
Date: 2005-01-28 17:39
Message:
Logged In: YES
user_id=741217
sk1p: can you please test CVS on various dates to find the
point at which it started crashing? In my experience it's
the easiest way to fix bugs. If you can narrow it to a
single commit it will be easy to see what broke.
----------------------------------------------------------------------
Comment By: MenTaLguY (mental)
Date: 2005-01-28 17:31
Message:
Logged In: YES
user_id=791
It looks like the backtrace may be lying to us;
sp_desktop_get_style() never calls synthesizeEvents(), and
the corrupt stack warning would also suggest that.
Probably sp_desktop_get_style() is involved, but what we
really need is someone to try to replicate the bug in a
build without optimization; if they can, then I need them to
step through it in gdb and note where it is really failing.
----------------------------------------------------------------------
Comment By: Alexander Clausen (sk1p)
Date: 2005-01-28 02:26
Message:
Logged In: YES
user_id=934032
sorry, forgot to log in: the message below is by me, too.
some versions:
gcc 3.3.5-1
libgtk2.0-0 2.4.14-2
libglib2.0-0 2.6.1-3
libgtkmm-2.4-1 2.4.8-1
libglibmm-2.4-1 2.4.5-1
please shout if you need to know any other versions
----------------------------------------------------------------------
Comment By: Nobody/Anonymous (nobody)
Date: 2005-01-28 01:58
Message:
Logged In: NO
this happens on ppc and on amd64 for me. it might be a
problem with the debian/unstable toolchain/libs/...
----------------------------------------------------------------------
Comment By: bulia byak (buliabyak)
Date: 2005-01-27 21:55
Message:
Logged In: YES
user_id=741217
This bug was just filed:
https://sourceforge.net/tracker/index.php?func=detail&aid=1111232&group_id=93438&atid=604306
Exactly the same crash on Linux/PPC. So, Mental, this bug is
for real, and it's obviously a stopper, and I assign it to
you since it happens in your code. Please have a look.
sk1p: are you using PPC also? If not, what
architecture/OS/compiler/libraries?
----------------------------------------------------------------------
Comment By: MenTaLguY (mental)
Date: 2005-01-26 22:26
Message:
Logged In: YES
user_id=791
synthesizeEvents shouldn't ever be getting called with a
NULL vector parameter
----------------------------------------------------------------------
Comment By: bulia byak (buliabyak)
Date: 2005-01-26 16:47
Message:
Logged In: YES
user_id=741217
sk1p: can you try today's CVS? If you still see it, can you
try timed checkouts (cvs update -D "date") to find out exact
point in time when it started doing that, and which commit
broke it?
----------------------------------------------------------------------
Comment By: Alexander Clausen (sk1p)
Date: 2005-01-23 14:55
Message:
Logged In: YES
user_id=934032
same on a fresh checkout/build
----------------------------------------------------------------------
Comment By: Alexander Clausen (sk1p)
Date: 2005-01-23 14:01
Message:
Logged In: YES
user_id=934032
strange, i'll retry a fresh checkout. this is on linux/ppc
btw., but happens on linux/amd64, too
----------------------------------------------------------------------
Comment By: bulia byak (buliabyak)
Date: 2005-01-23 13:55
Message:
Logged In: YES
user_id=741217
I don't see it with latest cvs.
----------------------------------------------------------------------
Comment By: Alexander Clausen (sk1p)
Date: 2005-01-23 13:30
Message:
Logged In: YES
user_id=934032
ok, same seems to happen when creating circles, rects, etc.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=604306&aid=1107922&group_id=93438
|
