crash on std::bad_alloc

[Alvin Penner]

running Windows XP, Inkscape rev 10610, the attached file crashes with the message:

terminate called after throwing an instance of 'std::bad_alloc'
  what(): std::bad_alloc

Emergency save activated!

this file was created with pathalongpath.py, following Bug 486920. It loads properly into Inkscape 0.48.2 although it takes about 30 seconds to load.

backtrace is:

warning: Lowest section in C:\WINDOWS\system32\odbcint.dll is .rsrc at 00401000
BFD: C:\WINDOWS\system32\WMVCore.dll: Warning: Ignoring section flag IMAGE_SCN_MEM_NOT_PAGED in section .reloc
BFD: C:\WINDOWS\system32\WMVCore.dll: Warning: Ignoring section flag IMAGE_SCN_MEM_NOT_PAGED in section .reloc
[New thread 3884.0xf88]
warning: Lowest section in C:\WINDOWS\system32\xpsp2res.dll is .rsrc at 00011000

warning: HEAP[inkscape.exe]:
warning: HEAP: Free Heap block 36832d8 modified at 3683480 after it was freed

Program received signal SIGTRAP, Trace/breakpoint trap.
0x7c90120f in ntdll!DbgUiConnectToDbg () from C:\WINDOWS\system32\ntdll.dll
(gdb) bt
#0 0x7c90120f in ntdll!DbgUiConnectToDbg ()
   from C:\WINDOWS\system32\ntdll.dll
Cannot access memory at address 0x22e6fc

[jazzynico (jazzynico)]

Crash confirmed on Ubuntu 11.04, Inkscape trunk revision 10612.
Not reproduced, but very very long to open with 0.48.1-2ubuntu2.

[jazzynico (jazzynico)]

Backtrace on Ubuntu 11.04.

[jazzynico (jazzynico)]

Crash reproduced on XP with trunk revision 10923.
Not reproduced with revision 11098 and 11859.

Needs confirmation on other systems.

[su_v (suv-lp)]

> Needs confirmation on other systems.

On OS X 10.7.4 with local 64bit builds:
- no crash with Inkscape 0.48+devel r11859
- no crash with Inkscape 0.48.x r9914, 0.48.3.1 r9886, 0.48.2 r9819

On OS X 10.7.4 with official 32bit packages, older deps:
- no crash with Inkscape 0.48.2 r9819, 0.48.1 r9760, 0.48.0 r9654, 0.47 r22583

(Note: preview in Gtk file dialog is turned off, Inkscape main window de-maximized as soon as it opens and responds to resize commands from the window manager)

[su_v (suv-lp)]

Alvin Penner wrote:
> running Windows XP, Inkscape rev 10610, the attached file crashes

JazzyNico wrote:
> Crash confirmed on Ubuntu 11.04, Inkscape trunk revision 10612.
> Crash reproduced on XP with trunk revision 10923.

Based on on the content of the drawing (stroked paths, with tightly arranged segments) and tests with archived builds on OS X 10.7.4 (64bit), the crashes and slow performance with earlier trunk builds seem related to
Bug #906952 “Low performance while manipulating spirograph paths”
<https://bugs.launchpad.net/inkscape/+bug/906952>

With 64bit builds on OS X 10.7.4 and 8GB RAM, Inkscape doesn't crash with revisions < 10835, 10935, but grabs huge amounts of memory (> 3 GB) to open the file, and for simple operations like selecting the stroked path. Current trunk opens the file rather quickly, and memory usage stays < 200 MB even when selecting and moving the large stroked path.

Possibly this earlier performance regression (due to increased precision of the visual bounding box calculations) exhausts the system resources with 32bit builds on Windows…

[Alvin Penner (apenner)]

- crash reproduced on Windows 7, Inkscape 0.48.3.1, with file preview enabled. Total memory requirements just before the crash was about 1.7 GB and rising

- crash not reproduced on the same computer with file preview disabled. Loading took 105 seconds, and a drag operation took very long also. Total memory requirements about 1.5 GB.

- crash not reproduced on Windows XP, Inkscape rev 11861. Loading took about 2 seconds and total memory requirements was about 347 MB. (343 before the load and 347 after the load). Dragging operation was significantly faster as well.

I would propose to close this report since the trunk memory requirements appear to be less and since this apparently was due to lack of memory.