Re: [Inkscape-devel] PDF import crash, but not in 0.46
Yes, I believe that all of the patches made it into the trunk. However, I don't think that any of them applied to Poppler. This sounds like simply a typical common regression here. ^^
We haven't done a "debug" upload in a few weeks. Time to do another one, I guess.
bob
-----Original Message-----
From: J.B.C.Engelen@...1578... Sent: May 14, 2008 8:10 AM To: inkscape-devel@lists.sourceforge.net Subject: [Inkscape-devel] PDF import crash, but not in 0.46
Hi all,
PDF import crashes for me on Windows with current trunk. It does not crash in the Inkscape 0.46 binary that I just downloaded.
Have all windows patches been applied to trunk?
Thanks, Johan
This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Bob Jamison-2 wrote:
Yes, I believe that all of the patches made it into the trunk. However, I don't think that any of them applied to Poppler. This sounds like simply a typical common regression here. ^^
We haven't done a "debug" upload in a few weeks. Time to do another one, I guess.
I reported this in the bug mentioned above - https://bugs.launchpad.net/inkscape/+bug/229524
There's a full backtrace in it, as well as the stderr output using inkcl.bat, so hopefully that helps someone with more programming nouse than me.
Johan, all the windows patches from 0.46 were applied to trunk before the 0.46 release. They haven't been integrated into the 0.46 branch because of Linux distro issues, but are all sitting obediently in RELEASE_0_46_BRANCH/packaging/win32/patches - there is an 0.46 folder for the ones that went in the release, and an 0.46.1 folder for those I hope will go into the .1 release sometime soon.
I've had a quick look at the SVN records for trunk/src/extension/internal/ - things like pdf-input-cairo.cpp and stuff in the pdfinput folder, but I couldn't find any obvious recent changes that might have caused this problem (but I'm just a c++ noob!).
Some other options that I haven't tried are to go back through the dev libs versions one at a time and rebuild each time, and also to build SVN revisions going back in leaps of about 50 commits per jump and see when it starts mucking up. If you're keen, that's two good starting points.
Cheers,
Rygle.
I pasted that backtrace below. If you look at it, you will see that the error appears to be in poppler's code. So it might be good to alert those guys to this, too.
But what is curious, is this: Error: Couldn't open 'nameToUnicode' file 'C:\target\share\poppler\nameToUnicode\a'
Is Poppler looking in the $PREFIX of its install for that resource? What should be the result of not finding it?
bob
rygle wrote:
Bob Jamison-2 wrote:
Yes, I believe that all of the patches made it into the trunk. However, I don't think that any of them applied to Poppler. This sounds like simply a typical common regression here. ^^
We haven't done a "debug" upload in a few weeks. Time to do another one, I guess.
I reported this in the bug mentioned above - https://bugs.launchpad.net/inkscape/+bug/229524
There's a full backtrace in it, as well as the stderr output using inkcl.bat, so hopefully that helps someone with more programming nouse than me.
Program received signal SIGSEGV, Segmentation fault. 0x70cc1559 in GfxPath::moveTo (this=0x0, x=21.6000000000
00001, y=18) at GfxState.cc:3746 3746 GfxState.cc: No such file or directory. in GfxState.cc Current language: auto; currently c++ (gdb) bt #0 0x70cc1559 in GfxPath::moveTo (this=0x0, x=21.600000000000001, y=18) at GfxState.cc:3746 #1 0x00703676 in PdfParser::opRectangle () #2 0x007004b5 in PdfParser::go () #3 0x0070089b in PdfParser::parse () #4 0x0052940d in _fu2__globalParams () #5 0x06f81c68 in ?? () #6 0x0022edc8 in ?? () #7 0x00000001 in ?? () #8 0x00000006 in ?? () Cannot access memory at address 0x3fe00004 (gdb) c Continuing.
Program received signal SIGSEGV, Segmentation fault. 0x70cc1559 in GfxPath::moveTo (this=0x0, x=21.600000000000001, y=18) at GfxState.cc:3746 3746 in GfxState.cc (gdb) bt #0 0x70cc1559 in GfxPath::moveTo (this=0x0, x=21.600000000000001, y=18) at GfxState.cc:3746 #1 0x00703676 in PdfParser::opRectangle () #2 0x007004b5 in PdfParser::go () #3 0x0070089b in PdfParser::parse () #4 0x0052940d in _fu2__globalParams () #5 0x06f81c68 in ?? () #6 0x0022edc8 in ?? () #7 0x00000001 in ?? () #8 0x00000006 in ?? () Cannot access memory at address 0x3fe00004 (gdb) c Continuing.
Program exited with code 030000000005. (gdb) ============
Using inkcl to catch the stderr output I get this; ======= return code: -1073741819 Error: Couldn't open 'nameToUnicode' file 'C:\target\share\poppler\nameToUnicode \a' =======
participants (2)
-
Bob Jamison
-
rygle