You already sent this very back trace. The frame value shown in it looks
bogus, in most cases we shouldn't end up with an empty rectangle. I am
also a bit surprised about the name of the libobjc used. It is reported
as cygobjc-2.dll. This doesn't sound like the libobjc1 you claim to be
using. Most likely you made a mistake when switching to the old libobjc.
You will have to reconfigure make and recompile everything from scratch,
deleting all generated files before that.
I already explained that to you in a previous mail, but the only result
was that you left out the frame #0 in the next mail, so I could not tell
any more.
When I did that compilation on Cygwin I had a problem with gscolors.c in
back. If this issue still happens to you you wont be able to compile
back directly and will end up using the old version of back that links
with the standard libobjc of Cygwin.
I would recommend that you remove all compiled GNUstep libraries and all
the different versions of libibjc first. Then follow the standard
procedure to compile GNUstep from scratch. That is:
- make distclean; configure; make; make install for GNUstep make
- make distclean; configure; make; make install for libobjc1
- make distclean; configure; make; make install for GNUstep make again
- make distclean; configure; make; make install for GNUstep base
- make distclean; configure; make; make install for GNUstep gui
- make distclean; configure; make; make install for GNUstep back
Judging from your base test logs there are a few more issues here. For
example your temp directory gets reported as "GNUmakefile base". Could
you please check the environment settings for "TMP" and "TEMP"? This
causes some other tests to fail later on.
More worrying is that the NSBundle mechanism isn't working correctly for
you. With that broken, we should not be surprised to see more complex
applications failing.
I also noticed that your user name and therefore the home directory as
reported by NSHomeDirectory() contain spaces. Perhaps you should start
anew with a user that has a simple user name?
We also don't seem to be able to get the proper time zone information.
Could you please check whether the required files exist and what their
contents is?
As for the hanging socket client server test, you may have to comment
this out for Cygwin for now.
Hope this helps
Fred
On 01.08.2011 21:01, Riccardo Mottola wrote:
Hi,
I don't know if I posted it already or if it is a variation, this is the
current error I get when starting Ink on cygwin, using our current
libobjc1 runtime slightly modified as Fred suggested.
Program received signal SIGSEGV, Segmentation fault.
0x68549db9 in objc_msg_lookup () from C:\cygwin\bin\cygobjc-2.dll
(gdb) bt
#0 0x68549db9 in objc_msg_lookup () from C:\cygwin\bin\cygobjc-2.dll
#1 0x6323dde1 in -[NSView setFrame:] (self=0x6ab768, _cmd=0x633ae3d0,
frameRect={origin = {x = 0, y = 0}, size = {width = 0, height = 0}})
at NSView.m:1225
#2 0x632b7fe7 in -[GSWindowDecorationView layout] (self=0x695020,
_cmd=0x633ae3b0) at GSWindowDecorationView.m:341
#3 0x630e0f67 in -[NSApplication(Private) _appIconInit] (self=0x60eef8,
_cmd=0x632d20f0) at NSApplication.m:3799
#4 0x630e2c55 in -[NSApplication _init] (self=0x60eef8, _cmd=0x632d2100)
at NSApplication.m:921
#5 0x678530b2 in -[NSObject(NSThreadPerformAdditions)
performSelectorOnMainThre
ad:withObject:waitUntilDone:modes:] (self=0x60eef8, _cmd=0x67937f38,
aSelector=0x632d2100, anObject=0x60eef8, aFlag=<value optimized out>,
anArray=0x614330) at NSThread.m:1199
#6 0x67854953 in -[NSObject(NSThreadPerformAdditions)
performSelectorOnMainThre
ad:withObject:waitUntilDone:] (self=0x60eef8, _cmd=0x632d2108,
aSelector=0x632d2100, anObject=0x60eef8, aFlag=1 '\001')
at NSThread.m:1210
#7 0x630e1732 in -[NSApplication init] (self=0x60eef8, _cmd=0x632d1e70)
at NSApplication.m:975
#8 0x630d9e92 in +[NSApplication sharedApplication] (self=0x632cf940,
_cmd=0x632c7d88) at NSApplication.m:847
#9 0x630c206d in NSApplicationMain (argc=1, argv=0x4c9a58) at
Functions.m:76
#10 0x004011d0 in main (argc=5242960, argv=0x410044, env=0x410054)
at main.m:33
(gdb)
a clue about what that could be? Where to look at ?
Riccardo
_______________________________________________
Gnustep-dev mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/gnustep-dev
