[Sumover-tech] Exceptions when compiling VIC for Windows
mimiller at ncsa.illinois.edu
Tue Jan 25 16:51:08 GMT 2011
The error actually occurs in gs_support.c:
128 #if defined (_WIN64)
129 cookie = systime.ft_scalar;
130 #else /* defined (_WIN64) */
131 cookie = systime.ft_struct.dwLowDateTime;
132 cookie ^= systime.ft_struct.dwHighDateTime;
133 #endif /* defined (_WIN64) */
When I look in the locals pane I can see that systime.ft_struct.dwLowDateTime has a value of 3907274123 and systime.ft_struct.dwHighDateTime has a value of 4294918093.
So it would seem the values are there and accessible, but the error is:
First-chance exception at 0x0021b586 in vic.exe: 0xC0000005: Access violation reading location 0x0021b586.
Unhandled exception at 0x0021b586 in vic.exe: 0xC0000005: Access violation reading location 0x0021b586.
Hopefully this helps lead to an answer. Would the GPL version be less problematic?
217 649 0747
"If you're clear in your vision and trust the people in your team with clear objectives, they will invariably do their best to achieve everything desired, and usually deliver everything you could have hoped for and even more." -Paul Debevec
----- Original Message -----
From: "Douglas Kosovic" <doug at uq.edu.au>
To: "Michael Miller" <mimiller at ncsa.illinois.edu>
Cc: sumover-tech at cs.ucl.ac.uk
Sent: Monday, January 24, 2011 7:46:44 PM
Subject: RE: [Sumover-tech] Exceptions when compiling VIC for Windows
> when running in debug mode. The debugger takes me to the Disassembly
> window and points to:
> 021b586 ???
> which is less than unhelpful.
In the bottom right of VisiualStudio I think the backtrace pane which can be selected would probably show more useful info.
> I followed the instructions here:
> and installed DXSDK June 2010, ActiveTCL 126.96.36.199 along with the Common
> I'm on WinXPsp3 with VS2008.
> Any suggestions?
Are you building GPL vic or nonGPL vic? See Build->'Configuration Manager...' menu. If you are building GPL vic and using zip 3.0, instead of zip 2.x or 3.1, you'll get a weird error similar to what you are seeing. zip 3.0 has a bug when setting the offset of the zip file appended to an exe file.
More information about the Sumover-tech