I have already fixed this problem for the CVS HEAD. "Revision 1.103 2006/10/22 17:48:24 tamas Changed OGRLayerGetAutoStyle not to pass NULL pointer to GetRGBFromString causing access violation."
However it have not been backported to 4.10 yet. Tamas 2007/3/2, Alessandro Pasotti <[EMAIL PROTECTED]>:
Il 16:46, venerdì 2 marzo 2007, Umberto Nicoletti ha scritto: > Did you reconfigure with --enable-debug? > > After that run shp2img in gdb (or xxgdb if it still exists) and put a > breakpoint in msOGRLayerGetAutoStyle. Then proceed with the step > instruction until it crashes. You should be able to inspect the local > variables and see what values were passed to GetRGBFromString. Ok, I've found the problem: http://mapserver.gis.umn.edu/bugs/show_bug.cgi?id=1950 I applied this patch and now it works, any chance it will be fixed in future releases? linux:~/webgis # diff mapogr.cpp mapserver-4.10.1/mapogr.cpp 2318a2319,2326 > // ABP: patch > const char *pszColor; > pszColor = poBrushStyle->BackColor(bIsNull); > if (!bIsNull && poBrushStyle->GetRGBFromString(pszColor,r,g,b,t)) > { > MS_INIT_COLOR(c->styles[0].backgroundcolor, r, g, b); > } > /* 2324a2333 > */ Still remains a mistery to me why it works without patches on my local machine. Thanks Umberto for pointing me in the right direction. -- Alessandro Pasotti itOpen - "Open Solutions for the Net Age" w3: www.itopen.it Linux User# 167502
