On 13 Jan 2001 [EMAIL PROTECTED] wrote:
> - Memory leak! We're loosing memory, apparently when moving from one scene
> to another. Lots of it, too (maybe it's just pic_buffer_size != 0 being
> interpreted as "infinite", but I don't think so).
In the heap, too. Hoyle (which is SCI0) sometimes dies with memory
problems and leave me in the debugger. When it doesn't (this seems to be
random), it still pseudo-hangs (though the main loop is still running), as
always (I still need to track that down).
>
> Reliably finding the memory leak(s) will require slighly modified versions
> of malloc() and friends, unless somebody already has a good tool for that.
>
> This one is release critical.
I also get a segfault on startup if I set the interpreter version in the
config file:
#0 gfxr_interpreter_get_resources (type=GFX_RESOURCE_TYPE_VIEW,
version=629,
entries_nr=0xbfffc810) at sci_resmgr.c:172
#1 0x8084604 in gfxr_new_resource_manager (version=629,
options=0x80b8de0,
driver=0x80b7be0) at resmgr.c:59
#2 0x80801de in _gfxop_init_common (state=0x80b8d40, options=0x80b8de0)
at operations.c:506
#3 0x808044c in gfxop_init (state=0x80b8d40, xfact=1, yfact=1,
bpp=GFX_COLOR_MODE_TRUE, options=0x80b8de0) at operations.c:573
#4 0x804ca54 in init_gfx (cl_options=0xbffff924, version=629,
driver=0x80b7be0) at main.c:533
#5 0x804ce6e in main (argc=4, argv=0xbffff9ac) at main.c:682
#6 0x403e0dcc in __libc_start_main () from /lib/libc.so.6
169 int top = sci_max_resource_nr[version] + 1;
It seems that the above line should use the SCI_VERSION_* macros to get it
right.
> -- Christoph
Lars
