> At the top of dmalloc.h, there's a test that determines the return type
> of all the wrapped functions. It ends up setting DMALLOC_PNT to void*
> in the C++ case. The strdup function is supposed to always return a
> char*, though. Because C++ is very picky about pointer conversion,
> this can cause compiler errors because strdup is not returning a char* .
> Does the following patch make sense?

Yes it does Dan. Like duh. Not sure why I didn't see this before.
I'll push it into the main line right now.
---
gray.eg3@...

Copyright 2003 by Gray Watson.
You may not copy, forward, transmit, publish, post or otherwise use
this work without the express written permission of Gray Watson.