On Fri, Nov 02, 2007 at 04:12:50AM +0000, [EMAIL PROTECTED] wrote: > Author: njn > -static UInt clo_heap_admin = 8; > + // clo_heap_admin is deliberately a word-sized type. At one point it was > + // a UInt, but this caused problems on 64-bit machines when it was > + // multiplied by a small negative number and then promoted to a > + // word-sized type -- it ended up with a value of 4.2 billion. Sigh. > +static SizeT clo_heap_admin = 8; > this is really odd ... the code in update_heap_stats() did this: (unsigned long) = (unsigned long) + (unsigned int) * (signed int); while the actual multiplication is signed, the result is taken to be unsigned and is accordingly zero-extended in the promotion. you converted this to (unsigned long) = (unsigned long) + (unsigned long) * (signed int); which is about the right thing to do. however, for *way* more clarity and less bloat i'd suggest writing the actual expression as update_alloc_stats(heap_szB_delta + (SizeT)clo_heap_admin*n_heap_blocks_delta); instead of making the variable bigger.
btw, in ms_post_clo_init() you'll get a warning, because, duh, an unsigned value cannot be smaller than zero. -- Hi! I'm a .signature virus! Copy me into your ~/.signature, please! -- Chaos, panic, and disorder - my work here is done. ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Valgrind-developers mailing list Valgrind-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/valgrind-developers
