------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugs.exim.org/show_bug.cgi?id=1295 --- Comment #24 from Philip Hazel <[email protected]> 2012-10-20 18:54:08 --- On Sat, 20 Oct 2012, Zoltan Herczeg wrote: > I fixed the issue in study: > > #if defined SUPPORT_UTF || !defined COMPILE_PCRE8 > case OP_XCLASS: > - cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS]; > + cc += GET(cc, 1); > + cc -= PRIV(OP_lengths)[OP_CLASS]; > /* Fall through */ > #endif > > If GET(cc, 1) < PRIV(OP_lengths)[OP_CLASS], the result overflows in 64 bit > mode > (both are unsigned values). Unfortunately I still see crashes in 64 bit mode > (when 32 bit tests are running). Perhaps a similar issue. I think that is worth a comment in the code to stop somebody from "tidying" it up. And wouldn't this be neater? cc = (cc + GET(cc, 1)) - PRIV(OP_lengths)[OP_CLASS]; Philip -- Configure bugmail: http://bugs.exim.org/userprefs.cgi?tab=email -- ## List details at https://lists.exim.org/mailman/listinfo/pcre-dev
