http://bugs.winehq.org/show_bug.cgi?id=6677
------- Additional Comments From the3dfxdude@gmail.com 2006-13-11 12:20 ------- Sam, there is a bug number 1789 for problems with float conversions. Yes, the method we use for float is rather brain dead and clearly has bugs. For problems with integers, I'd like to see your test case.
Additionally, when extra memory is allocated, pf_vsnprintf uses the pointer to the static buffer instead of the dynamically allocated one in the HeapFree call.
No it should use a dyanmic buffer when the field size is greater than 40. FieldLength should match the number of actual digit outputs at the end of conversion I believe. I should have created test cases against precision at least (with I64), but if it does not match what you are talking about could you create one to show a buffer overrun problem?