Re: UB status of snprintf on invalid ptr+size combination?

[Siddhesh Poyarekar <siddhesh@gotplt.org>]

On 2023-03-20 13:00, Vincent Lefevre wrote:
> It doesn't. *Currently*, MPFR does not use snprintf. With the buggy
> version, on a typical 64-bit machine (where int = 32 bits), the size
> given to mpfr_snprintf became the value modulo 2^32, so if n is 2^32
> (possible as size_t has 64 bits), the implementation of mpfr_snprintf
> assumed that the size was 0 (instead of 2^32). Hence the incorrect
> behavior.
> 
> The test helped to catch this bug because it checks mpfr_snprintf
> on this value n = (size_t) UINT_MAX + 1, which is 2^32 here. But in
> order not to use much memory, the test is done with a small buffer.
> Using a buffer of size n would need 4 GB, and this amount of memory
> is not available everywhere.

In glibc we tend to skip a test as UNSUPPORTED when resources to run a 
test are considered uncommon (from the perspective of a test system) and 
cannot be expected to be met in all environments.

It looks like mpfr_snprintf and glibc snprintf are incompatible in this 
context and the latter should not be used to implement the former if the 
n > __bos(dest) use case is important to it.

Then again, I wish mpfr_snprintf also similarly tightened its 
requirements, but that's a discussion for another day.

Thanks,
Sid