Florian Weimer <fw@deneb.enyo.de> writes: > Current test locks at the values, which is good for comparing > O_CLOEXEC, where the token sequences differ. Sounds like I can just remove that part of the test, then.