On Fri, 29 Sep 2006 14:40:04 -0600 Theo de Raadt <[EMAIL PROTECTED]> wrote:
> > On 9/28/06, Paul Stoeber <[EMAIL PROTECTED]> wrote: > > > Should a userland program be allowed to depend on errno==0 on entry > > > to main()? (At least one in the tree does.) > > > > checking errno is the wrong way to ascertain whether a function had a > > problem, so this shouldn't be a problem for a well written program. > > > > on the other hand, code running before main is outside the scope of C > > standard, and i think the standard at least implies errno should be 0 > > on entry, so we should fix/verify that too. > > there is an even easier way to look at this: > > errno is only valid if something indicates that you should > be checking it > > some functions are documented & defined so that when they return NULL, > errno is set. > > some functions are documented & defined so that when they return -1, > errno is set. > > if a function has not indicated that errno is now valid, then errno contains > some previous result code which is absolutely irrelevant. > > so fundamentally, the question being asked by the original poster is > "can i check what is in errno whenever i want" has a simple answer: no. > > therefore it is irrelevant what value is currently stored in errno. > > if nothing told you to look at it, don't look at it. I guess that means the answer to my original question is No. I've sent a diff to [EMAIL PROTECTED] Note especially usr.bin/head/head.c
