joey@tessi.UUCP (08/30/87)
Maybe someone can answer this for me... Why doesn't isatty() clear errno before it returns? It is determining something by purposely trying to generate an error. Seems reasonable that it should then clean up after itself if it actually generated an error. Think of all the code that would stop behaving semi-irrationally if this "fix" were installed. -- Joe Pruett Test Systems Strategies ...!tektronix!tessi!joey
gwyn@brl-smoke.ARPA (Doug Gwyn ) (09/02/87)
In article <408@tessi.UUCP> joey@tessi.UUCP (Joe Pruett) writes: >Think of all the code that would stop behaving semi-irrationally if >this "fix" were installed. The code would of course still be irrational, in that it would be perror()ing when it shouldn't. The most likely symptom would then be the "No error" error message, which is even sillier.
charles@hpcvcd.HP (Charles Brown) (09/04/87)
> > Think of all the code that would stop behaving semi-irrationally if > > this "fix" were installed. > The code would of course still be irrational, in that it would be > perror()ing when it shouldn't. The most likely symptom would then > be the "No error" error message, which is even sillier. An error message should be informative, accurately revealing the true nature of the problem. In that line I suggest the above error message should be changed to read "Programmer is a twit." Charles :-) Brown hplabs!hp-pcd!charles