[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [csmith-dev] gcc crash and csmith bug

[Repost: Whoops, meant to send this to the list, not to Eric specifically]

On Mon, Jul 18, 2011 at 6:33 PM, Eric Eide <eeide@cs.utah.edu> wrote:
>        Arthur> John, etc.: I suggest that csmith not print *any* output until
>        Arthur> it's ready to print *all* its output. That way, if it's killed
>        Arthur> by ulimit, it won't leave a partial file behind.
> Hmm.  One of my to-do's is to flush the output right after the header comment,
> to make it easier for people to report bugs that we can reproduce.  That way,
> if Csmith crashes, the identifying info is persisted.
> But I can see the value of no output, too.  Something to ponder.

I guess a trivial way to get the best of both worlds is to flush the
output right after the boilerplate header text, but make sure that the
last line of that header text is something like

/* CAUTION: If this comment is the last thing in this file, that means
Csmith crashed or was killed
 * before it got a chance to print the rest of the program. This might
be due to a timeout in someone's
 * wrapper script, or it might indicate a bug. */