Re: cvs exit status

[Paul Sander]

>--- Forwarded mail from address@hidden

>[ On Wednesday, October 10, 2001 at 10:57:50 (-0700), Paul Sander wrote: ]
>> Subject: Re: cvs exit status
>>
>> Under some error conditions, the exit status gets incremented rather than
>> being set to a fixed value.  That means that under those error conditions,
>> the exit status varies and may sometimes indicate success (due to overflow)
>> if there are a lot of problems.
>> 
>> So, exit status is not a reliable way to detect problems.  You must parse
>> the CVS standard output and standard error streams for relevant messages.

>What's wrong with using relational tests?!?!?!?  Don't you do that
>anyway just as a matter of careful robust programming?

What happens on a Unix system when the exit status exceeds 127?  It overflows.
An incremented exit status can (and does) report success in the presence of
failures.

>--- End of forwarded message from address@hidden