*cabe logging?

[Tim Fischer]

Great - I'll probably look at doing this.  All I really need is logging 
and this will give it to me.  I'd like to stay away from the code 
management application - just one more thing for us to work on & maintain.

Thanks to all who responded!

Tim Fischer




Fairlight wrote:

>Simon--er, no...it was Tim Fischer--said:
>  
>
>>Is there a way to log when someone opens *cabe and when they save it?  
>>Is there anyway to activate some kind of "checkout" policy with *cabe so 
>>that no 2 people can edit the same processing table?  If no to 
>>both/either of these, is this a planned addition?
>>    
>>
>
>It doesn't do it itself, but you could rename it and stick it inside a
>wrapper.  The wrapper would just log the username, date, time, and tty
>of the person using it and log which table was being worked on, open
>a semaphore file, and try to flock() it.  If it can get a lock on the
>semaphore for *cabe on the requested file (ie, grab that from the command
>line arguments as well), then it goes ahead and launches the real *cabe,
>since renamed.  On exit from the real *cabe, clear down the lock and exit,
>optionally perhaps logging the end time of use for easy reading if you
>wanted.
>
>It would probably be best to use non-blocking flock() so that you can
>bail programmatically after too long a time without having to hit INTR or
>similar.
>
>This is pretty trivial stuff to do.
>
>mark->
>  
>