Re: Fwd: VCP error code

On Tue, May 31, 2005 at 10:44:57PM +0100, Robert Cowham wrote:
> Personally I like to use P4Perl and print the error messages array.

VCP has an option to use P4::Client, but there are a couple of issues
that arise when doing so.  Now that P4Perl is independant of that
module, perhaps we should convert the P4 API portions of VCP (see
VCP::Utils::p4) to use it.  I recall several issues with P4::Client, one
of which was for larger commands (VCP can replay some very large lists
of files for edit/submit/etc and I don't recall the P4::Client
limitation on large commands being well documented enough to work
around).  Then there was a buffer identification problem and a memory
leak when dealing with overlapped requests (I think the buffer id
problem was an issue with overlapping requests reusing the same buffer).

Does P4Perl handle large commands and overlapped requests properly?  I
haven't used it.  I'd much rather use it than the command line if it's
as bug-free as the command line.

> It's all to easy to miss things on stderr otherwise.

VCP captures all of p4's stderr and logs it; can't miss that (if we
get the logs).  It also reports it to stderr if there's an error.

- Barrie