[PATCH 1/2] new routines for socket based synchronisation

Petr Muller pmuller at redhat.com
Mon Nov 25 14:54:26 UTC 2013


On Wed, 2013-11-20 at 13:37 +0100, Jiri Jaburek wrote:
> On 11/19/2013 06:46 PM, Hubert Kario wrote:
> >>
> >>> ---
> >>> is the fatal error handling OK?
> >>
> >> No: by directly exiting, you are e.g. preventing any cleanups from
> >> running (and logs from uploading when run on beaker), unfortunately. My
> >> preference would be to just return a RC and let the user handle the
> >> failure e.g. with rlRun. Use journal logs for messages. It's done this
> >> way everywhere BL.
> > 
> > I'm assuming that by "RC" you mean "return code".
> > I'll change it to do rlLogFatal and "return big-n" on unrecoverable errors
> > then.
> 
> IMHO rlLogError would be better - it's used by other BeakerLib code
> (rlLogFatal being never used). Also returning 0 for success and 1 for
> fail (or 2/3/4/.. for special purposes) is widely accepted.

+1

PM



More information about the beakerlib-devel mailing list