On Sun, 16 May 2004, [UTF-8] Branko Ä^Libej wrote:
> Tobias Ringström wrote:
>
> > Peter N. Lundblad wrote:
> >
> >> b) Ditch the return value of the new functions in favour of void and
> >> ignore failures (hoping for SIGPIPE kicking in as we do today).
> >>
> >> Opinions?
> >
> >
> > FWIW, we currently set the SIGPIPE handler to SIG_IGN, which means
> > that we depend on getting an error return value from the write functions.
>
Why do we ignore SIGPIPE? If we do that and at the same time ignore printf
errors, we won't notice a broken pipe. So if I do
svn ls -R |less
and then quit less after one page, won't svn keep retrieving the whole
tree then?
> This is for printing progress and error messages to the command line. I
> think we should act like normal printf, FWIW, apr_vformatter returns the
> character count.
>
You mean returning an int which can be negative in case of failure?
Do we agree that the current return value of svn_error_t* is wrong?
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun May 16 23:21:15 2004