Daniel Shahaf wrote on Fri, 17 Oct 2008 at 20:14 +0200:
> Stefan Sperling wrote on Fri, 17 Oct 2008 at 18:40 +0100:
> > On Fri, Oct 17, 2008 at 05:57:30PM +0200, Daniel Shahaf wrote:
> > > > + if (nb->had_print_error)
> > > > + return SVN_NO_ERROR;
> > >
> > > The docstring of 'had_print_error' says it's for error messages, not for
> > > informational messages.
> >
> > That's not how it's being used.
> > It's also used for messages like _("--- Merging differences between "
> > "repository URLs into '%s':\n"),
> >
>
> Are we talking about the same thing? had_print_error in notify.c is
> used in exactly one place in the code -- in the 'print_error' label:
>
In other words, the other informational messages do try to print
themselves even when had_print_error == TRUE. It is only the error from
printf() that is suppressed.
(I'm just saying that it's inconsistent, not that your new code should
be fixed.)
Daniel
> print_error:
> /* If we had no errors before, print this error to stderr. Else, don't print
> anything. The user already knows there were some output errors,
> so there is no point in flooding her with an error per notification. */
> if (!nb->had_print_error)
> {
> nb->had_print_error = TRUE;
> svn_handle_error2(err, stderr, FALSE, "svn: ");
> }
> svn_error_clear(err);
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-10-17 20:22:37 CEST