[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: client help messages

From: Karl Fogel <kfogel_at_collab.net>
Date: 2001-06-26 17:12:06 CEST

Kieran Barry <kieran@esperi.demon.co.uk> writes:
> I'd think that the options 1 will turn to be easier, and more reliable.

Hmmm. Could you describe your proposal in more detail? I don't quite
understand it, and wouldn't expect such a change to be so
far-reaching...

>From what I can see, only the print_command_info() function in
help-cmd.c needs to know whether to send to STDERR or STDOUT.
Presumably it would take an argument telling it where to print to, and
its callers would would pass that argument down, but that doesn't
imply changing the prototypes of most svn_cl__*() functions.

We'd only need to add a parameter to svn_cl__subcommand_help(), which
is one of two callers of print_command_info(). The other is
print_generic_help(), which is called from svn_cl__help(). Neither of
the latter two need their prototypes changed, it's simply that
print_generic_help() would need to pass the new argument to
print_command_info().

Is there more? I may be missing something, haven't studied this as
much as you have, probably...

-Karl

Kieran Barry <kieran@esperi.demon.co.uk> writes:
> On 19 Jun 2001, Karl Fogel wrote:
> > "B. W. Fitzpatrick" <fitz@red-bean.com> writes:
> > > Well, I for one often find myself doing this in a big CVS working
> > > copy:
> > >
> > > $ cvs up -d > delta.txt
> > >
> > > and then I look at delta.txt for added/modded files, etc. I think I'd
> > > prefer errors to go to STDERR so that I would see them right away.
> >
> > Bingo, thank you. I understand now.
> >
> > +1 on Michael's proposal:
> >
> > > exactly,
> > > svn --nosuchcommand -> STDERR
> > > svn --help -> STDOUT
> >
> > with the former exiting with non-zero code, and the latter with zero,
> > of course. Have made new issue #402, so we don't forget about this.
> > I'll do it as soon as get a chance, if no one else beats me to it
> > (please, someone beat me to it).
> >
> I've had a look, and this looks do-able. (I've been looking for
> something trivial and boring to hack at.)
>
> A question: how to do?
>
> A number of approaches appear viable:
> 1. Add a parameter to the svn_cl__* functions, either an int to be
> used as a flag, a FILE * (and change the printf calls to fprintf),
> or an apr_file_t *, and change the printf calls to apr_file_printf.
>
> 2. Wade through the logic of all the function calls, and work out the
> correct stream to write to. I'm not sure that this is going to be
> easy.
>
> I'd think that the options 1 will turn to be easier, and more reliable.
>
> Regards
>
> Kieran

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:36:32 2006

This is an archived mail posted to the Subversion Dev mailing list.