Re: [PATCH v4] add-patch: response to unknown command

[Jeff King <peff@peff.net>]

On Thu, Apr 25, 2024 at 05:04:45AM +0200, Rubén Justo wrote:

> > The test_cmp output looks
> > like this:
> > 
> >   -(1/1) Stage addition [y,n,q,a,d,e,p,?]? Unknown command 'W' (use '?' for help)
> >   -(1/1) Stage addition [y,n,q,a,d,e,p,?]?
> >   +(1/1) Stage addition [y,n,q,a,d,e,p,?]? (1/1) Stage addition [y,n,q,a,d,e,p,?]?
> >   +Unknown command 'W' (use '?' for help)
> > 
> > which makes me suspect a race. Perhaps because the prompt is going to
> > stdout, but the "Unknown command" message goes to stderr?
> 
> I have to read the thread pointed by Eric, but my knee-jerk reaction has
> been to think in something like:
> 
> diff --git a/add-patch.c b/add-patch.c
> index 447e8166d2..0090543f89 100644
> --- a/add-patch.c
> +++ b/add-patch.c
> @@ -292,6 +292,7 @@ static void err(struct add_p_state *s, const char *fmt, ...)
>  {
>         va_list args;
> 
>         va_start(args, fmt);
> +       fflush(stdout);
>         fputs(s->s.error_color, stderr);
>         vfprintf(stderr, fmt, args);

I think the "just send it all to stdout" approach makes the most sense
here, but in case we don't do that: I don't think this will fix it. In
the output above it is the "Unknown command" output which is delayed,
which is sent to stderr via err(). So flushing stdout again won't help.
Flushing stderr after the vfprintf _might_ help (though I'm confused why
stderr would be fully buffered in the first place).

-Peff