Re: [RFC PATCH 4/7] fetch: --stdin

From: Jonathan Tan <jonathantanmy@google.com>
Date: Aug 7, 2020, 21:10 UTC
Message-ID: <20200807211028.2736644-1-jonathantanmy@google.com>
In-Reply-To: <xmqqft90omsw.fsf@gitster.c.googlers.com>

Show 20 quoted lines

> > @@ -1741,6 +1745,13 @@ static int fetch_one(struct remote *remote, int argc, const char **argv, int pru
> >  		}
> >  	}
> >  
> > +	if (use_stdin_refspecs) {
> > +		struct strbuf line = STRBUF_INIT;
> > +		while (strbuf_getline_lf(&line, stdin) != EOF)
> > +			refspec_append(&rs, line.buf);
> > +		strbuf_release(&line);
> > +	}
> 
> This will use refspecs both from the command line and the standard
> input by appending?  IOW, these refspecs that came from the standard
> input are treated otherwise identically to those that came from the
> command line?
> 
> I do not particularly care whether it is "append to command line" or
> "replace command line", as I do not think it makes much difference
> in usability.  Just wanted to be sure you coded the behaviour you
> wanted.

Yes, except that I didn't plan to support the "tag foo" format. (My aim with this is just to allow "git fetch" to take large numbers of refspecs, because when we lazy fetch, the number of objects we fetch might be large.)

Show 18 quoted lines

> > @@ -1849,6 +1860,10 @@ int cmd_fetch(int argc, const char **argv, const char *prefix)
> >  			die(_("--filter can only be used with the remote "
> >  			      "configured in extensions.partialclone"));
> >  
> > +		if (stdin_refspecs)
> > +			die(_("--stdin can only be used when fetching "
> > +			      "from one remote"));
> 
> Is that only because you happened to have implemented the reading in
> fetch_one() that is designed to be called once per remote?  
> 
> You could read them here to a refspec for everybody, and then pass a
> pointer to that refspec as the extra parameter to fetch_one(), and
> fetch_one() can use that by duplicating and appending to its "rs",
> if we wanted to, no?  I do not know how important to support such a
> use case, though.  It just feels a bit of shame if this restriction
> is purely imposed by the implementation, when lifting the refstiction
> does not seem too involved.

Yes, and I only implemented the reading in fetch_one() because fetch_multiple() does not read additional refspecs from the command-line (it does not take "argv"). Looking at the code, this seems to be on purpose - there is the error message "fetch --all does not make sense with refspecs", and when --multiple is set, all args are assumed to be remotes.