Re: [PATCH 2/2] ref-filter: implement "quick" formats

From: Jeff King <peff@peff.net>
Date: Sep 5, 2021, 13:07 UTC
Message-ID: <YTTBCWlsC1cBZMio@coredump.intra.peff.net>
In-Reply-To: <CAOLTT8QYe3PBPxSH8CYY+FatSfT7C5m6nccR2xMZ1yxSDFh5OQ@mail.gmail.com>

On Sun, Sep 05, 2021 at 04:20:07PM +0800, ZheNing Hu wrote:

Show 15 quoted lines

> > +       case REF_FORMAT_QUICK_OBJECTNAME_REFNAME:
> > +               printf("%s %s\n", oid_to_hex(oid), refname);
> > +               return 0;
> > +       }
> > +       BUG("unknown ref_format_quick value: %d", format->quick);
> > +}
> > +
> 
> So as a fast path, we actually avoided format_ref_array_item() when we are using
> %(objectname) and %(refname). But the problem is that it’s not very elegant
> (using string compare), and it is no optimization for other atoms that
> require in-depth
> parsing. I remember the "fast path" used by Ævar last time, and it
> seems that Junio doesn't
> like them. [1][2]

Yes, I did say it was "pretty horrible". :)

It was mostly meant as a proof-of-concept to see where the time was going, and what was possible. It _could_ be used as a stop-gap while improving the general performance, but it's gross enough that it's probably not a good idea (it's increased maintenance, but also it dis-incentivizes fixing the real problems).

-Peff