Re: RFC GSoC idea: new "git config" features

From: Jeff King <peff@peff.net>
Date: Mar 14, 2014, 04:43 UTC
Message-ID: <20140314044326.GE31906@sigill.intra.peff.net>
In-Reply-To: <vpq38j2tc6f.fsf@anie.imag.fr>

On Sat, Mar 01, 2014 at 12:01:44PM +0100, Matthieu Moy wrote:

Show 11 quoted lines

> Jeff King <peff@peff.net> writes:
> 
> > If we had the keys in-memory, we could reverse this: config code asks
> > for keys it cares about, and we can do an optimized lookup (binary
> > search, hash, etc).
> 
> I'm actually dreaming of a system where a configuration variable could
> be "declared" in Git's source code, with associated type (list/single
> value, boolean/string/path/...), default value and documentation (and
> then Documentation/config.txt could become a generated file). One could
> imagine a lot of possibilities like

Yes, I think something like that would be very nice. I am not a big fan of code generation, but if we had config queries like "config_get_bool", then I think it would be reasonably pleasant to take a spec like:

  Key: help.browser
  Type: string
  Description: Specify the browser for help...

and turn it into:

  const char *config_get_help_browser(void)
  {
          return config_get_string("help.browser");
  }

So technically code generation, but all the heavy lifting is done behind the scenes. We're not saving lines in the result so much as avoiding repeating ourselves (that is, the generated code is only mapping the config-type from the spec into a C type and function name that gives us extra compile-time safety).

However, I skimmed through config.txt looking for a key to use in my example above, and there are a surprising number of one-off semantics (e.g., things that are mostly bool, but can be "auto" or take some other special value). We may find that the "Type" field has a surprising number of variants that makes a technique like this annoying. But I'd reserve judgement until somebody actually tries encoding a significant chunk of the config keys and we see what it looks like.

> Migrating the whole code to such system would take time, but creating
> the system and applying it to a few examples might be feasible as a GSoC
> project.

Agreed, as long as we have enough examples to feel confident that the infrastructure is sufficient.

-Peff