Re: [PATCH 2/3] archive: implement protocol v2 archive command

From: Jonathan Nieder <jrnieder@gmail.com>
Date: Sep 14, 2018, 06:05 UTC
Message-ID: <20180914060552.GD219147@aiede.svl.corp.google.com>
In-Reply-To: <87k1npkzh5.fsf@evledraar.gmail.com>

Ævar Arnfjörð Bjarmason wrote:

> On Wed, Sep 12 2018, Stefan Beller wrote:

Show 6 quoted lines

>> Would asking for a setlocale() on the server side be an unreasonable
>> feature request for the capabilities (in a follow up patch, and then not
>> just for archive but also fetch/push, etc.)?
>
> This would be very nice to have, but as you suggest in some follow-up
> change.

Indeed, I think we've gone pretty far afield from the goal of this patch series.

> I think though that instead of doing setlocale() it would be better to
> pass some flag saying we're operating in a machine-readable mode, and
> then we'd (as part of the protocol defintion) say we're going to emit
> GIT_ERR_UPLOAD_ARCHIVE_EXPECTED_DELIM_PACKET or whatever.

I think you're suggesting client-side message generation, and that is one way to handle internationalization of server output.

The main downside is when the server really does want to provide a custom error message. For that, we'd need

 1. To propagate LANG to the server, so it knows what human language
    to generate messages in.

 2. On the server side, to produce messages in that language if
    available, with an appropriate fallback if not.

We've been thinking of doing at least (1) using the same trick as server-options use (cramming it into client capabilities).

It is difficult to use setlocale for this because it affects the whole program (problematic for a threaded server) and affects features like collation order instead of just message generation (problematic for many things). Does gettext have a variant that takes a locale_t argument?

[...]

Show 5 quoted lines

>  4) Aside from translation purposes, getting a machine-readable
>     "push/pull" etc. mode would be very handy. E.g. now you need to
>     parse stderr to see why exactly your push failed (hook denied, or
>     non-fast-forward, or non-fast-forward where there was a lock race
>     condition? ...).

Indeed, this is a good reason to provide error codes instead of (in the case where the message doesn't add anything to it) or alongside (in case the error message is more specialized) human-oriented error messages.

Thanks, Jonathan