Re: [PATCH] git-send-email: add ~/.authinfo parsing

From: Ted Zlatanov <tzz@lifelogs.com>
To: Matthieu Moy <Matthieu.Moy@grenoble-inp.fr>
Cc: Junio C Hamano <gitster@pobox.com>, Jeff King <peff@peff.net>,
	Michal Nazarewicz <mpn@google.com>,
	git@vger.kernel.org, Krzysztof Mazur <krzysiek@podlesie.net>,
	Michal Nazarewicz <mina86@mina86.com>
Subject: Re: [PATCH] git-send-email: add ~/.authinfo parsing
Date: Wed, 06 Feb 2013 10:58:13 -0500	[thread overview]
Message-ID: <87sj59mo2y.fsf@lifelogs.com> (raw)
In-Reply-To: <vpqmwvhxyuj.fsf@grenoble-inp.fr> (Matthieu Moy's message of "Wed, 06 Feb 2013 16:10:12 +0100")

On Wed, 06 Feb 2013 16:10:12 +0100 Matthieu Moy <Matthieu.Moy@grenoble-inp.fr> wrote: 

MM> Ted Zlatanov <tzz@lifelogs.com> writes:
MM> [...] so the way to go for send-email is probably to libify the
MM> credential support in git-remote-mediawiki, and to use it in send-email.
>> 
>> I looked and that's indeed very useful.  If it's put in a library, I'd
>> use credential_read() and credential_write() in my netrc credential
>> helper.  But I would formalize it a little more about the token names
>> and output,

MM> Can you elaborate on this? The idea of the Perl code was to mimick a
MM> call to the C API, keeping essentially the same names.

None of these are a big deal, and Michal said he's working on libifying
this anyhow:

- making 'fill' a special operation is weird
- anchor the key regex to beginning of line (not strictly necessary)
- sort the output tokens (after 'url' is extracted) so the output is consistent and testable

>> and I wouldn't necessarily die() on error. 

MM> Sure, die()ing in a library is bad.

>> Maybe this can be merged with the netrc credential helper's
>> read_credential_data_from_stdin() and print_credential_data()?

MM> I don't know about the netrc credential helper, but I guess that's
MM> another layer. The git-remote-mediawiki code is the code to call the
MM> credential C API, that in turn may (or may not) call a credential
MM> helper.

Yup.  But what you call "read" and "write" are, to the credential
helper, "write" and "read" but it's the same protocol :)  So maybe the
names should be changed to reflect that, e.g. "query" and "response."

MM> One thing to be careful about: git-remote-mediawiki is currently a
MM> standalone script, so it can be installed with a plain "cp
MM> git-remote-mediawiki $somewhere/".  One consequence of libification
MM> is that it adds a dependency on the library (e.g. Git.pm). We should
MM> be carefull to keep it easy for the user to install it (e.g. some
MM> kind of "make install", or update the doc).

I don't know--it's up to the `git-remote-mediawiki' maintainers...  But
I think anywhere you have Git, you also have Git.pm, right?  Maybe?  But
then you also have to look at whether Git.pm has the functionality you
need... so I better go quiet :)

Ted