Re: [PATCH v7 2/2] json-writer: t0019: add Python unit test

From: Jeff King <peff@peff.net>
To: Todd Zullinger <tmz@pobox.com>
Cc: git@jeffhostetler.com, git@vger.kernel.org, gitster@pobox.com,
	Jeff Hostetler <jeffhost@microsoft.com>
Subject: Re: [PATCH v7 2/2] json-writer: t0019: add Python unit test
Date: Wed, 6 Jun 2018 17:03:00 -0400	[thread overview]
Message-ID: <20180606210300.GA1879@sigill.intra.peff.net> (raw)
In-Reply-To: <20180606171052.GI3094@zaya.teonanacatl.net>

On Wed, Jun 06, 2018 at 01:10:52PM -0400, Todd Zullinger wrote:

> git@jeffhostetler.com wrote:
> > +# As a sanity check, ask Python to parse our generated JSON.  Let Python
> > +# recursively dump the resulting dictionary in sorted order.  Confirm that
> > +# that matches our expectations.
> > +test_expect_success PYTHON 'parse JSON using Python' '
> [...]
> > +	python "$TEST_DIRECTORY"/t0019/parse_json_1.py <output.json >actual &&
> 
> Would this be better using $PYTHON_PATH rather than
> hard-coding python as the command?

Probably. We may want to go the same route as we did for perl in 
a0e0ec9f7d (t: provide a perl() function which uses $PERL_PATH,
2013-10-28) so that test writers don't have to remember this.

That said, I wonder if it would be hard to simply do the python bits
here in perl. This is the first use of python in our test scripts (and
really the only user in the whole code base outside of a few fringe
commands). Leaving aside any perl vs python flame-war, I think there's
value in keeping the number of languages limited when there's not a
compelling reason to do otherwise.

-Peff