Re: [PATCH 3/3] tests: warn against appending non-httpd-specific tests at the end

From: SZEDER Gábor <szeder.dev@gmail.com>
Date: Aug 2, 2019, 10:09 UTC
Message-ID: <20190802100956.GV20404@szeder.dev>
In-Reply-To: <xmqqzhksn3zx.fsf@gitster-ct.c.googlers.com>

On Thu, Aug 01, 2019 at 11:18:42AM -0700, Junio C Hamano wrote:

Show 14 quoted lines

> SZEDER Gábor <szeder.dev@gmail.com> writes:
> 
> > On Thu, Aug 01, 2019 at 05:53:09PM +0200, SZEDER Gábor wrote:
> >> Subject: Re: [PATCH 3/3] tests: warn against appending non-httpd-specific
> >>  tests at the end
> >
> > This subject line kind of sucks, doesn't it?!
> >
> > Alas I had a bit of a hard time coming up with something better.  So
> > far the best (well, least bad) I could think of is:
> >
> >   t: warn against adding non-httpd-specific tests after sourcing 'lib-httpd'
> 
> That reads well.

Ok. Should I resend, or will you amend? (I see that 'pu' contains the old subject line).

> Thanks.  It must have been tedious to make sure moving tests around
> won't break them due to some (hidden) inter-dependency among them.
> Very much appreciated.

Luckily, all the moved non-httpd tests and all the httpd tests in 't5510' and 't5703' work in their own directories, meaning that they neither influence other tests nor are influenced by other tests, with the exception of that $LOCAL_PRISTINE directory that I noted in the commit message of patch 2.

I did actually diff-ed the output of the involved tests before and after these patches, and they were essentially identical (the only differences were that extra 'rm -rf', a couple of different timestamps, different commit oids shown by 'git commit' or 'git fetch', and the occasional races between the trace output and actual command output).