From: konglu@minatec.inpg.fr
Date: Mon, 04 Jun 2012 10:04:06 GMT
Subject: Re: [PATCHv6 2/4] t7512-status-help.sh: better advices for git status
Message-ID: <20120604120406.Horde.ge3VMHwdC4BPzIgWN72l8oA@webmail.minatec.grenoble-inp.fr>
In-Reply-To: <7vvcj8xffp.fsf@alter.siamese.dyndns.org>


Junio C Hamano <gitster@pobox.com> a écrit :

> I thought Matthieu always called himself "Matthieu Moy", not the
> other way around.  I also inferred the convention used for naming
> the mailboxes at your institution to be "$First.$Last@$Domain",
> which tells me that perhaps these names are all spelled backwards,
> i.e. "$Last $First", which probably goes against the convention used
> around you and Matthieu?

Hm, yeah it's spelled backwards, I didn't take notice of that.

>> +. ./test-lib.sh
>> +
>> +. "$TEST_DIRECTORY"/lib-rebase.sh
>> +
>> +set_fake_editor
>> +
>> +test_expect_success 'status when conflicts unresolved' '
>> +	git init git &&
>> +	cd git &&
>
> Eek.
>
> Later tests seem to use
>
> 	git init git &&
> 	(
> 		cd git &&
>                 ... everything is done inside git subdirectory ...
> 	)
>
> with either "test_when_finished 'rm -fr git'" clean-up before the
> first "git init git" or "&& rm -fr git" at the end of the && chain.
>
> Please be consistent.  If you want _me_ to pick one, I'd say
>
> 	test_when_finished 'rm -fr git' &&
>         git init git &&
>         (
>         	cd git &&
>                 ... do everything in git subdirectory ...
> 	)
>
> would be my choice.

Btw, should each test be independant with the others ? We saw that there
are already a lot of tests that are dependant. In this case, can we apply
this method with the parenthesis ?

Wouldn't it be better if we remove the "git init git" part?