Re: [PATCH 6/7] Fix tests under GETTEXT_POISON on git-remote

From: Jiang Xin <worldhello.net@gmail.com>
Date: Aug 22, 2012, 14:56 UTC
Message-ID: <CANYiYbE9sJNOXxxFX_xdow=tnWfUwBYOx4S0_CgaB+bUQxDSRg@mail.gmail.com>
In-Reply-To: <7v628bdhol.fsf@alter.siamese.dyndns.org>

2012/8/22 Junio C Hamano <gitster@pobox.com>:

Show 9 quoted lines

> Which part of the output from "git remote show" does this test
> expect to be translated?  Specifically, does "tracked" ever get
> translated?
>
> It appears that _(" tracked") is indeed marked for translation in
> the source, so how can we expect value in $actual be any useful for
> any comparison?
>
> Confused...

Yes, because _("tracked") is marked for translation, check_remote_track() is hard to do right.

Show 6 quoted lines

>  check_remote_track () {
>         actual=$(git remote show "$1" | sed -ne 's|^    \(.*\) tracked$|\1|p')
>         shift &&
> -       tokens_match "$*" "$actual"
> +       tokens_i18nmatch "$*" "$actual"
>  }

Output of command "git remote show origin" is:

    # GETTEXT POISON #
    # GETTEXT POISON #
    # GETTEXT POISON #
    # GETTEXT POISON #
    # GETTEXT POISON #
        master# GETTEXT POISON #
        side  # GETTEXT POISON #

Maybe we should bypass all testcases which calling check_remote_track(). Changing 'tokens_match' to 'tokens_i18nmatch' won't fix but hide the problem.

- test_expect_success 'remote information for the origin' '
+ test_expect_success C_LOCALE_OUTPUT 'remote information for the origin' '

- test_expect_succes 'remove remote' '
+ test_expect_success C_LOCALE_OUTPUT 'remove remote' '

- test_expect_success 'add another remote' '
+ test_expect_success C_LOCALE_OUTPUT 'add another remote' '

-- 
Jiang Xin