On Sun, Jun 24, 2012 at 11:28 PM, Jonathan Nieder <jrnieder@xxxxxxxxx> wrote:
> Nguyễn Thái Ngọc Duy wrote:
>
>> This fixes all GETTEXT_POISON breakages caused by recent i18n changes.
>
> First, thanks much for this.
>
> Lots of these could be fixed in a more targetted way by using
> test_i18ngrep, but the C_LOCALE_OUTPUT prereq works just as well as a
> way to double-check that the newly translated strings are hopefully
> not disrupting any functionality people rely on.
I probably still don't grasp the principles behind gettext-poison.
Pretending success quietly in test_i18ngrep and test_i18ncmp makes me
uncomfortable. At least C_LOCAL_OUTPUT clearly shows the test is
skipped, which is why I prefer it over test_i18n*. Letting it run to
catch crashes is good, but not enough (it's more likely to crash with
wrong printf format specifiers for example, which gettext-poison won't
catch). Will update it according to your comments.
> [...]
>> --- a/t/t0006-date.sh
>> +++ b/t/t0006-date.sh
>> @@ -9,7 +9,7 @@ TEST_DATE_NOW=1251660000; export TEST_DATE_NOW
>> check_show() {
>> t=$(($TEST_DATE_NOW - $1))
>> echo "$t -> $2" >expect
>> - test_expect_${3:-success} "relative date ($2)" "
>> + test_expect_${3:-success} C_LOCALE_OUTPUT "relative date ($2)" "
>> test-date show $t >actual &&
>> test_cmp expect actual
>> "
>
> Could use test_i18ncmp so we catch if test-date crashes, but anyway,
> yeah, we can't expect to be able to meaningfully test date formatting
> in another language.
This reminds me of an inkscape crash I encountered long ago because of
a mistake in Vietnamese translation. Too bad we can't test all locales
for similar bugs (or we would need to i18n-ize expected results in the
test suite too)
--
Duy
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html