Re: [PATCH v2 1/6] i18n: factorize generic failure messages

From: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Date: Apr 3, 2022, 14:34 UTC
Message-ID: <220403.86wng6kzby.gmgdl@evledraar.gmail.com>
In-Reply-To: <4b46ba19-3a39-7317-42fb-4fd4bba09b36@gmail.com>

On Sun, Apr 03 2022, Bagas Sanjaya wrote:

Show 22 quoted lines

> On 02/04/22 23.10, Jean-Noël Avila via GitGitGadget wrote:
> [...]
>> index 04ce33cdcd4..30a4b81257d 100644
>> --- a/setup.c
>> +++ b/setup.c
>> @@ -1506,12 +1506,12 @@ int daemonize(void)
>>   		case 0:
>>   			break;
>>   		case -1:
>> -			die_errno(_("fork failed"));
>> +			die_errno(_("the function '%s' failed"), "fork");
>>   		default:
>>   			exit(0);
>>   	}
>>   	if (setsid() == -1)
>> -		die_errno(_("setsid failed"));
>> +		die_errno(_("the function '%s' failed"), "setsid");
>>   	close(0);
>>   	close(1);
>>   	close(2);
>
> Why not simply "'%s' failed"?

I think saying what failed is helpful in any case, a user who knows nothing about *nix APIs might be quite perplexed at their source control telling them their fork failed, why not the spoon? :)

So including more context helps.

But also because there's languages where adjectives like "failed" are different depending on the grammatical gender of the subject.

So even if the original and translation would be needlessly terse in any case, the message would also become imppossible to translate in a gramatically correct way.