From: Atharva Raykar Date: Sat, 10 Jul 2021 07:52:16 GMT Subject: Re: [GSoC] [PATCH v2 2/4] submodule: prefix die messages with 'fatal' Message-ID: <598E78EB-48B1-4C2E-BD89-90EF003A15F6@gmail.com> In-Reply-To: > On 09-Jul-2021, at 20:22, Đoàn Trần Công Danh wrote: > > On 2021-07-08 15:25:31+0530, Atharva Raykar wrote: >> The standard `die()` function that is used in C code prefixes all the >> messages passed to it with 'fatal: '. This does not happen with the >> `die` used in 'git-submodule.sh'. >> >> Let's prefix each of the shell die messages with 'fatal: ' so that when >> they are converted to C code, the error messages stay the same as before >> the conversion. > > That sounds good. > >> --- a/git-submodule.sh >> +++ b/git-submodule.sh >> @@ -147,7 +147,7 @@ cmd_add() >> >> if ! git submodule--helper config --check-writeable >/dev/null 2>&1 >> then >> - die "$(eval_gettext "please make sure that the .gitmodules file is in the working tree")" >> + die "$(eval_gettext "fatal: please make sure that the .gitmodules file is in the working tree")" > > Except that, "fatal: " isn't subjected to translation. And this will > create new translatable item for translator. Perhaps: > > - die "$(eval_gettext "please make sure that the .gitmodules file is in the working tree")" > + die "fatal: $(eval_gettext "please make sure that the .gitmodules file is in the working tree")" Okay, I have made the change. I was wondering if there any specific reason as to why 'fatal' should not be translated? Is it because an intermediate change like this should not create more work for translators?