Re: [PATCH 0/7] More i18n fixes

From: Johannes Sixt <j6t@kdbg.org>
Date: Mar 21, 2022, 06:48 UTC
Message-ID: <e44b6ccf-21a2-72c6-4d40-dc0004895255@kdbg.org>
In-Reply-To: <pull.1181.git.1647813291.gitgitgadget@gmail.com>

Am 20.03.22 um 22:54 schrieb Jean-Noël Avila via GitGitGadget:

Show 5 quoted lines

> This is another i18n PR (and hopefully the last for a while).
> 
> As usual, the intent is kept the same: curbing the number of strings to
> translate, remove constant, error prone parts out of the way, trying in some
> sense to "put a precedent" so that the template strings can be reused later.

I feel that many of the example conversions look like sentence lego because there remains only one English word, e.g., "'%s' failed". The converted code does not leave a hint for the translators what the %s will be. Is it a command, a function name, somehting else? Even if the hint was provided, different translations may be required depending on the substituted entity. Did you investigate the existing translations whether all of them can be converted to the new scheme?

> This series has also a RFC status: can "bad argument" messages be merged
> with unrecognized argument?

The cases that patch 7/7 transforms look like they need not keep "unrecognized argument", but can be converted to "bad argument".

Disclaimer: neither am I a translator nor a user of a translated Git.

-- Hannes