From: Jonathan Nieder Date: Wed, 25 Jul 2012 04:26:33 GMT Subject: Re: [PATCH v4 5/7] i18n: am: mark more strings for translation Message-ID: <20120725042633.GB3055@burratino> In-Reply-To: <1493bd09efd0ba6e1d56500305d989047da245ec.1343188013.git.worldhello.net@gmail.com> Jiang Xin wrote: > Mark strings in 'git-am.sh' for translation. In the last chunk, I > changed '$1' to '-b/--binary' for this reason: > > * First, if there is a variable in the l10n message, we could not use > gettext. Because the variable will be expanded (evaluated) and will > never match the entry in the po file. > > * Second, if there is a positional parameter ($1, $2,...) in the > message, we could not use eval_gettext either. Because > eval_gettext may be a wapper for gettext, and the positional > parameter would loose it's original context. Yes, I think it's a good change. > --- a/git-am.sh > +++ b/git-am.sh [...] > @@ -387,8 +386,8 @@ do > -i|--interactive) > interactive=t ;; > -b|--binary) > - echo >&2 "The $1 option has been a no-op for long time, and" > - echo >&2 "it will be removed. Please do not use it anymore." > + echo >&2 $(gettext "The -b/--binary option has been a no-op for long time, and > +it will be removed. Please do not use it anymore.") Could this be simplified to gettextln >&2 'The -b/--binary option... ... any more.' or gettextln 'The -b/--binary option ... ... any more.' >&2 or would that confuse xgettext? That would be more comforting than "echo" because if the translated string includes backslashes then the behavior of echo can be unpredictable. Sorry I missed that last time. The rest is indeed Reviewed-by: Jonathan Nieder Thanks.