Re: [PATCH] i18n: move the trailing space out of translatable strings

[Junio C Hamano <gitster@pobox.com>]

Nguyễn Thái Ngọc Duy  <pclouds@gmail.com> writes:

> I've got this with Vietnamese translation
>
>     $ git status
>     HEAD được tách rời từorigin/master
>
> One does not need to understand Vietnamese to see that a space is
> missing before "origin/master"...

Not really.  Only if one guesses (or knows) that Vietnamese is among
the languages that use SP to separate words.  There are languages
that do not use inter-word spaces.

Because those languages are in the minority, it would be easier to
unconditionally add SP after translatable string regardless of the
l10n target languages, but I am afraid that it is going in a wrong
direction in the longer term.

>  		clean_print_color(CLEAN_COLOR_PROMPT);
> -		printf(_("Input ignore patterns>> "));
> +		printf("%s ", _("Input ignore patterns>>"));
>  		clean_print_color(CLEAN_COLOR_RESET);

As this space after the prompt may be different from inter-word
space after all, this one may probably be on the borderline.

> @@ -754,7 +754,8 @@ static int ask_each_cmd(void)
>  		/* Ctrl-D should stop removing files */
>  		if (!eof) {
>  			qname = quote_path_relative(item->string, NULL, &buf);
> -			printf(_("remove %s? "), qname);
> +			printf(_("remove %s?"), qname);
> +			putchar(' ');

As is this change.

But it is an example that can be used to illustrate my earlier
point.  The l10n target language may want to have _no_ space between
words, i.e. to turn the above into:

	printf("%sを削除しますか?", qname);
        putchar(' ');

i.e. no space between the object of the verb ("the path being
removed") and the verb ("to remove"), while still wanting to have SP
between the prompt and its answer like everybody else.  And having
SP after "remove" in the gettext key _is_ a good thing, as l10n
people can choose not to have any SP between the verb and its
object.

> diff --git a/builtin/clone.c b/builtin/clone.c
> index 874e0fd..d48ccee 100644
> --- a/builtin/clone.c
> +++ b/builtin/clone.c
> @@ -551,12 +551,12 @@ static void update_remote_refs(const struct ref *refs,
>  
>  	if (check_connectivity) {
>  		if (transport->progress)
> -			fprintf(stderr, _("Checking connectivity... "));
> +			fprintf(stderr, _("Checking connectivity..."));
>  		if (check_everything_connected_with_transport(iterate_ref_map,
>  							      0, &rm, transport))
>  			die(_("remote did not send all necessary objects"));
>  		if (transport->progress)
> -			fprintf(stderr, _("done.\n"));
> +			fprintf(stderr, " %s", _("done.\n"));

But I think that this changes enforces the inter-word SP policy that
could go against convention by specific l10n target languages.