[GSoC PATCH] fsck: use Q_() for the verbose connectivity message

[GSoC PATCH] fsck: use Q_() for the verbose connectivity message

[Aniket Gupta]

The verbose "Checking connectivity" message has the word "objects"
hardcoded as a plural inside a plain _(), so translators can't give the
right form for the number. Use Q_() like the other count messages so
gettext can pick it. English output stays the same.

Signed-off-by: Aniket Gupta <aniketgupta3001@gmail.com>
---
The number printed here comes from the hash table size in
get_max_object_index() and not a real object count, so it never
says 1. Left that alone, can look at it separately if people think
it's worth it.

 builtin/fsck.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/builtin/fsck.c b/builtin/fsck.c
index 9af4cc085..fedf8f853 100644
--- a/builtin/fsck.c
+++ b/builtin/fsck.c
@@ -379,7 +379,10 @@ static void check_connectivity(struct repository *repo)
 	/* Look up all the requirements, warn about missing objects.. */
 	max = get_max_object_index(repo);
 	if (verbose)
-		fprintf_ln(stderr, _("Checking connectivity (%d objects)"), max);
+		fprintf_ln(stderr, Q_("Checking connectivity (%d object)",
+				      "Checking connectivity (%d objects)",
+				      max),
+			   max);
 
 	for (i = 0; i < max; i++) {
 		struct object *obj = get_indexed_object(repo, i);
-- 
2.56.0.186.gc51408889.dirty

Re: [GSoC PATCH] fsck: use Q_() for the verbose connectivity message

[Junio C Hamano]

Aniket Gupta <aniketgupta3001@gmail.com> writes:

> The verbose "Checking connectivity" message has the word "objects"
> hardcoded as a plural inside a plain _(), so translators can't give the
> right form for the number. Use Q_() like the other count messages so
> gettext can pick it. English output stays the same.
>
> Signed-off-by: Aniket Gupta <aniketgupta3001@gmail.com>
> ---
> The number printed here comes from the hash table size in
> get_max_object_index() and not a real object count, so it never
> says 1. Left that alone, can look at it separately if people think
> it's worth it.

The size of the hashtable we use ought to be comfortably larger than
the actual number of objects we are looking at for the hashtable to
perform well, so the number we are showing indeed is not the actual
number of objects.  Nevertheless, using Q_() is the right thing to
do here, especially to support languages with more complex plural
forms.

We _could_ add a NEEDSWORK comment near there to say that future
developers may want to think about the number being inaccurate and
do something about it (or not), but I think your decision not to
address the inaccuracies in this patch even though you noticed it is
the right one.

Will queue.  Thanks.

>  builtin/fsck.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/builtin/fsck.c b/builtin/fsck.c
> index 9af4cc085..fedf8f853 100644
> --- a/builtin/fsck.c
> +++ b/builtin/fsck.c
> @@ -379,7 +379,10 @@ static void check_connectivity(struct repository *repo)
>  	/* Look up all the requirements, warn about missing objects.. */
>  	max = get_max_object_index(repo);
>  	if (verbose)
> -		fprintf_ln(stderr, _("Checking connectivity (%d objects)"), max);
> +		fprintf_ln(stderr, Q_("Checking connectivity (%d object)",
> +				      "Checking connectivity (%d objects)",
> +				      max),
> +			   max);
>  
>  	for (i = 0; i < max; i++) {
>  		struct object *obj = get_indexed_object(repo, i);