Re: [PATCH/RFC] inconsistent error messages for translation
From: Junio C Hamano <hidden>
Date: 2016-06-15 22:51:06
Motiejus Jakštys [off-list ref] writes:
There are lots of variants of the same message: msgid "cannot stat '%s'" msgid "failed to stat '%s'" msgid "failed to stat %s\n" msgid "Could not stat '%s'"
I am not sure what to do with the trailing LF (it may be a bug in the
message written without being aware that die/warn will give their own LF
at the end), but the first one ("cannot $verb '$name'") is preferred.
Also same trivial fix:
- return error(_("path '%s' does not have all three versions"),
+ return error(_("path '%s' does not have all 3 versions"),I would not call this a "fix", though. What problem does it solve?
quoted hunk
diff --git a/builtin/grep.c b/builtin/grep.c index 10a1f65..24d19b8 100644 --- a/builtin/grep.c +++ b/builtin/grep.c@@ -413,7 +413,7 @@ static void *load_file(const char *filename, size_t *sz) if (lstat(filename, &st) < 0) { err_ret: if (errno != ENOENT) - error(_("'%s': %s"), filename, strerror(errno)); + error("'%s': %s", filename, strerror(errno)); return NULL; } if (!S_ISREG(st.st_mode))
This hunk is a fix for mismarked message and is unrelated to the error message unification, no? I prefer to have only this part as a separate patch.