[RFC/PATCH] git-shortlog: respect i18n.logOutputEncoding config setting

As git-shortlog can be used as a filter as well, we do not really have
the encoding info to do a reencode_string(), but in case
i18n.logOutputEncoding is set, we can try to convert to the given value
from utf-8.

Signed-off-by: Miklos Vajna <redacted>
---

It was just annoying for me that git log respected
i18n.logOutputEncoding, but I still saw unencoded utf-8 in git-shortlog
output. This patches fixes my problem.

Yes, I'm aware that hardwiring that utf-8 value is a bit hackish. But at
least this way the output encoding is correct in case the original
encoding is utf-8, which is true in most cases.

Or do you have a better idea?

 builtin-shortlog.c |    8 +++++++-
 1 files changed, 7 insertions(+), 1 deletions(-)

diff --git a/builtin-shortlog.c b/builtin-shortlog.c
index badd912..cd95858 100644
--- a/builtin-shortlog.c
+++ b/builtin-shortlog.c

@@ -45,6 +45,7 @@ static void insert_one_record(struct shortlog *log,
 	const char *eol;
 	const char *boemail, *eoemail;
 	struct strbuf subject = STRBUF_INIT;
+	char *encoded_name = NULL;
 
 	boemail = strchr(author, '<');
 	if (!boemail)

@@ -84,7 +85,12 @@ static void insert_one_record(struct shortlog *log,
 		snprintf(namebuf + len, room, " <%.*s>", maillen, emailbuf);
 	}
 
-	item = string_list_insert(namebuf, &log->list);
+	if (git_log_output_encoding)
+		encoded_name = reencode_string(namebuf, git_log_output_encoding, "utf-8");
+	if (!encoded_name)
+		encoded_name = xstrdup(namebuf);
+	item = string_list_insert(encoded_name, &log->list);
+	free(encoded_name);
 	if (item->util == NULL)
 		item->util = xcalloc(1, sizeof(struct string_list));
 

-- 
1.6.1.3

On Fri, Feb 20, 2009 at 02:12:38AM +0100, Miklos Vajna [off-list ref] wrote:

As git-shortlog can be used as a filter as well, we do not really have
the encoding info to do a reencode_string(), but in case
i18n.logOutputEncoding is set, we can try to convert to the given value
from utf-8.

Should I resend this patch without RFC in the subject?

I just hit this bug again today, then realised that I already posted a
patch for this problem. :-)