git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH] blame.c: prepare_lines should not call xrealloc for every line

From
Junio C Hamano <gitster@pobox.com>
Date
Feb 4, 2014, 21:00 UTC
Message-ID
<xmqq4n4e8uf1.fsf@gitster.dls.corp.google.com>
In-Reply-To
<xmqq8utq8uxn.fsf@gitster.dls.corp.google.com>
Junio C Hamano <gitster@pobox.com> writes:
Show 8 quoted lines
> David Kastrup <dak@gnu.org> writes:
>
>> Whitespace error in line 1778.  Should I be reposting?
>
> Heh, let me try to clean it up first and then repost for your
> review.
>
> Thanks.
-- >8 --
From: David Kastrup <dak@gnu.org>

Making a single preparation run for counting the lines will avoid memory fragmentation. Also, fix the allocated memory size which was wrong when sizeof(int *) != sizeof(int), and would have been too small for sizeof(int *) < sizeof(int), admittedly unlikely.

Signed-off-by: David Kastrup <dak@gnu.org>
---
 One logic difference from what was posted is that sb->lineno[num]
 is filled with the length of the entire buffer when the file ends
 with a complete line.  I do not remember if the rest of the logic
 actually depends on it (I think I use lineno[n+1] - lineno[n] to
 find the end of line, so this may matter for the last line),
 though.
 The original code dates back to 2006 when the author of the code
 was not paid for doing anything for Git but was doing it as a
 weekend and evening hobby, so it may not be so surprising to find
 this kind of "what was I thinking when I wrote it" inefficiency in
 such a code with $0 monetary value ;-)
 builtin/blame.c | 37 +++++++++++++++++++++----------------
 1 file changed, 21 insertions(+), 16 deletions(-)
diff --git a/builtin/blame.c b/builtin/blame.c
index e44a6bb..2364661 100644
--- a/builtin/blame.c
+++ b/builtin/blame.c
@@ -1772,25 +1772,30 @@ static int prepare_lines(struct scoreboard *sb)
 {
 	const char *buf = sb->final_buf;
 	unsigned long len = sb->final_buf_size;
-	int num = 0, incomplete = 0, bol = 1;
+	const char *end = buf + len;
+	const char *p;
+	int *lineno;
+	int num = 0, incomplete = 0;
+
+	for (p = buf; p < end; p++) {
+		p = memchr(p, '\n', end - p);
+		if (!p)
+			break;
+		num++;
+	}
 
-	if (len && buf[len-1] != '\n')
+	if (len && end[-1] != '\n')
 		incomplete++; /* incomplete line at the end */
-	while (len--) {
-		if (bol) {
-			sb->lineno = xrealloc(sb->lineno,
-					      sizeof(int *) * (num + 1));
-			sb->lineno[num] = buf - sb->final_buf;
-			bol = 0;
-		}
-		if (*buf++ == '\n') {
-			num++;
-			bol = 1;
-		}
+
+	sb->lineno = lineno = xmalloc(sizeof(int) * (num + incomplete + 1));
+
+	for (p = buf; p < end; p++) {
+		*lineno++ = p - buf;
+		p = memchr(p, '\n', end - p);
+		if (!p)
+			break;
 	}
-	sb->lineno = xrealloc(sb->lineno,
-			      sizeof(int *) * (num + incomplete + 1));
-	sb->lineno[num + incomplete] = buf - sb->final_buf;
+	sb->lineno[num + incomplete] = len;
 	sb->num_lines = num + incomplete;
 	return sb->num_lines;
 }
Previous: Junio C HamanoNext: David Kastrup
Message 4 of 23 in “blame.c: prepare_lines should not call xrealloc for every line”
  1. blame.c: prepare_lines should not call xrealloc for every lineDavid Kastrup, Feb 4, 2014
  2. David KastrupFeb 4, 2014
  3. Junio C HamanoFeb 4, 2014
  4. Junio C HamanoFeb 4, 2014
  5. David KastrupFeb 4, 2014
  6. Philip OakleyFeb 4, 2014
  7. Philip OakleyFeb 4, 2014
  8. Junio C HamanoFeb 4, 2014
  9. David KastrupFeb 4, 2014
  10. Junio C HamanoFeb 4, 2014
  11. David KastrupFeb 4, 2014
  12. Junio C HamanoFeb 4, 2014
  13. David KastrupFeb 4, 2014
  14. Junio C HamanoFeb 4, 2014
  15. David KastrupFeb 4, 2014
  16. Junio C HamanoFeb 4, 2014
  17. David KastrupFeb 5, 2014
  18. Junio C HamanoFeb 5, 2014
  19. David KastrupFeb 6, 2014
  20. David KastrupFeb 6, 2014
  21. David KastrupFeb 5, 2014
  22. Junio C HamanoFeb 5, 2014
  23. David KastrupFeb 5, 2014

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.