[PATCH] strbuf_read: skip unnecessary strbuf_grow at eof

Subsystems: the rest

DORMANTno replies

2 messages, 2 authors, 2016-06-15 · open the first message on its own page

[PATCH] strbuf_read: skip unnecessary strbuf_grow at eof

From: Jim Hill <hidden>
Date: 2016-06-15 23:05:04

Make strbuf_read not try to do read_in_full's job too.  If xread returns
less than was requested it can be either eof or an interrupted read.  If
read_in_full returns less than was requested, it's eof. Use read_in_full
to detect eof and not iterate when eof has been seen.

Signed-off-by: Jim Hill <redacted>
---
 strbuf.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/strbuf.c b/strbuf.c
index 88cafd4..c70733e 100644
--- a/strbuf.c
+++ b/strbuf.c
@@ -364,19 +364,19 @@ ssize_t strbuf_read(struct strbuf *sb, int fd, size_t hint)
 
 	strbuf_grow(sb, hint ? hint : 8192);
 	for (;;) {
-		ssize_t cnt;
+		ssize_t want = sb->alloc - sb->len - 1;
+		ssize_t got = read_in_full(fd, sb->buf + sb->len, want);
 
-		cnt = xread(fd, sb->buf + sb->len, sb->alloc - sb->len - 1);
-		if (cnt < 0) {
+		if (got < 0) {
 			if (oldalloc == 0)
 				strbuf_release(sb);
 			else
 				strbuf_setlen(sb, oldlen);
 			return -1;
 		}
-		if (!cnt)
+		sb->len += got;
+		if (got < want)
 			break;
-		sb->len += cnt;
 		strbuf_grow(sb, 8192);
 	}
 
-- 
2.4.1.4.gfc728c2

Re: [PATCH] strbuf_read: skip unnecessary strbuf_grow at eof

From: Jeff King <hidden>
Date: 2016-06-15 23:05:04

On Sun, May 31, 2015 at 11:16:45AM -0700, Jim Hill wrote:
Make strbuf_read not try to do read_in_full's job too.  If xread returns
less than was requested it can be either eof or an interrupted read.  If
read_in_full returns less than was requested, it's eof. Use read_in_full
to detect eof and not iterate when eof has been seen.
I think this makes sense. I somehow had to read this over several times
to understand that the main point is not the cleanup, but rather the
space savings from not doing an extra strbuf_grow. Perhaps it is because
the main idea is mentioned only in the subject. Or perhaps I was just
being dense.

-Peff
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help