Re: [PATCH] builtin-bundle.c - use stream buffered input for rev-list

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

Re: [PATCH] builtin-bundle.c - use stream buffered input for rev-list

From: Junio C Hamano <hidden>
Date: 2016-06-15 22:43:28

Mark Levedahl [off-list ref] writes:

git-bundle create on cygwin was nearly unusable due to 1 character
at a time (unbuffered) reading from an exec'ed process.

Gaah, well spotted.  I was wondering what that 20x difference
was, but did not notice that read_string() thing.

That function is probably sometimes necessary if you are reading
from a non-seekable fd and need to stop immediately after
reading the header (so that you can hand the rest to pack
handling code), but it really should be the last-ditch thing to
read one-byte-at-a-time.

Also, I even suspect that we are always reading from a seekable
bundle file in this program and it should be able to use
buffered input with seeking back as needed after you read
enough.

[PATCH] builtin-bundle - use buffered reads for bundle header

From: Mark Levedahl <hidden>
Date: 2016-06-15 22:43:28

This eliminates all use of byte-at-a-time reading of data in this
function: as Junio noted, a bundle file is seekable so we can
reset the file position to the first part of the pack-file using lseek
after reading the header.

Signed-off-by: Mark Levedahl <redacted>
---
 builtin-bundle.c |   37 +++++++++++++------------------------
 1 files changed, 13 insertions(+), 24 deletions(-)

diff --git a/builtin-bundle.c b/builtin-bundle.c
index b954213..6651c51 100644
--- a/builtin-bundle.c
+++ b/builtin-bundle.c

@@ -44,38 +44,21 @@ struct bundle_header {
 	struct ref_list references;
 };
 
-/* this function returns the length of the string */
-static int read_string(int fd, char *buffer, int size)
-{
-	int i;
-	for (i = 0; i < size - 1; i++) {
-		ssize_t count = xread(fd, buffer + i, 1);
-		if (count < 0)
-			return error("Read error: %s", strerror(errno));
-		if (count == 0) {
-			i--;
-			break;
-		}
-		if (buffer[i] == '\n')
-			break;
-	}
-	buffer[i + 1] = '\0';
-	return i + 1;
-}
-
 /* returns an fd */
 static int read_header(const char *path, struct bundle_header *header) {
 	char buffer[1024];
-	int fd = open(path, O_RDONLY);
+	int fd;
+    long fpos;
+    FILE *ffd = fopen(path, "r");
 
-	if (fd < 0)
+	if (!ffd)
 		return error("could not open '%s'", path);
-	if (read_string(fd, buffer, sizeof(buffer)) < 0 ||
+    if (!fgets(buffer, sizeof(buffer), ffd) ||
 			strcmp(buffer, bundle_signature)) {
-		close(fd);
+        fclose(ffd);
 		return error("'%s' does not look like a v2 bundle file", path);
 	}
-	while (read_string(fd, buffer, sizeof(buffer)) > 0
+    while (fgets(buffer, sizeof(buffer), ffd)
 			&& buffer[0] != '\n') {
 		int is_prereq = buffer[0] == '-';
 		int offset = is_prereq ? 1 : 0;

@@ -97,6 +80,12 @@ static int read_header(const char *path, struct bundle_header *header) {
 		add_to_ref_list(sha1, isspace(delim) ?
 				buffer + 41 + offset : "", list);
 	}
+    fpos = ftell(ffd);
+    fclose(ffd);
+    fd = open(path, O_RDONLY);
+    if (fd < 0)
+        return error("could not open '%s'", path);
+    lseek(fd, fpos, SEEK_SET);
 	return fd;
 }

-- 
1.5.3.rc4.54.gbe00

[PATCH] builtin-bundle - use buffered reads for bundle header

From: Mark Levedahl <hidden>
Date: 2016-06-15 22:43:28

This eliminates all use of byte-at-a-time reading of data in this
function: as Junio noted, a bundle file is seekable so we can
reset the file position to the first part of the pack-file using lseek
after reading the header.

Signed-off-by: Mark Levedahl <redacted>
---

 Try again - editor had its tab settings mangled..grrrr

 builtin-bundle.c |   37 +++++++++++++------------------------
 1 files changed, 13 insertions(+), 24 deletions(-)

diff --git a/builtin-bundle.c b/builtin-bundle.c
index b954213..b8f8e78 100644
--- a/builtin-bundle.c
+++ b/builtin-bundle.c

@@ -44,38 +44,21 @@ struct bundle_header {
 	struct ref_list references;
 };

-/* this function returns the length of the string */
-static int read_string(int fd, char *buffer, int size)
-{
-	int i;
-	for (i = 0; i < size - 1; i++) {
-		ssize_t count = xread(fd, buffer + i, 1);
-		if (count < 0)
-			return error("Read error: %s", strerror(errno));
-		if (count == 0) {
-			i--;
-			break;
-		}
-		if (buffer[i] == '\n')
-			break;
-	}
-	buffer[i + 1] = '\0';
-	return i + 1;
-}
-
 /* returns an fd */
 static int read_header(const char *path, struct bundle_header *header) {
 	char buffer[1024];
-	int fd = open(path, O_RDONLY);
+	int fd;
+	long fpos;
+	FILE *ffd = fopen(path, "r");

-	if (fd < 0)
+	if (!ffd)
 		return error("could not open '%s'", path);
-	if (read_string(fd, buffer, sizeof(buffer)) < 0 ||
+	if (!fgets(buffer, sizeof(buffer), ffd) ||
 			strcmp(buffer, bundle_signature)) {
-		close(fd);
+        fclose(ffd);
 		return error("'%s' does not look like a v2 bundle file", path);
 	}
-	while (read_string(fd, buffer, sizeof(buffer)) > 0
+	while (fgets(buffer, sizeof(buffer), ffd)
 			&& buffer[0] != '\n') {
 		int is_prereq = buffer[0] == '-';
 		int offset = is_prereq ? 1 : 0;

@@ -97,6 +80,12 @@ static int read_header(const char *path, struct bundle_header *header) {
 		add_to_ref_list(sha1, isspace(delim) ?
 				buffer + 41 + offset : "", list);
 	}
+	fpos = ftell(ffd);
+	fclose(ffd);
+	fd = open(path, O_RDONLY);
+	if (fd < 0)
+		return error("could not open '%s'", path);
+	lseek(fd, fpos, SEEK_SET);
 	return fd;
 }

--

1.5.3.rc4.54.gbe00

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help