threads / patch / 20406

patchgit-describe: Die early if there are no possible descriptions

Subject: [PATCH] git-describe: Die early if there are no possible descriptions

## tl;dr

4 messages between Aug 5, 2009 and Aug 6, 2009. Diffs are folded; open one to read it.

replies: 3people: 2as markdown or json

Björn Steinbrink· Aug 5, 2009, 14:17 UTC · lore

If we found no refs that may be used for git-describe with the current options, then die early instead of pointlessly walking the whole history.

Signed-off-by: Björn Steinbrink <B.Steinbrink@gmx.de>
---
In git.git with all the tags dropped, this makes "git describe" go down
from 0.244 to 0.003 seconds for me. This is especially noticeable with
"git submodule" which calls describe with increasing levels of allowed
refs to be matched. Without tags, this means that it walks the whole
history in the submodule twice (first annotated, then plain tags), just
to find out that it can't describe the thing anyway.

I'm not particularly sure about found_names actually counting the found names, it was just out of the thought that maybe the walking code could make use of it, but I didn't actually check that and ran out of time, so I'm sending this version, hoping that it doesn't suck too much.

 builtin-describe.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)

Show changes to builtin-describe.c +6 −0

diff --git a/builtin-describe.c b/builtin-describe.c
index 7a66298..cf802d1 100644
--- a/builtin-describe.c
+++ b/builtin-describe.c
@@ -20,6 +20,7 @@ static int tags;	/* Allow lightweight tags */
 static int longformat;
 static int abbrev = DEFAULT_ABBREV;
 static int max_candidates = 10;
+static int found_names = 0;
 static const char *pattern;
 static int always;
 
@@ -39,6 +40,8 @@ static void add_to_known_names(const char *path,
 			       const unsigned char *sha1)
 {
 	struct commit_name *e = commit->util;
+	if (!e)
+		found_names++;
 	if (!e || e->prio < prio) {
 		size_t len = strlen(path)+1;
 		free(e);
@@ -195,6 +198,9 @@ static void describe(const char *arg, int last_one)
 		for_each_ref(get_name, NULL);
 	}
 
+	if (!found_names)
+		die("cannot describe '%s'", sha1_to_hex(sha1));
+
 	n = cmit->util;
 	if (n) {
 		/*

-- 
1.6.4.19.g42af.dirty

Shawn O. Pearce· Aug 5, 2009, 15:34 UTC · re: Björn Steinbrink · lore

Re: [PATCH] git-describe: Die early if there are no possible descriptions

Bj?rn Steinbrink <B.Steinbrink@gmx.de> wrote:

Show 16 quoted lines

> If we found no refs that may be used for git-describe with the current
> options, then die early instead of pointlessly walking the whole history.
> 
> Signed-off-by: Bj?rn Steinbrink <B.Steinbrink@gmx.de>
> ---
> In git.git with all the tags dropped, this makes "git describe" go down
> from 0.244 to 0.003 seconds for me. This is especially noticeable with
> "git submodule" which calls describe with increasing levels of allowed
> refs to be matched. Without tags, this means that it walks the whole
> history in the submodule twice (first annotated, then plain tags), just
> to find out that it can't describe the thing anyway.
> 
> I'm not particularly sure about found_names actually counting the found
> names, it was just out of the thought that maybe the walking code could
> make use of it, but I didn't actually check that and ran out of time, so
> I'm sending this version, hoping that it doesn't suck too much.

This seems reasonable to me. Really you don't need found_names to be a counter, but could just always set it to 1 every time the add_to_known_names function is called. All you care about is that add_to_known_names was invoked at least once.

Also, I really think that first paragraph after the --- should have been part of the commit message. The message above doesn't justify the change, even if it is fairly trivial, without that additional explanation.

Show 9 quoted lines

> @@ -39,6 +40,8 @@ static void add_to_known_names(const char *path,
>  			       const unsigned char *sha1)
>  {
>  	struct commit_name *e = commit->util;
> +	if (!e)
> +		found_names++;
>  	if (!e || e->prio < prio) {
>  		size_t len = strlen(path)+1;
>  		free(e);

-- 
Shawn.

Björn Steinbrink· Aug 5, 2009, 15:42 UTC · re: Shawn O. Pearce · lore

Re: [PATCH] git-describe: Die early if there are no possible descriptions

On 2009.08.05 08:34:12 -0700, Shawn O. Pearce wrote:

Show 22 quoted lines

> Bj?rn Steinbrink <B.Steinbrink@gmx.de> wrote:
> > If we found no refs that may be used for git-describe with the current
> > options, then die early instead of pointlessly walking the whole history.
> > 
> > Signed-off-by: Bj?rn Steinbrink <B.Steinbrink@gmx.de>
> > ---
> > In git.git with all the tags dropped, this makes "git describe" go down
> > from 0.244 to 0.003 seconds for me. This is especially noticeable with
> > "git submodule" which calls describe with increasing levels of allowed
> > refs to be matched. Without tags, this means that it walks the whole
> > history in the submodule twice (first annotated, then plain tags), just
> > to find out that it can't describe the thing anyway.
> > 
> > I'm not particularly sure about found_names actually counting the found
> > names, it was just out of the thought that maybe the walking code could
> > make use of it, but I didn't actually check that and ran out of time, so
> > I'm sending this version, hoping that it doesn't suck too much.
> 
> This seems reasonable to me.  Really you don't need found_names
> to be a counter, but could just always set it to 1 every time the
> add_to_known_names function is called.  All you care about is that
> add_to_known_names was invoked at least once.

OK.

> Also, I really think that first paragraph after the --- should
> have been part of the commit message.  The message above doesn't
> justify the change, even if it is fairly trivial, without that
> additional explanation.

Oh, d'oh, that's even what I intended to do (the original commit message was written in even more of a hurry than the mail itself), but messed up. Will re-send (also with found_names turned boolean)

Thanks, Björn

Björn Steinbrink· Aug 6, 2009, 12:15 UTC · re: Björn Steinbrink · lore

[PATCH v2] git-describe: Die early if there are no possible descriptions

If we find no refs that may be used for git-describe with the current options, then die early instead of pointlessly walking the whole history.

In git.git with all the tags dropped, this makes "git describe" go down from 0.244 to 0.003 seconds for me. This is especially noticeable with "git submodule status" which calls describe with increasing levels of allowed refs to be matched. For a submodule without tags, this means that it walks the whole history in the submodule twice (first annotated, then plain tags), just to find out that it can't describe the commit anyway.

Signed-off-by: Björn Steinbrink <B.Steinbrink@gmx.de>
---
v2 improved the commit message and turned found_names into a boolean,
instead of a counter for the found names.

 builtin-describe.c |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)

Show changes to builtin-describe.c +5 −0

diff --git a/builtin-describe.c b/builtin-describe.c
index 7a66298..75feb11 100644
--- a/builtin-describe.c
+++ b/builtin-describe.c
@@ -20,6 +20,7 @@ static int tags;	/* Allow lightweight tags */
 static int longformat;
 static int abbrev = DEFAULT_ABBREV;
 static int max_candidates = 10;
+static int found_names = 0;
 static const char *pattern;
 static int always;
 
@@ -49,6 +50,7 @@ static void add_to_known_names(const char *path,
 		memcpy(e->path, path, len);
 		commit->util = e;
 	}
+	found_names = 1;
 }
 
 static int get_name(const char *path, const unsigned char *sha1, int flag, void *cb_data)
@@ -195,6 +197,9 @@ static void describe(const char *arg, int last_one)
 		for_each_ref(get_name, NULL);
 	}
 
+	if (!found_names)
+		die("cannot describe '%s'", sha1_to_hex(sha1));
+
 	n = cmit->util;
 	if (n) {
 		/*

-- 
1.6.4.20.gd25bb.dirty