Globbing for ignored branches?

Globbing for ignored branches?

[Markus Trippelsdorf]

I would like to ignore branches that match a certain pattern, e.g.:

markus@x4 gcc % git branch -a
  gcc-4.8
* master
  remotes/origin/HEAD -> origin/master
  remotes/origin/aldyh/cilk-in-gomp
...
  remotes/origin/hjl/arch
  remotes/origin/hjl/asan
...
  remotes/origin/hjl/x86/m16
  remotes/origin/ifunc
  remotes/origin/jason/4.6-cxx0x
  remotes/origin/jason/alias-decl
  remotes/origin/jason/comdat-debug
  remotes/origin/lw-ipo
  remotes/origin/master
...

Is it possible to ignore all branches that match "hjl"?

-- 
Markus

RE: Globbing for ignored branches?

[Jim Garrison]

> -----Original Message-----
> Behalf Of Markus Trippelsdorf
> Sent: Friday, January 24, 2014 1:01 AM
> Subject: Globbing for ignored branches?
> 
> I would like to ignore branches that match a certain pattern, e.g.:
[snip]
> 
> Is it possible to ignore all branches that match "hjl"?
 

If you mean ignore them when you do "git branch -a", then

git branch -a |grep -v hjl

If you mean "ignore" in some other scenario you need to be more specific about what you want.

Re: Globbing for ignored branches?

[Markus Trippelsdorf]

On 2014.01.24 at 16:37 +0000, Jim Garrison wrote:
> > -----Original Message-----
> > Behalf Of Markus Trippelsdorf
> > Sent: Friday, January 24, 2014 1:01 AM
> > Subject: Globbing for ignored branches?
> > 
> > I would like to ignore branches that match a certain pattern, e.g.:
> [snip]
> > 
> > Is it possible to ignore all branches that match "hjl"?
>  
> 
> If you mean ignore them when you do "git branch -a", then
> 
> git branch -a |grep -v hjl
> 
> If you mean "ignore" in some other scenario you need to be more
> specific about what you want.

I want to them when I run "git pull".

-- 
Markus

Re: Globbing for ignored branches?

[Markus Trippelsdorf]

On 2014.01.24 at 18:07 +0100, Markus Trippelsdorf wrote:
> On 2014.01.24 at 16:37 +0000, Jim Garrison wrote:
> > > -----Original Message-----
> > > Behalf Of Markus Trippelsdorf
> > > Sent: Friday, January 24, 2014 1:01 AM
> > > Subject: Globbing for ignored branches?
> > > 
> > > I would like to ignore branches that match a certain pattern, e.g.:
> > [snip]
> > > 
> > > Is it possible to ignore all branches that match "hjl"?
> >  
> > 
> > If you mean ignore them when you do "git branch -a", then
> > 
> > git branch -a |grep -v hjl
> > 
> > If you mean "ignore" in some other scenario you need to be more
> > specific about what you want.
> 
> I want to them when I run "git pull".
           ignore
-- 
Markus

Re: Globbing for ignored branches?

[Jeff King]

On Fri, Jan 24, 2014 at 06:09:09PM +0100, Markus Trippelsdorf wrote:

> > > If you mean "ignore" in some other scenario you need to be more
> > > specific about what you want.
> > 
> > I want to them when I run "git pull".
>            ignore

I assume you mean that you do not want to fetch them at all, not that
you want to avoid merging them. The set of branches that git fetches is
configured by the fetch "refspec" in your config file. It usually looks
like this:

  $ git config remote.origin.fetch
  +refs/heads/*:refs/remotes/origin/*

But you can specify a specific list of branches you want to fetch
instead:

  $ git config --unset remote.origin.fetch
  $ for i in master other-branch; do
      git config --add remote.origin.fetch \
                       +refs/heads/$i:refs/remotes/origin/$i
    done

However, you do have to specify each branch individually. You probably
want to say "all branches except X", and you cannot currently specify
a negative refspec like that.

-Peff

Re: Globbing for ignored branches?

[Markus Trippelsdorf]

On 2014.01.24 at 13:23 -0500, Jeff King wrote:
> On Fri, Jan 24, 2014 at 06:09:09PM +0100, Markus Trippelsdorf wrote:
> 
> > > > If you mean "ignore" in some other scenario you need to be more
> > > > specific about what you want.
> > > 
> > > I want to them when I run "git pull".
> >            ignore
> 
> I assume you mean that you do not want to fetch them at all, not that
> you want to avoid merging them. The set of branches that git fetches is
> configured by the fetch "refspec" in your config file. It usually looks
> like this:
> 
>   $ git config remote.origin.fetch
>   +refs/heads/*:refs/remotes/origin/*
> 
> But you can specify a specific list of branches you want to fetch
> instead:
> 
>   $ git config --unset remote.origin.fetch
>   $ for i in master other-branch; do
>       git config --add remote.origin.fetch \
>                        +refs/heads/$i:refs/remotes/origin/$i
>     done
> 
> However, you do have to specify each branch individually. You probably
> want to say "all branches except X", and you cannot currently specify
> a negative refspec like that.

Thanks.
Yes, that was the question I wanted to ask (, sorry for not formulating
it more clearly). 
Is this "negative refspec for branches" a feature that is planned for
the future?

-- 
Markus

Re: Globbing for ignored branches?

[Jeff King]

On Fri, Jan 24, 2014 at 07:32:22PM +0100, Markus Trippelsdorf wrote:

> > However, you do have to specify each branch individually. You probably
> > want to say "all branches except X", and you cannot currently specify
> > a negative refspec like that.
> 
> Yes, that was the question I wanted to ask (, sorry for not formulating
> it more clearly). 
> Is this "negative refspec for branches" a feature that is planned for
> the future?

It is something that has been talked about before, but I do not think
anybody is actively working on. It would probably not be too hard a
feature if you are interested in getting your feet wet in git
development. :)

-Peff

Re: Globbing for ignored branches?

[Junio C Hamano]

Jeff King <peff@peff.net> writes:

> On Fri, Jan 24, 2014 at 07:32:22PM +0100, Markus Trippelsdorf wrote:
>
>> > However, you do have to specify each branch individually. You probably
>> > want to say "all branches except X", and you cannot currently specify
>> > a negative refspec like that.
>> 
>> Yes, that was the question I wanted to ask (, sorry for not formulating
>> it more clearly). 
>> Is this "negative refspec for branches" a feature that is planned for
>> the future?
>
> It is something that has been talked about before, but I do not think
> anybody is actively working on. It would probably not be too hard a
> feature if you are interested in getting your feet wet in git
> development. :)

The end result might be not so hard in the mechanical sense, but
designing the interface would be hard.  I do not offhand think of a
good way to do this.

Re: Globbing for ignored branches?

[Markus Trippelsdorf]

On 2014.01.24 at 12:00 -0800, Junio C Hamano wrote:
> Jeff King <peff@peff.net> writes:
> 
> > On Fri, Jan 24, 2014 at 07:32:22PM +0100, Markus Trippelsdorf wrote:
> >
> >> > However, you do have to specify each branch individually. You probably
> >> > want to say "all branches except X", and you cannot currently specify
> >> > a negative refspec like that.
> >> 
> >> Yes, that was the question I wanted to ask (, sorry for not formulating
> >> it more clearly). 
> >> Is this "negative refspec for branches" a feature that is planned for
> >> the future?
> >
> > It is something that has been talked about before, but I do not think
> > anybody is actively working on. It would probably not be too hard a
> > feature if you are interested in getting your feet wet in git
> > development. :)
> 
> The end result might be not so hard in the mechanical sense, but
> designing the interface would be hard.  I do not offhand think of a
> good way to do this.

I don't know if the in-tree regex engine supports negative lookaheads.
If it does, then something like the following should work (to use my
"hjl" example):

^(.(?!hjl))*

-- 
Markus

Re: Globbing for ignored branches?

[Junio C Hamano]

Markus Trippelsdorf <markus@trippelsdorf.de> writes:

> On 2014.01.24 at 12:00 -0800, Junio C Hamano wrote:
>> Jeff King <peff@peff.net> writes:
>> 
>> > On Fri, Jan 24, 2014 at 07:32:22PM +0100, Markus Trippelsdorf wrote:
>> >
>> >> > However, you do have to specify each branch individually. You probably
>> >> > want to say "all branches except X", and you cannot currently specify
>> >> > a negative refspec like that.
>> >> 
>> >> Yes, that was the question I wanted to ask (, sorry for not formulating
>> >> it more clearly). 
>> >> Is this "negative refspec for branches" a feature that is planned for
>> >> the future?
>> >
>> > It is something that has been talked about before, but I do not think
>> > anybody is actively working on. It would probably not be too hard a
>> > feature if you are interested in getting your feet wet in git
>> > development. :)
>> 
>> The end result might be not so hard in the mechanical sense, but
>> designing the interface would be hard.  I do not offhand think of a
>> good way to do this.
>
> I don't know if the in-tree regex engine supports negative lookaheads.
> If it does, then something like the following should work (to use my
> "hjl" example):
>
> ^(.(?!hjl))*

refspec wildcards are *NOT* regular expressions.

Re: Globbing for ignored branches?

[Jeff King]

On Fri, Jan 24, 2014 at 12:00:16PM -0800, Junio C Hamano wrote:

> Jeff King <peff@peff.net> writes:
> 
> > On Fri, Jan 24, 2014 at 07:32:22PM +0100, Markus Trippelsdorf wrote:
> >
> >> > However, you do have to specify each branch individually. You probably
> >> > want to say "all branches except X", and you cannot currently specify
> >> > a negative refspec like that.
> [...]
> The end result might be not so hard in the mechanical sense, but
> designing the interface would be hard.  I do not offhand think of a
> good way to do this.

I had imagined a "not" token at the front of the refspec, like:

  git fetch origin +refs/heads/*:refs/remotes/origin/* ^refs/heads/foo

In this case, a colon in the refspec would be an error. An alternative
would be:

  git fetch origin +refs/heads/*:refs/remotes/origin/* refs/heads/foo:

I.e., to say "put foo to nowhere". But generally refspecs do not affect
each other. So "refs/heads/foo:refs/heads/bar" would generally work _in
addition_ to the other refspec. Making the "null destination" work
differently might be confusing.

I dunno. I have not thought very hard on the topic, so maybe there are
some subtle cases I am missing.

-Peff

Re: Globbing for ignored branches?

[Markus Trippelsdorf]

On 2014.01.24 at 12:44 -0800, Junio C Hamano wrote:
> Markus Trippelsdorf <markus@trippelsdorf.de> writes:
> 
> > On 2014.01.24 at 12:00 -0800, Junio C Hamano wrote:
> >> Jeff King <peff@peff.net> writes:
> >> 
> >> > On Fri, Jan 24, 2014 at 07:32:22PM +0100, Markus Trippelsdorf wrote:
> >> >
> >> >> > However, you do have to specify each branch individually. You probably
> >> >> > want to say "all branches except X", and you cannot currently specify
> >> >> > a negative refspec like that.
> >> >> 
> >> >> Yes, that was the question I wanted to ask (, sorry for not formulating
> >> >> it more clearly). 
> >> >> Is this "negative refspec for branches" a feature that is planned for
> >> >> the future?
> >> >
> >> > It is something that has been talked about before, but I do not think
> >> > anybody is actively working on. It would probably not be too hard a
> >> > feature if you are interested in getting your feet wet in git
> >> > development. :)
> >> 
> >> The end result might be not so hard in the mechanical sense, but
> >> designing the interface would be hard.  I do not offhand think of a
> >> good way to do this.
> >
> > I don't know if the in-tree regex engine supports negative lookaheads.
> > If it does, then something like the following should work (to use my
> > "hjl" example):
> >
> > ^(.(?!hjl))*
> 
> refspec wildcards are *NOT* regular expressions.

Yes, but that's the point. If they were, the "negative refspec"
interface issue would be solved.

-- 
Markus

Re: Globbing for ignored branches?

[Junio C Hamano]

Jeff King <peff@peff.net> writes:

> I had imagined a "not" token at the front of the refspec, like:
>
>   git fetch origin +refs/heads/*:refs/remotes/origin/* ^refs/heads/foo
>
> In this case, a colon in the refspec would be an error. An alternative
> would be:
>
>   git fetch origin +refs/heads/*:refs/remotes/origin/* refs/heads/foo:
>
> I.e., to say "put foo to nowhere". But generally refspecs do not affect
> each other.

Not really.  You do not have to view it as "'not refs/heads/foo' is
affecting the previous '+refs/heads/*:refs/remotes/origin/*'".

You can think of two refspecs "refs/heads/foo refs/heads/bar" are
both affecting the "end result"; so far we only had a single way for
multiple refspecs to affect the end result and that was a "union".
Introducing "subtract" as another mode of combining is not too bad,
I would think, at the conceptual level.

> ... Making the "null destination" work
> differently might be confusing.

I tend to agree that "refs/heads/foo:" is being too cute and may be
confusing, at least if it will be the only way to express this in
the end-user-facing UI.  Even some people were confused enough on a
very sensible "push nothing to ref means deletion" to make us add
another explicit way, "push --delete", to ask for the same thing.

Re: Globbing for ignored branches?

[Jeff King]

On Fri, Jan 24, 2014 at 01:08:42PM -0800, Junio C Hamano wrote:

> Not really.  You do not have to view it as "'not refs/heads/foo' is
> affecting the previous '+refs/heads/*:refs/remotes/origin/*'".
> 
> You can think of two refspecs "refs/heads/foo refs/heads/bar" are
> both affecting the "end result"; so far we only had a single way for
> multiple refspecs to affect the end result and that was a "union".
> Introducing "subtract" as another mode of combining is not too bad,
> I would think, at the conceptual level.

OK, I buy that line of reasoning. I assume that ordering should not
matter (that is, we would apply all positive refspecs, and then subtract
all negative refspecs).

I took a quick look at the refspec code, and how bad it would be to
implement this feature. It's rather a bit of a mess. It looks like there
are three separate code paths to apply refspecs:

  - fetch uses get_ref_map, which calls get_fetch_map for each refspec;
    each refspec than expands into 0, 1, or multiple refs (if it's a
    pattern). You can mention a ref multiple times on the LHS of a
    refspec, and it may be fetched multiple times. After we have the
    whole list, we detect duplicate destinations, and either drop the
    duplicates (if all sources are the same) or complain (if there are
    different sources).

  - push uses match_push_refs, which calls get_ref_match for each ref
    (not refspec). So the loop is inside-out from fetch, and it looks
    like we do weird things with multiple matches. We seem to handle
    multiple explicit matches like:

      $ git push --dry-run tmp master:foo master:bar
      To tmp
       * [new branch]      master -> foo
       * [new branch]      master -> bar

    but we don't seem to do the same for pattern matches:

      $ git push --dry-run tmp refs/heads/*:refs/foo/* \
                               refs/heads/*:refs/bar/*
      To tmp
       * [new branch]      master -> refs/foo/master

    we just take the first match, even though the two did not conflict.
    I doubt this comes up that much, but I do not see any reason this
    should not be doing the same as fetch: apply all refspecs to come up
    with a complete list, then cull duplicates.

  - @{upstream} uses apply_refspecs to convert a single name. This is
    also used by transport-helper's fetch_with_import and
    push_with_import. Which makes me think they do not handle
    overlapping refspecs at all, unlike the builtin counterparts.

    There is also query_refspecs, which underlies apply_refspecs. I'm
    not even sure I understand all of the uses there.

The patch below implements negative refspecs for fetch, but does nothing
for push and apply_refspecs (in fact, it probably makes them worse,
because they've learned to parse negative refspecs, but not handle them
properly).

The helpers in the patch could probably be used to build support for the
other code paths, but it really seems like there could stand to be some
refactoring. I'm not sure if I have the time/stomach for it at the
moment. But I'll post this here anyway in case somebody else is
interested.

> I tend to agree that "refs/heads/foo:" is being too cute and may be
> confusing, at least if it will be the only way to express this in
> the end-user-facing UI.  Even some people were confused enough on a
> very sensible "push nothing to ref means deletion" to make us add
> another explicit way, "push --delete", to ask for the same thing.

Agreed. I went with "^refs/heads/master" in the patch below, but I am
open to other suggestions.

---
diff --git a/builtin/fetch.c b/builtin/fetch.c
index 025bc3e..47f25e9 100644
--- a/builtin/fetch.c
+++ b/builtin/fetch.c
@@ -363,6 +363,8 @@ static struct ref *get_ref_map(struct transport *transport,
 		tail = &rm->next;
 	}
 
+	ref_map = apply_negative_refspecs(ref_map, refspecs, refspec_count);
+
 	return ref_remove_duplicates(ref_map);
 }
 
diff --git a/remote.c b/remote.c
index a89efab..b7b20de 100644
--- a/remote.c
+++ b/remote.c
@@ -16,6 +16,7 @@ static struct refspec s_tag_refspec = {
 	1,
 	0,
 	0,
+	0,
 	"refs/tags/*",
 	"refs/tags/*"
 };
@@ -533,8 +534,14 @@ static struct refspec *parse_refspec_internal(int nr_refspec, const char **refsp
 			rs[i].force = 1;
 			lhs++;
 		}
+		else if (*lhs == '^') {
+			rs[i].negative = 1;
+			lhs++;
+		}
 
 		rhs = strrchr(lhs, ':');
+		if (rs[i].negative && rhs)
+			goto invalid;
 
 		/*
 		 * Before going on, special case ":" (or "+:") as a refspec
@@ -1663,6 +1670,9 @@ int get_fetch_map(const struct ref *remote_refs,
 {
 	struct ref *ref_map, **rmp;
 
+	if (refspec->negative)
+		return 0;
+
 	if (refspec->pattern) {
 		ref_map = get_expanded_map(remote_refs, refspec);
 	} else {
@@ -1705,6 +1715,48 @@ int get_fetch_map(const struct ref *remote_refs,
 	return 0;
 }
 
+static int refspec_match(const struct refspec *refspec,
+			 const char *name)
+{
+	if (refspec->pattern)
+		return match_name_with_pattern(refspec->src, name, NULL, NULL);
+
+	return !strcmp(refspec->src, name);
+}
+
+static int omit_name_by_refspec(const char *name,
+				const struct refspec *refspecs,
+				int nr_refspec)
+{
+	int i;
+
+	for (i = 0; i < nr_refspec; i++) {
+		if (refspecs[i].negative && refspec_match(&refspecs[i], name))
+			return 1;
+	}
+	return 0;
+}
+
+struct ref *apply_negative_refspecs(struct ref *ref_map,
+				    const struct refspec *refspecs,
+				    int nr_refspec)
+{
+	struct ref **tail;
+
+	for (tail = &ref_map; *tail; ) {
+		struct ref *ref = *tail;
+
+		if (omit_name_by_refspec(ref->name, refspecs, nr_refspec)) {
+			*tail = ref->next;
+			free(ref->peer_ref);
+			free(ref);
+		} else
+			tail = &ref->next;
+	}
+
+	return ref_map;
+}
+
 int resolve_remote_symref(struct ref *ref, struct ref *list)
 {
 	if (!ref->symref)
diff --git a/remote.h b/remote.h
index fb7647f..534d365 100644
--- a/remote.h
+++ b/remote.h
@@ -67,6 +67,7 @@ struct refspec {
 	unsigned pattern : 1;
 	unsigned matching : 1;
 	unsigned exact_sha1 : 1;
+	unsigned negative : 1;
 
 	char *src;
 	char *dst;
@@ -157,6 +158,10 @@ int ref_newer(const unsigned char *new_sha1, const unsigned char *old_sha1);
  */
 struct ref *ref_remove_duplicates(struct ref *ref_map);
 
+struct ref *apply_negative_refspecs(struct ref *ref_map,
+				    const struct refspec *refspecs,
+				    int nr_refspecs);
+
 int valid_fetch_refspec(const char *refspec);
 struct refspec *parse_fetch_refspec(int nr_refspec, const char **refspec);
 

Re: Globbing for ignored branches?

[Markus Trippelsdorf]

On 2014.01.24 at 20:34 -0500, Jeff King wrote:
> On Fri, Jan 24, 2014 at 01:08:42PM -0800, Junio C Hamano wrote:
> 
> > Not really.  You do not have to view it as "'not refs/heads/foo' is
> > affecting the previous '+refs/heads/*:refs/remotes/origin/*'".
> > 
> > You can think of two refspecs "refs/heads/foo refs/heads/bar" are
> > both affecting the "end result"; so far we only had a single way for
> > multiple refspecs to affect the end result and that was a "union".
> > Introducing "subtract" as another mode of combining is not too bad,
> > I would think, at the conceptual level.
> 
> > I tend to agree that "refs/heads/foo:" is being too cute and may be
> > confusing, at least if it will be the only way to express this in
> > the end-user-facing UI.  Even some people were confused enough on a
> > very sensible "push nothing to ref means deletion" to make us add
> > another explicit way, "push --delete", to ask for the same thing.
> 
> Agreed. I went with "^refs/heads/master" in the patch below, but I am
> open to other suggestions.

Many thanks for the patch. It seems to work as advertised, but only if
the negative refspec appears on a separate line. For example:

[remote "origin"]
        url = git://gcc.gnu.org/git/gcc.git
        fetch = +refs/heads/*:refs/remotes/origin/*
        fetch = ^refs/remotes/hjl

works fine, but:

[remote "origin"]
        url = git://gcc.gnu.org/git/gcc.git
        fetch = +refs/heads/*:refs/remotes/origin/* ^refs/remotes/hjl 

doesn't. (I think this happens because bad_ref_char in refs.c checks for '^'.)

-- 
Markus

Re: Globbing for ignored branches?

[Markus Trippelsdorf]

On 2014.01.25 at 15:15 +0100, Markus Trippelsdorf wrote:
> On 2014.01.24 at 20:34 -0500, Jeff King wrote:
> > On Fri, Jan 24, 2014 at 01:08:42PM -0800, Junio C Hamano wrote:
> > 
> > > Not really.  You do not have to view it as "'not refs/heads/foo' is
> > > affecting the previous '+refs/heads/*:refs/remotes/origin/*'".
> > > 
> > > You can think of two refspecs "refs/heads/foo refs/heads/bar" are
> > > both affecting the "end result"; so far we only had a single way for
> > > multiple refspecs to affect the end result and that was a "union".
> > > Introducing "subtract" as another mode of combining is not too bad,
> > > I would think, at the conceptual level.
> > 
> > > I tend to agree that "refs/heads/foo:" is being too cute and may be
> > > confusing, at least if it will be the only way to express this in
> > > the end-user-facing UI.  Even some people were confused enough on a
> > > very sensible "push nothing to ref means deletion" to make us add
> > > another explicit way, "push --delete", to ask for the same thing.
> > 
> > Agreed. I went with "^refs/heads/master" in the patch below, but I am
> > open to other suggestions.
> 
> Many thanks for the patch. It seems to work as advertised, but only if
> the negative refspec appears on a separate line. For example:

I've posted a wrong negative refspec. Sorry. Correction below.

> [remote "origin"]
>         url = git://gcc.gnu.org/git/gcc.git
>         fetch = +refs/heads/*:refs/remotes/origin/*
>         fetch = ^refs/remotes/hjl
          fetch = ^refs/remotes/origin/hjl

> works fine, but:
> 
> [remote "origin"]
>         url = git://gcc.gnu.org/git/gcc.git
>         fetch = +refs/heads/*:refs/remotes/origin/* ^refs/remotes/hjl 
          fetch = +refs/heads/*:refs/remotes/origin/* ^refs/remotes/origin/hjl

> doesn't. (I think this happens because bad_ref_char in refs.c checks for '^'.)

-- 
Markus

Re: Globbing for ignored branches?

[Jeff King]

On Sat, Jan 25, 2014 at 03:15:42PM +0100, Markus Trippelsdorf wrote:

> Many thanks for the patch. It seems to work as advertised, but only if
> the negative refspec appears on a separate line. For example:
> 
> [remote "origin"]
>         url = git://gcc.gnu.org/git/gcc.git
>         fetch = +refs/heads/*:refs/remotes/origin/*
>         fetch = ^refs/remotes/hjl
> 
> works fine, but:
> 
> [remote "origin"]
>         url = git://gcc.gnu.org/git/gcc.git
>         fetch = +refs/heads/*:refs/remotes/origin/* ^refs/remotes/hjl 

That does not have anything to do with the negative refspec. The config
format is one refspec per "fetch" key, but you may have as many keys as
you like. Doing:

  [remote "origin"]
  fetch = refs/heads/a:refs/heads/a refs/heads/b:refs/heads/b

is similarly wrong. You need to do:

  [remote "origin"]
  fetch = refs/heads/a:refs/heads/a
  fetch = refs/heads/b:refs/heads/b

instead.  I believe that since space is forbidden in refnames, it should
also be forbidden in refspecs, which means that we could interpret the
first one as you expected without losing backwards compatibility. But I
do not think there is any real advantage to doing so, aside from being
more forgiving. I suspect the documentation in that area could be
improved, though.

-Peff