thr3ads.net - Xen devel - [PATCH] Try to use git commit id if hg changeset is unavailable [Apr 2013]

If this information is useful, please help other people find it:
Share via:

Marek Marczykowski

2013-Apr-10 23:32 UTC

[PATCH] Try to use git commit id if hg changeset is unavailable

As Xen uses git as primary repository, get git commit id for
xen_changeset info.

Maybe hg code should be removed completely?

Signed-off-by: Marek Marczykowski <marmarek@invisiblethingslab.com>
---
 xen/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/xen/Makefile b/xen/Makefile
index 0fb3db7..9e4e6f3 100644
--- a/xen/Makefile
+++ b/xen/Makefile
@@ -126,7 +126,7 @@ include/xen/compile.h: include/xen/compile.h.in .banner
 	    -e ''s/@@version@@/$(XEN_VERSION)/g'' \
 	    -e ''s/@@subversion@@/$(XEN_SUBVERSION)/g'' \
 	    -e ''s/@@extraversion@@/$(XEN_EXTRAVERSION)/g'' \
-	    -e ''s!@@changeset@@!$(shell ((hg parents --template
"{date|date} {rev}:{node|short}" >/dev/null && hg parents
--template "{date|date} {rev}:{node|short}") || echo
"unavailable") 2>/dev/null)!g'' \
+	    -e ''s!@@changeset@@!$(shell ((hg parents --template
"{date|date} {rev}:{node|short}" >/dev/null && hg parents
--template "{date|date} {rev}:{node|short}") || git show -s
--pretty="%ad %h" || echo "unavailable") 2>/dev/null | tr
-d ''\n'')!g'' \
 	    < include/xen/compile.h.in > $@.new
 	@grep \" .banner >> $@.new
 	@grep -v \" .banner
-- 
1.8.1.4

Keir Fraser

2013-Apr-11 06:40 UTC

head link

Re: [PATCH] Try to use git commit id if hg changeset is unavailable

On 11/04/2013 00:32, "Marek Marczykowski"
<marmarek@invisiblethingslab.com>
wrote:
> As Xen uses git as primary repository, get git commit id for
> xen_changeset info.
> 
> Maybe hg code should be removed completely?
> 
> Signed-off-by: Marek Marczykowski <marmarek@invisiblethingslab.com>
I think we could just remove the hg code. Any other opinions?

 -- Keir
> ---
>  xen/Makefile | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/xen/Makefile b/xen/Makefile
> index 0fb3db7..9e4e6f3 100644
> --- a/xen/Makefile
> +++ b/xen/Makefile
> @@ -126,7 +126,7 @@ include/xen/compile.h: include/xen/compile.h.in .banner
>    -e ''s/@@version@@/$(XEN_VERSION)/g'' \
>    -e ''s/@@subversion@@/$(XEN_SUBVERSION)/g'' \
>    -e ''s/@@extraversion@@/$(XEN_EXTRAVERSION)/g'' \
> -     -e ''s!@@changeset@@!$(shell ((hg parents --template
"{date|date}
> {rev}:{node|short}" >/dev/null && hg parents --template
"{date|date}
> {rev}:{node|short}") || echo "unavailable")
2>/dev/null)!g'' \
> +     -e ''s!@@changeset@@!$(shell ((hg parents --template
"{date|date}
> {rev}:{node|short}" >/dev/null && hg parents --template
"{date|date}
> {rev}:{node|short}") || git show -s --pretty="%ad %h" ||
echo "unavailable")
> 2>/dev/null | tr -d ''\n'')!g'' \
>    < include/xen/compile.h.in > $@.new
> @grep \" .banner >> $@.new
> @grep -v \" .banner

Jan Beulich

2013-Apr-11 07:28 UTC

head link

Re: [PATCH] Try to use git commit id if hg changeset is unavailable

>>> On 11.04.13 at 08:40, Keir Fraser <keir.xen@gmail.com> wrote:
> On 11/04/2013 00:32, "Marek Marczykowski"
<marmarek@invisiblethingslab.com> wrote:
> 
>> As Xen uses git as primary repository, get git commit id for
>> xen_changeset info.
>> 
>> Maybe hg code should be removed completely?
>> 
>> Signed-off-by: Marek Marczykowski
<marmarek@invisiblethingslab.com>
> 
> I think we could just remove the hg code. Any other opinions?
No. But if we someone wants it to be retained, the git invocation
should precede the hg one in any case.

Jan
>> ---
>>  xen/Makefile | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>> 
>> diff --git a/xen/Makefile b/xen/Makefile
>> index 0fb3db7..9e4e6f3 100644
>> --- a/xen/Makefile
>> +++ b/xen/Makefile
>> @@ -126,7 +126,7 @@ include/xen/compile.h: include/xen/compile.h.in
.banner
>>    -e ''s/@@version@@/$(XEN_VERSION)/g'' \
>>    -e ''s/@@subversion@@/$(XEN_SUBVERSION)/g'' \
>>    -e ''s/@@extraversion@@/$(XEN_EXTRAVERSION)/g'' \
>> -     -e ''s!@@changeset@@!$(shell ((hg parents --template
"{date|date}
>> {rev}:{node|short}" >/dev/null && hg parents --template
"{date|date}
>> {rev}:{node|short}") || echo "unavailable")
2>/dev/null)!g'' \
>> +     -e ''s!@@changeset@@!$(shell ((hg parents --template
"{date|date}
>> {rev}:{node|short}" >/dev/null && hg parents --template
"{date|date}
>> {rev}:{node|short}") || git show -s --pretty="%ad %h" ||
echo "unavailable")
>> 2>/dev/null | tr -d ''\n'')!g'' \
>>    < include/xen/compile.h.in > $@.new
>> @grep \" .banner >> $@.new
>> @grep -v \" .banner
> 
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org 
> http://lists.xen.org/xen-devel

Sander Eikelenboom

2013-Apr-11 07:33 UTC

head link

Re: [PATCH] Try to use git commit id if hg changeset is unavailable

Thursday, April 11, 2013, 1:32:05 AM, you wrote:
> As Xen uses git as primary repository, get git commit id for
> xen_changeset info.
> Maybe hg code should be removed completely?
> Signed-off-by: Marek Marczykowski <marmarek@invisiblethingslab.com>
> ---
>  xen/Makefile | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> diff --git a/xen/Makefile b/xen/Makefile
> index 0fb3db7..9e4e6f3 100644
> --- a/xen/Makefile
> +++ b/xen/Makefile
> @@ -126,7 +126,7 @@ include/xen/compile.h: include/xen/compile.h.in .banner
>             -e ''s/@@version@@/$(XEN_VERSION)/g'' \
>             -e ''s/@@subversion@@/$(XEN_SUBVERSION)/g'' \
>             -e ''s/@@extraversion@@/$(XEN_EXTRAVERSION)/g''
\
> -           -e ''s!@@changeset@@!$(shell ((hg parents --template
"{date|date} {rev}:{node|short}" >/dev/null && hg parents
--template "{date|date} {rev}:{node|short}") || echo
"unavailable") 2>/dev/null)!g'' \
> +           -e ''s!@@changeset@@!$(shell ((hg parents --template
"{date|date} {rev}:{node|short}" >/dev/null && hg parents
--template "{date|date} {rev}:{node|short}") || git show -s
--pretty="%ad %h" || echo "unavailable") 2>/dev/null | tr
-d ''\n'')!g'' \
>             < include/xen/compile.h.in > $@.new
>         @grep \" .banner >> $@.new
>         @grep -v \" .banner
Hi Marek,

Yesterday i was looking into the same, i tried to also reflect if the tree was
dirty or not (after this discussion:
http://lists.xen.org/archives/html/xen-devel/2013-02/msg02077.html).

Haven''t got it in a nice oneliner though.

Following that discussion the tree can be a: git tree, mercurial tree, tarball

For both the git and mercurial tree, changes are tracked and it should be able
to mark the tree as dirty when there are changes not in the upstream tree:
a) uncommited changes (and untracked files)
b) local commits (outgoing/not pushed)

Git does have a option to mark the tree dirty for uncommitted changes,
but probably due it''s philosophy of "my tree is as good as
yours",
this doesn''t work if the changes have been only locally committed and
not pushed.

Mercurial doesn''t have a facility to mark the tree as dirty.

Another thing is the rev/node/hash to show, to me it would be preferable to:
- get the last upstream commit this local tree was based on.
- appended with -dirty if local commits or uncommitted changes dirty the tree.

That''s what i have tried with the script below,
but getting the upstream commit for mercurial proves difficult because
some functions from hg''s revset (like remote() or outgoing())
seem unavailable on Debian squeeze.


script:

        local dirty=false
        local changeset=""

        # Check for git and a git repo.
        if test -d "../.git" && changeset="`git rev-parse
--verify --long origin/HEAD 2>/dev/null`"; then

                # If there are differences with remote, commited, uncommited or
untracked, mark as dirty
                if [ -n "`git diff-index origin/HEAD 2>/dev/null`"
-o -n "`git status --porcelain 2>/dev/null`" ]; then
                    changeset="$changeset-dirty"
                fi

                changeset="git $changeset"
        fi

        # Check for mercurial and a mercurial repo.
        if [ -z "$changeset" ] && test -d "../.hg"
&& hg parents --template ''{date|date}
{rev}:{node|short}'' >/dev/null && changeset="`hg
parents --template ''{rev}:{node|short}''`"; then

                # If there are differences with remote, outgoing commits,
uncommited or untracked, mark as dirty
                if [ -z "`hg outgoing | grep ''no changes
found'' 2> /dev/null`" -o -n "`hg id | grep
''+''`" ]; then
                    changeset="$changeset-dirty"
                fi

                changeset="mercurial $changeset"
        fi

        if [ -z "$changeset" ] && test -f
"../.hg_archival.txt"; then

            # Determine changeset from node entry in
".hg_archival.txt"
            changeset="tarball $changeset"
        fi

        if [ -z "$changeset" ]; then
            changeset="unavailable"
        fi

echo "changeset: $changeset"






  --
  Sander

Ian Campbell

2013-Apr-11 07:41 UTC

head link

Re: [PATCH] Try to use git commit id if hg changeset is unavailable

On Thu, 2013-04-11 at 08:33 +0100, Sander Eikelenboom
wrote:> Thursday, April 11, 2013, 1:32:05 AM, you wrote:
> 
> > As Xen uses git as primary repository, get git commit id for
> > xen_changeset info.
> 
> > Maybe hg code should be removed completely?
> 
> > Signed-off-by: Marek Marczykowski
<marmarek@invisiblethingslab.com>
> > ---
> >  xen/Makefile | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> > diff --git a/xen/Makefile b/xen/Makefile
> > index 0fb3db7..9e4e6f3 100644
> > --- a/xen/Makefile
> > +++ b/xen/Makefile
> > @@ -126,7 +126,7 @@ include/xen/compile.h: include/xen/compile.h.in
.banner
> >             -e ''s/@@version@@/$(XEN_VERSION)/g'' \
> >             -e
''s/@@subversion@@/$(XEN_SUBVERSION)/g'' \
> >             -e
''s/@@extraversion@@/$(XEN_EXTRAVERSION)/g'' \
> > -           -e ''s!@@changeset@@!$(shell ((hg parents
--template "{date|date} {rev}:{node|short}" >/dev/null &&
hg parents --template "{date|date} {rev}:{node|short}") || echo
"unavailable") 2>/dev/null)!g'' \
> > +           -e ''s!@@changeset@@!$(shell ((hg parents
--template "{date|date} {rev}:{node|short}" >/dev/null &&
hg parents --template "{date|date} {rev}:{node|short}") || git show -s
--pretty="%ad %h" || echo "unavailable") 2>/dev/null | tr
-d ''\n'')!g'' \
> >             < include/xen/compile.h.in > $@.new
> >         @grep \" .banner >> $@.new
> >         @grep -v \" .banner
> 
> Hi Marek,
> 
> Yesterday i was looking into the same, i tried to also reflect if the tree
was
> dirty or not (after this discussion:
> http://lists.xen.org/archives/html/xen-devel/2013-02/msg02077.html).
> 
> Haven''t got it in a nice oneliner though.
The one liner is getting a bit silly. Perhaps we should import something
like Linux''s scripts/setlocalversion.
> Following that discussion the tree can be a: git tree, mercurial tree,
tarball
> 
> For both the git and mercurial tree, changes are tracked and it should be
able
> to mark the tree as dirty when there are changes not in the upstream tree:
> a) uncommited changes (and untracked files)
> b) local commits (outgoing/not pushed)
> 
> Git does have a option to mark the tree dirty for uncommitted changes,
> but probably due it''s philosophy of "my tree is as good as
yours",
> this doesn''t work if the changes have been only locally committed
and not pushed.
> 
> Mercurial doesn''t have a facility to mark the tree as dirty.
> 
> Another thing is the rev/node/hash to show, to me it would be preferable
to:
> - get the last upstream commit this local tree was based on.
> - appended with -dirty if local commits or uncommitted changes dirty the
tree.
I think it is best to show the actual literal commit which was used,
whether or not that has been pushed. That''s useful to me as a dev when
I
am building locally and when users report things with commits which
don''t appear in the main tree we can easily say "what additinal
commits
do you have"

Ian.

Sander Eikelenboom

2013-Apr-11 08:25 UTC

head link

Re: [PATCH] Try to use git commit id if hg changeset is unavailable

Thursday, April 11, 2013, 9:41:15 AM, you wrote:
> Delivery-date: Thu, 11 Apr 2013 09:41:29 +0200
> Received: from Debian-exim by smtp.eikelenboom.it with spam-scanned (Exim
4.72)
>         id 1UQC8D-0005xX-Oc
>         for linux@eikelenboom.it; Thu, 11 Apr 2013 09:41:29 +0200
> X-Spam-Checker-Version: SpamAssassin 3.3.2 (2011-06-06) on
>         vserver.eikelenboom.it
> X-Spam-Level: 
> X-Spam-Status: No, score=-11.8 required=5.0 tests=BAYES_00=-1.9,
>         RP_MATCHES_RCVD=-2.373,SPF_HELO_PASS=-0.001,SPF_PASS=-0.001,
>         USER_IN_WHITELIST=-7.5 autolearn=ham version=3.3.2
> X-Spam-Language: en
> Received: from smtp.eu.citrix.com ([46.33.159.39]:29476)
>         by smtp.eikelenboom.it with esmtps (TLS1.0:RSA_ARCFOUR_SHA1:16)
>         (Exim 4.72)
>         (envelope-from <Ian.Campbell@citrix.com>)
>         id 1UQC8D-0005xS-Ll
>         for linux@eikelenboom.it; Thu, 11 Apr 2013 09:41:17 +0200
> X-IronPort-AV: E=Sophos;i="4.87,454,1363132800"; 
>    d="scan''208";a="3446646"
> Received: from lonpmailmx01.citrite.net ([10.30.203.162])
>   by LONPIPO01.EU.CITRIX.COM with ESMTP/TLS/RC4-MD5; 11 Apr 2013 07:41:17
+0000
> Received: from [10.80.2.42] (10.80.2.42) by LONPMAILMX01.citrite.net
>  (10.30.203.162) with Microsoft SMTP Server id 8.3.298.1; Thu, 11 Apr 2013
>  08:41:16 +0100
> Message-ID: <1365666075.27868.111.camel@zakaz.uk.xensource.com>
> Subject: Re: [Xen-devel] [PATCH] Try to use git commit id if hg changeset
is
>  unavailable
> From: Ian Campbell <Ian.Campbell@citrix.com>
> To: Sander Eikelenboom <linux@eikelenboom.it>
> CC: Marek Marczykowski <marmarek@invisiblethingslab.com>,
>         "xen-devel@lists.xen.org" <xen-devel@lists.xen.org>
> Date: Thu, 11 Apr 2013 08:41:15 +0100
> In-Reply-To: <3310359051.20130411093331@eikelenboom.it>
> References: <20130410233918.66DF61D3@duch.mimuw.edu.pl>
>          <3310359051.20130411093331@eikelenboom.it>
> Organization: Citrix Systems, Inc.
> Content-Type: text/plain; charset="UTF-8"
> X-Mailer: Evolution 3.4.4-1 
> MIME-Version: 1.0
> Content-Transfer-Encoding: 7bit
> On Thu, 2013-04-11 at 08:33 +0100, Sander Eikelenboom wrote:
>> Thursday, April 11, 2013, 1:32:05 AM, you wrote:
>> 
>> > As Xen uses git as primary repository, get git commit id for
>> > xen_changeset info.
>> 
>> > Maybe hg code should be removed completely?
>> 
>> > Signed-off-by: Marek Marczykowski
<marmarek@invisiblethingslab.com>
>> > ---
>> >  xen/Makefile | 2 +-
>> >  1 file changed, 1 insertion(+), 1 deletion(-)
>> 
>> > diff --git a/xen/Makefile b/xen/Makefile
>> > index 0fb3db7..9e4e6f3 100644
>> > --- a/xen/Makefile
>> > +++ b/xen/Makefile
>> > @@ -126,7 +126,7 @@ include/xen/compile.h:
include/xen/compile.h.in .banner
>> >             -e ''s/@@version@@/$(XEN_VERSION)/g''
\
>> >             -e
''s/@@subversion@@/$(XEN_SUBVERSION)/g'' \
>> >             -e
''s/@@extraversion@@/$(XEN_EXTRAVERSION)/g'' \
>> > -           -e ''s!@@changeset@@!$(shell ((hg parents
--template "{date|date} {rev}:{node|short}" >/dev/null &&
hg parents --template "{date|date} {rev}:{node|short}") || echo
"unavailable") 2>/dev/null)!g'' \
>> > +           -e ''s!@@changeset@@!$(shell ((hg parents
--template "{date|date} {rev}:{node|short}" >/dev/null &&
hg parents --template "{date|date} {rev}:{node|short}") || git show -s
--pretty="%ad %h" || echo "unavailable") 2>/dev/null | tr
-d ''\n'')!g'' \
>> >             < include/xen/compile.h.in > $@.new
>> >         @grep \" .banner >> $@.new
>> >         @grep -v \" .banner
>> 
>> Hi Marek,
>> 
>> Yesterday i was looking into the same, i tried to also reflect if the
tree was
>> dirty or not (after this discussion:
>> http://lists.xen.org/archives/html/xen-devel/2013-02/msg02077.html).
>> 
>> Haven''t got it in a nice oneliner though.
> The one liner is getting a bit silly. Perhaps we should import something
> like Linux''s scripts/setlocalversion.
>> Following that discussion the tree can be a: git tree, mercurial tree,
tarball
>> 
>> For both the git and mercurial tree, changes are tracked and it should
be able
>> to mark the tree as dirty when there are changes not in the upstream
tree:
>> a) uncommited changes (and untracked files)
>> b) local commits (outgoing/not pushed)
>> 
>> Git does have a option to mark the tree dirty for uncommitted changes,
>> but probably due it''s philosophy of "my tree is as good
as yours",
>> this doesn''t work if the changes have been only locally
committed and not pushed.
>> 
>> Mercurial doesn''t have a facility to mark the tree as dirty.
>> 
>> Another thing is the rev/node/hash to show, to me it would be
preferable to:
>> - get the last upstream commit this local tree was based on.
>> - appended with -dirty if local commits or uncommitted changes dirty
the tree.
> I think it is best to show the actual literal commit which was used,
> whether or not that has been pushed. That''s useful to me as a dev
when I
> am building locally and when users report things with commits which
> don''t appear in the main tree we can easily say "what
additinal commits
> do you have"
Ok, the linux script "setlocalversion" seems to also use tags to
derive the version,
but since the last tag is of the previous tagged release (for unstable) this
seems to confuse.
(as is with linux during the merge window, which generates a kernel with the
name of the previous version)

And since the version, subversion and extraversion are already derived by other
means,
i don''t think that the git tags are needed for describing the
changeset.
So the changeset should only reflect the hash of the last commit with a dirty
flag appended to make it easy
to see if the tree was dirty.

So the script shouldn''t be a verbatim copy from linux i guess.

--

Sander


> Ian.

Xen devel - Apr 2013 - [PATCH] Try to use git commit id if hg changeset is unavailable

[PATCH] Try to use git commit id if hg changeset is unavailable

Re: [PATCH] Try to use git commit id if hg changeset is unavailable

Re: [PATCH] Try to use git commit id if hg changeset is unavailable

Re: [PATCH] Try to use git commit id if hg changeset is unavailable

Re: [PATCH] Try to use git commit id if hg changeset is unavailable

Re: [PATCH] Try to use git commit id if hg changeset is unavailable