[PATCH] wireproto: allow unbundle with hashed heads parameter (issue2126)

[Steven Brown]

On 16 April 2011 11:26, Shuhei Takahashi <takahashi.shuhei at gmail.com> wrote:
> # HG changeset patch
> # User Shuhei Takahashi <takahashi.shuhei at gmail.com>
> # Date 1302883556 -32400
> # Node ID 497aab7b3f8fc038daa6a542f3014477fa8451cf
> # Parent  ed22a26fc7c6d7ee436c8b4dc33ff1efc349a522
> wireproto: allow unbundle with hashed heads parameter (issue2126)
>
> Current wire protocol of unbundle sends the list of all heads in the remote
> repository to avoid race condition.  This causes "URL too long" error on HTTP
> server when the repository has many heads.
>
> This change allows clients to send SHA1 hash of sorted head hashes instead.
> Also, this introduces "unbundlehash" capability to inform them that the server
> accepts hashed heads parameter.
>
> diff -r ed22a26fc7c6 -r 497aab7b3f8f mercurial/wireproto.py
> --- a/mercurial/wireproto.py    Fri Apr 15 20:35:59 2011 -0500
> +++ b/mercurial/wireproto.py    Sat Apr 16 01:05:56 2011 +0900
> @@ -139,7 +139,13 @@
>         remote server as a bundle. Return an integer indicating the
>         result of the push (see localrepository.addchangegroup()).'''
>
> -        ret, output = self._callpush("unbundle", cg, heads=encodelist(heads))
> +        if self.capable('unbundlehash'):
> +            heads = encodelist(['hashed',
> +                                util.sha1(''.join(sorted(heads))).digest()])
> +        else:
> +            heads = encodelist(heads)
> +
> +        ret, output = self._callpush("unbundle", cg, heads=heads)
>         if ret == "":
>             raise error.ResponseError(
>                 _('push failed:'), output)
> @@ -216,7 +222,8 @@
>     return "".join(r)
>
>  def capabilities(repo, proto):
> -    caps = 'lookup changegroupsubset branchmap pushkey known getbundle'.split()
> +    caps = ('lookup changegroupsubset branchmap pushkey known getbundle ' +
> +            'unbundlehash').split()
>     if _allowstream(repo.ui):
>         requiredformats = repo.requirements & repo.supportedformats
>         # if our local revlogs are just revlogv1, add 'stream' cap
> @@ -353,7 +360,9 @@
>
>     def check_heads():
>         heads = repo.heads()
> -        return their_heads == ['force'] or their_heads == heads
> +        heads_hash = util.sha1(''.join(sorted(heads))).digest()
> +        return (their_heads == ['force'] or their_heads == heads or
> +                their_heads == ['hashed', heads_hash])
>
>     proto.redirect()
>
> diff -r ed22a26fc7c6 -r 497aab7b3f8f tests/test-hgweb-commands.t
> --- a/tests/test-hgweb-commands.t       Fri Apr 15 20:35:59 2011 -0500
> +++ b/tests/test-hgweb-commands.t       Sat Apr 16 01:05:56 2011 +0900
> @@ -943,7 +943,7 @@
>   $ "$TESTDIR/get-with-headers.py" 127.0.0.1:$HGPORT '?cmd=capabilities'; echo
>   200 Script output follows
>
> -  lookup changegroupsubset branchmap pushkey known getbundle unbundle=HG10GZ,HG10BZ,HG10UN
> +  lookup changegroupsubset branchmap pushkey known getbundle unbundlehash unbundle=HG10GZ,HG10BZ,HG10UN
>
>  heads
>
> diff -r ed22a26fc7c6 -r 497aab7b3f8f tests/test-unbundlehash.t
> --- /dev/null   Thu Jan 01 00:00:00 1970 +0000
> +++ b/tests/test-unbundlehash.t Sat Apr 16 01:05:56 2011 +0900
> @@ -0,0 +1,31 @@
> +
> +Test wire protocol unbundle with hashed heads (capability: unbundlehash)
> +
> +Create a remote repository.
> +
> +  $ hg init remote
> +  $ hg serve -R remote --config web.push_ssl=False --config web.allow_push=* -p $HGPORT -d --pid-file=hg1.pid -E error.log -A access.log
> +  $ cat hg1.pid >> $DAEMON_PIDS
> +
> +Clone the repository and push a change.
> +
> +  $ hg clone http://localhost:$HGPORT/ local
> +  no changes found
> +  updating to branch default
> +  0 files updated, 0 files merged, 0 files removed, 0 files unresolved
> +  $ touch local/README
> +  $ hg ci -R local -A -m hoge
> +  adding README
> +  $ hg push -R local
> +  pushing to http://localhost:$HGPORT/
> +  searching for changes
> +  remote: adding changesets
> +  remote: adding manifests
> +  remote: adding file changes
> +  remote: added 1 changesets with 1 changes to 1 files
> +
> +Ensure hashed heads format is used.
> +The hash here is always the same since the remote repository only has the null head.
> +
> +  $ cat access.log | grep unbundle
> +  * - - [*] "POST /?cmd=unbundle&heads=686173686564+6768033e216468247bd031a0a2d9876d79818f8f HTTP/1.1" 200 - (glob)
> _______________________________________________
> Mercurial-devel mailing list
> Mercurial-devel at selenic.com
> http://selenic.com/mailman/listinfo/mercurial-devel
>

This is similar to a patch that I submitted in January.

Differences in my patch:
a) It generated the unbundle hash on the server, instead of the client.
   This makes it trivial to change the value in later releases.
b) It added more tests.

I now think that a) is unnecessary, but would you please include my tests
in your patch?

- test-push-cgi.t: successful push, heads in sha1 format (see issue2126)
- test-push-http.t: expect success, server lacks the prepush capability

See:
http://thread.gmane.org/gmane.comp.version-control.mercurial.devel/37759