This is an archive of the discontinued Mercurial Phabricator instance.

Differential D6825

hgext: start building a library for simple hooks
ClosedPublic

Authored by joerg.sonnenberger on Sep 7 2019, 8:50 AM.

Details

Reviewers
baymax
durin42
Group Reviewers
hg-reviewers
Commits
rHG4cabeea6d214: hgext: start building a library for simple hooks
Summary

Many workflows depend on hooks to enforce certain policies, e.g. to
prevent forced pushes. The Mercurial Guide includes some cases and
Google can help finding others, but it can save users a lot of time
if hg itself has a couple of examples for further customization.

Diff Detail

Repository
rHG Mercurial
Lint
Automatic diff as part of commit; lint not applicable.
Unit
Automatic diff as part of commit; unit tests not applicable.

Event Timeline

pulkit added a subscriber: pulkit.Sep 7 2019, 12:20 PM

Thanks a lot for putting efforts here. I am +1 on this. I haven't looked at the code though, will try to review if no one else beats me to it.

indygreg added a subscriber: indygreg.Sep 7 2019, 1:27 PM

I think we should go a step further and have some built-in hooks in Mercurial itself, like we do extensions. A good first implementation would be to define those hooks somewhere where they can be imported, add tests like they are standalone hooks. As a follow-up (read: slightly more work), we can define a new config syntax for using these hooks. e.g. builtin:reject_new_heads.

What do others think?

durin42 added a subscriber: durin42.Sep 9 2019, 11:29 AM

I like where this is going, but I wonder if these specific hooks can be written without being in-process Python hooks. Reason being that Python hooks use the unstable internals of hg and are semi-discouraged if they're avoidable.

Thoughts?

If you do stick with in-process for this, I definitely want a test included so we don't regress our own examples. ;)

indygreg added a comment.Sep 9 2019, 9:54 PM
In D6825#100146, @durin42 wrote:

I like where this is going, but I wonder if these specific hooks can be written without being in-process Python hooks. Reason being that Python hooks use the unstable internals of hg and are semi-discouraged if they're avoidable.

This is exactly why I think the Python hooks should be part of the modules shipped as part of the core distribution :)

baymax requested changes to this revision.Jan 24 2020, 12:32 PM

There seems to have been no activities on this Diff for the past 3 Months.

By policy, we are automatically moving it out of the need-review state.

Please, move it back to need-review without hesitation if this diff should still be discussed.

:baymax:need-review-idle:

This revision now requires changes to proceed.Jan 24 2020, 12:32 PM
joerg.sonnenberger retitled this revision from contrib: start building a library for simple hooks to hgext: start building a library for simple hooks.Feb 16 2020, 12:03 AM
joerg.sonnenberger updated this revision to Diff 20249.
indygreg added inline comments.Feb 28 2020, 12:50 AM
hgext/hooklib/__init__.py
7

Assuming this is related to the IRC page.

When the demand importer is active, an import will materialize a proxy module object: the code in that module won't get executed until an attribute of the module is accessed.

So anything relying on run-on-import side-effects may not work correctly. Since our configuration option registration mechanism relies on code running as a side-effect of module import, in order to ensure this code runs, you need to ensure the module is "loaded" by accessing any of its attributes.

In addition, the extension mechanism may be looking for a configtable attribute in the module in order to associate config options with extensions. If this isn't actually happening (I'm too lazy to look at the source code), then you won't need to assign to a local variable here: simply accessing changeset_published.configtable will be sufficient to trigger module execution and register the config options.

durin42 added subscribers: marmoute, yuja, martinvonz.Feb 28 2020, 11:50 AM
durin42 accepted this revision as: durin42.

I'm happy with this, modulo my one question for other maintainers. If anyone else is happy, please feel encouraged to push it.

hgext/hooklib/__init__.py
2

Should we mark this as (EXPERIMENTAL) or are we comfortable holding on to this long-term?

/cc @marmoute @martinvonz @indygreg @yuja

joerg.sonnenberger added inline comments.Feb 28 2020, 4:42 PM
hgext/hooklib/__init__.py
2

I'm perfectly fine with adding such a marker and a comment that the functionality is meant for long-term support, but e.g. might migrate into the notify extension. One of the big goals here is to ensure in-tree testing after all.

martinvonz added inline comments.Feb 28 2020, 4:49 PM
hgext/hooklib/__init__.py
2

I'd prefer to mark it experimental to start with.

joerg.sonnenberger added a commit: rHG4cabeea6d214: hgext: start building a library for simple hooks.Mar 3 2020, 1:48 PM
This revision was not accepted when it landed; it landed in state Needs Review.
Closed by commit rHG4cabeea6d214: hgext: start building a library for simple hooks (authored by joerg.sonnenberger). · Explain Why
This revision was automatically updated to reflect the committed changes.

Revision Contents
Changeset List

Diff 20430

contrib/import-checker.py

prefix = absname + '.' prefix = absname + '.'
for n in node.names: for n in node.names:
found = fromlocal(prefix + n.name) found = fromlocal(prefix + n.name)
if not found: if not found:
# this should be a function or a property of "node.module" # this should be a function or a property of "node.module"
modnotfound = True modnotfound = True
continue continue
yield found[1] yield found[1]
if modnotfound: if modnotfound and dottedpath != modulename:
# "dottedpath" is a package, but imported because of non-module # "dottedpath" is a package, but imported because of non-module
# lookup # lookup
# specifically allow "from . import foo" from __init__.py
yield dottedpath yield dottedpath
def verify_import_convention(module, source, localmods): def verify_import_convention(module, source, localmods):
"""Verify imports match our established coding convention. """Verify imports match our established coding convention.
We have 2 conventions: legacy and modern. The modern convention is in We have 2 conventions: legacy and modern. The modern convention is in
effect when using absolute imports. effect when using absolute imports.

hgext/hooklib/__init__.py

  • This file was added.
"""collection of simple hooks for common tasks (EXPERIMENTAL)
durin42Unsubmitted
Not Done

Should we mark this as (EXPERIMENTAL) or are we comfortable holding on to this long-term?

/cc @marmoute @martinvonz @indygreg @yuja

durin42: Should we mark this as (EXPERIMENTAL) or are we comfortable holding on to this long-term? /cc…
joerg.sonnenbergerAuthorUnsubmitted
Done

I'm perfectly fine with adding such a marker and a comment that the functionality is meant for long-term support, but e.g. might migrate into the notify extension. One of the big goals here is to ensure in-tree testing after all.

joerg.sonnenberger: I'm perfectly fine with adding such a marker and a comment that the functionality is meant for…
martinvonzUnsubmitted
Not Done

I'd prefer to mark it experimental to start with.

martinvonz: I'd prefer to mark it experimental to start with.
This extension provides a number of simple hooks to handle issues
commonly found in repositories with many contributors:
- email notification when changesets move from draft to public phase
- email notification when changesets are obsoleted
- enforcement of draft phase for all incoming changesets
indygregUnsubmitted
Not Done

Assuming this is related to the IRC page.

When the demand importer is active, an import will materialize a proxy module object: the code in that module won't get executed until an attribute of the module is accessed.

So anything relying on run-on-import side-effects may not work correctly. Since our configuration option registration mechanism relies on code running as a side-effect of module import, in order to ensure this code runs, you need to ensure the module is "loaded" by accessing any of its attributes.

In addition, the extension mechanism may be looking for a configtable attribute in the module in order to associate config options with extensions. If this isn't actually happening (I'm too lazy to look at the source code), then you won't need to assign to a local variable here: simply accessing changeset_published.configtable will be sufficient to trigger module execution and register the config options.

indygreg: Assuming this is related to the IRC page. When the demand importer is active, an `import` will…
- enforcement of a no-branch-merge policy
- enforcement of a no-multiple-heads policy
The implementation of the hooks is subject to change, e.g. whether to
implement them as individual hooks or merge them into the notify
extension as option. The functionality itself is planned to be supported
long-term.
"""
from __future__ import absolute_import
from . import (
changeset_obsoleted,
changeset_published,
)
# configtable is only picked up from the "top-level" module of the extension,
# so expand it here to ensure all items are properly loaded
configtable = {}
configtable.update(changeset_published.configtable)
configtable.update(changeset_obsoleted.configtable)

hgext/hooklib/changeset_obsoleted.py

  • This file was added.
# Copyright 2020 Joerg Sonnenberger <joerg@bec.de>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
"""changeset_obsoleted is a hook to send a mail when an
existing draft changeset is obsoleted by an obsmarker without successor.
Correct message threading requires the same messageidseed to be used for both
the original notification and the new mail.
Usage:
[notify]
messageidseed = myseed
[hooks]
pretxnclose.changeset_obsoleted = \
python:hgext.hooklib.changeset_obsoleted.hook
"""
from __future__ import absolute_import
import email.errors as emailerrors
import email.utils as emailutils
from mercurial.i18n import _
from mercurial import (
encoding,
error,
logcmdutil,
mail,
obsutil,
pycompat,
registrar,
)
from mercurial.utils import dateutil
from .. import notify
configtable = {}
configitem = registrar.configitem(configtable)
configitem(
b'notify_obsoleted', b'domain', default=None,
)
configitem(
b'notify_obsoleted', b'messageidseed', default=None,
)
configitem(
b'notify_obsoleted',
b'template',
default=b'''Subject: changeset abandoned
This changeset has been abandoned.
''',
)
def _report_commit(ui, repo, ctx):
domain = ui.config(b'notify_obsoleted', b'domain') or ui.config(
b'notify', b'domain'
)
messageidseed = ui.config(
b'notify_obsoleted', b'messageidseed'
) or ui.config(b'notify', b'messageidseed')
template = ui.config(b'notify_obsoleted', b'template')
spec = logcmdutil.templatespec(template, None)
templater = logcmdutil.changesettemplater(ui, repo, spec)
ui.pushbuffer()
n = notify.notifier(ui, repo, b'incoming')
subs = set()
for sub, spec in n.subs:
if spec is None:
subs.add(sub)
continue
revs = repo.revs(b'%r and %d:', spec, ctx.rev())
if len(revs):
subs.add(sub)
continue
if len(subs) == 0:
ui.debug(
b'notify_obsoleted: no subscribers to selected repo and revset\n'
)
return
templater.show(
ctx,
changes=ctx.changeset(),
baseurl=ui.config(b'web', b'baseurl'),
root=repo.root,
webroot=n.root,
)
data = ui.popbuffer()
try:
msg = mail.parsebytes(data)
except emailerrors.MessageParseError as inst:
raise error.Abort(inst)
msg['In-reply-to'] = notify.messageid(ctx, domain, messageidseed)
msg['Message-Id'] = notify.messageid(
ctx, domain, messageidseed + b'-obsoleted'
)
msg['Date'] = encoding.strfromlocal(
dateutil.datestr(format=b"%a, %d %b %Y %H:%M:%S %1%2")
)
if not msg['From']:
sender = ui.config(b'email', b'from') or ui.username()
if b'@' not in sender or b'@localhost' in sender:
sender = n.fixmail(sender)
msg['From'] = mail.addressencode(ui, sender, n.charsets, n.test)
msg['To'] = ', '.join(sorted(subs))
msgtext = msg.as_bytes() if pycompat.ispy3 else msg.as_string()
if ui.configbool(b'notify', b'test'):
ui.write(msgtext)
if not msgtext.endswith(b'\n'):
ui.write(b'\n')
else:
ui.status(_(b'notify_obsoleted: sending mail for %d\n') % ctx.rev())
mail.sendmail(
ui, emailutils.parseaddr(msg['From'])[1], subs, msgtext, mbox=n.mbox
)
def hook(ui, repo, hooktype, node=None, **kwargs):
if hooktype != b"pretxnclose":
raise error.Abort(
_(b'Unsupported hook type %r') % pycompat.bytestr(hooktype)
)
for rev in obsutil.getobsoleted(repo, repo.currenttransaction()):
_report_commit(ui, repo, repo.unfiltered()[rev])

hgext/hooklib/changeset_published.py

  • This file was added.
# Copyright 2020 Joerg Sonnenberger <joerg@bec.de>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
"""changeset_published is a hook to send a mail when an
existing draft changeset is moved to the public phase.
Correct message threading requires the same messageidseed to be used for both
the original notification and the new mail.
Usage:
[notify]
messageidseed = myseed
[hooks]
txnclose-phase.changeset_published = \
python:hgext.hooklib.changeset_published.hook
"""
from __future__ import absolute_import
import email.errors as emailerrors
import email.utils as emailutils
from mercurial.i18n import _
from mercurial import (
encoding,
error,
logcmdutil,
mail,
pycompat,
registrar,
)
from mercurial.utils import dateutil
from .. import notify
configtable = {}
configitem = registrar.configitem(configtable)
configitem(
b'notify_published', b'domain', default=None,
)
configitem(
b'notify_published', b'messageidseed', default=None,
)
configitem(
b'notify_published',
b'template',
default=b'''Subject: changeset published
This changeset has been published.
''',
)
def _report_commit(ui, repo, ctx):
domain = ui.config(b'notify_published', b'domain') or ui.config(
b'notify', b'domain'
)
messageidseed = ui.config(
b'notify_published', b'messageidseed'
) or ui.config(b'notify', b'messageidseed')
template = ui.config(b'notify_published', b'template')
spec = logcmdutil.templatespec(template, None)
templater = logcmdutil.changesettemplater(ui, repo, spec)
ui.pushbuffer()
n = notify.notifier(ui, repo, b'incoming')
subs = set()
for sub, spec in n.subs:
if spec is None:
subs.add(sub)
continue
revs = repo.revs(b'%r and %d:', spec, ctx.rev())
if len(revs):
subs.add(sub)
continue
if len(subs) == 0:
ui.debug(
b'notify_published: no subscribers to selected repo and revset\n'
)
return
templater.show(
ctx,
changes=ctx.changeset(),
baseurl=ui.config(b'web', b'baseurl'),
root=repo.root,
webroot=n.root,
)
data = ui.popbuffer()
try:
msg = mail.parsebytes(data)
except emailerrors.MessageParseError as inst:
raise error.Abort(inst)
msg['In-reply-to'] = notify.messageid(ctx, domain, messageidseed)
msg['Message-Id'] = notify.messageid(
ctx, domain, messageidseed + b'-published'
)
msg['Date'] = encoding.strfromlocal(
dateutil.datestr(format=b"%a, %d %b %Y %H:%M:%S %1%2")
)
if not msg['From']:
sender = ui.config(b'email', b'from') or ui.username()
if b'@' not in sender or b'@localhost' in sender:
sender = n.fixmail(sender)
msg['From'] = mail.addressencode(ui, sender, n.charsets, n.test)
msg['To'] = ', '.join(sorted(subs))
msgtext = msg.as_bytes() if pycompat.ispy3 else msg.as_string()
if ui.configbool(b'notify', b'test'):
ui.write(msgtext)
if not msgtext.endswith(b'\n'):
ui.write(b'\n')
else:
ui.status(_(b'notify_published: sending mail for %d\n') % ctx.rev())
mail.sendmail(
ui, emailutils.parseaddr(msg['From'])[1], subs, msgtext, mbox=n.mbox
)
def hook(ui, repo, hooktype, node=None, **kwargs):
if hooktype != b"txnclose-phase":
raise error.Abort(
_(b'Unsupported hook type %r') % pycompat.bytestr(hooktype)
)
ctx = repo.unfiltered()[node]
if kwargs['oldphase'] == b'draft' and kwargs['phase'] == b'public':
_report_commit(ui, repo, ctx)

hgext/hooklib/enforce_draft_commits.py

  • This file was added.
# Copyright 2020 Joerg Sonnenberger <joerg@bec.de>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
"""enforce_draft_commits us a hook to ensure that all new changesets are
in the draft phase. This allows enforcing policies for work-in-progress
changes in overlay repositories, i.e. a shared hidden repositories with
different views for work-in-progress code and public history.
Usage:
[hooks]
pretxnclose-phase.enforce_draft_commits = \
python:hgext.hooklib.enforce_draft_commits.hook
"""
from __future__ import absolute_import
from mercurial.i18n import _
from mercurial import (
error,
pycompat,
)
def hook(ui, repo, hooktype, node=None, **kwargs):
if hooktype != b"pretxnclose-phase":
raise error.Abort(
_(b'Unsupported hook type %r') % pycompat.bytestr(hooktype)
)
ctx = repo.unfiltered()[node]
if kwargs['oldphase']:
raise error.Abort(
_(b'Phase change from %r to %r for %s rejected')
% (
pycompat.bytestr(kwargs['oldphase']),
pycompat.bytestr(kwargs['phase']),
ctx,
)
)
elif kwargs['phase'] != b'draft':
raise error.Abort(
_(b'New changeset %s in phase %r rejected')
% (ctx, pycompat.bytestr(kwargs['phase']))
)

hgext/hooklib/reject_merge_commits.py

  • This file was added.
# Copyright 2020 Joerg Sonnenberger <joerg@bec.de>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
"""reject_merge_commits is a hook to check new changesets for merge commits.
Merge commits are allowed only between different branches, i.e. merging
a feature branch into the main development branch. This can be used to
enforce policies for linear commit histories.
Usage:
[hooks]
pretxnchangegroup.reject_merge_commits = \
python:hgext.hooklib.reject_merge_commits.hook
"""
from __future__ import absolute_import
from mercurial.i18n import _
from mercurial import (
error,
pycompat,
)
def hook(ui, repo, hooktype, node=None, **kwargs):
if hooktype != b"pretxnchangegroup":
raise error.Abort(
_(b'Unsupported hook type %r') % pycompat.bytestr(hooktype)
)
ctx = repo.unfiltered()[node]
for rev in repo.changelog.revs(start=ctx.rev()):
rev = repo[rev]
parents = rev.parents()
if len(parents) < 2:
continue
if all(repo[p].branch() == rev.branch() for p in parents):
raise error.Abort(
_(
b'%s rejected as merge on the same branch. '
b'Please consider rebase.'
)
% rev
)

hgext/hooklib/reject_new_heads.py

  • This file was added.
# Copyright 2020 Joerg Sonnenberger <joerg@bec.de>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
"""reject_new_heads is a hook to check that branches touched by new changesets
have at most one open head. It can be used to enforce policies for
merge-before-push or rebase-before-push. It does not handle pre-existing
hydras.
Usage:
[hooks]
pretxnclose.reject_new_heads = \
python:hgext.hooklib.reject_new_heads.hook
"""
from __future__ import absolute_import
from mercurial.i18n import _
from mercurial import (
error,
pycompat,
)
def hook(ui, repo, hooktype, node=None, **kwargs):
if hooktype != b"pretxnclose":
raise error.Abort(
_(b'Unsupported hook type %r') % pycompat.bytestr(hooktype)
)
ctx = repo.unfiltered()[node]
branches = set()
for rev in repo.changelog.revs(start=ctx.rev()):
rev = repo[rev]
branches.add(rev.branch())
for branch in branches:
if len(repo.revs("head() and not closed() and branch(%s)", branch)) > 1:
raise error.Abort(
_(b'Changes on branch %r resulted in multiple heads')
% pycompat.bytestr(branch)
)

setup.py

'mercurial.revlogutils', 'mercurial.revlogutils',
'mercurial.testing', 'mercurial.testing',
'hgext', 'hgext',
'hgext.convert', 'hgext.convert',
'hgext.fsmonitor', 'hgext.fsmonitor',
'hgext.fastannotate', 'hgext.fastannotate',
'hgext.fsmonitor.pywatchman', 'hgext.fsmonitor.pywatchman',
'hgext.highlight', 'hgext.highlight',
'hgext.hooklib',
'hgext.infinitepush', 'hgext.infinitepush',
'hgext.largefiles', 'hgext.largefiles',
'hgext.lfs', 'hgext.lfs',
'hgext.narrow', 'hgext.narrow',
'hgext.remotefilelog', 'hgext.remotefilelog',
'hgext.zeroconf', 'hgext.zeroconf',
'hgext3rd', 'hgext3rd',
'hgdemandimport', 'hgdemandimport',

tests/test-hooklib-changeset_obsoleted.t

  • This file was added.
$ cat <<EOF >> $HGRCPATH
> [experimental]
> evolution = true
>
> [extensions]
> notify =
> hooklib =
>
> [phases]
> publish = False
>
> [notify]
> sources = pull
> diffstat = False
> messageidseed = example
> domain = example.com
>
> [reposubs]
> * = baz
> EOF
$ hg init a
$ hg --cwd a debugbuilddag +2
$ hg init b
$ cat <<EOF >> b/.hg/hgrc
> [hooks]
> incoming.notify = python:hgext.notify.hook
> pretxnclose.changeset_obsoleted = python:hgext.hooklib.changeset_obsoleted.hook
> EOF
$ hg --cwd b pull ../a | "$PYTHON" $TESTDIR/unwrap-message-id.py
pulling from ../a
requesting all changes
adding changesets
adding manifests
adding file changes
added 2 changesets with 0 changes to 0 files
new changesets 1ea73414a91b:66f7d451a68b (2 drafts)
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Date: * (glob)
Subject: changeset in * (glob)
From: debugbuilddag@example.com
X-Hg-Notification: changeset 1ea73414a91b
Message-Id: <hg.81c297828fd2d5afaadf2775a6a71b74143b6451dfaac09fac939e9107a50d01@example.com>
To: baz@example.com
changeset 1ea73414a91b in $TESTTMP/b
details: $TESTTMP/b?cmd=changeset;node=1ea73414a91b
description:
r0
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Date: * (glob)
Subject: changeset in * (glob)
From: debugbuilddag@example.com
X-Hg-Notification: changeset 66f7d451a68b
Message-Id: <hg.364d03da7dc13829eb779a805be7e37f54f572e9afcea7d2626856a794d3e8f3@example.com>
To: baz@example.com
changeset 66f7d451a68b in $TESTTMP/b
details: $TESTTMP/b?cmd=changeset;node=66f7d451a68b
description:
r1
(run 'hg update' to get a working copy)
$ hg --cwd a debugobsolete 1ea73414a91b0920940797d8fc6a11e447f8ea1e
1 new obsolescence markers
obsoleted 1 changesets
1 new orphan changesets
$ hg --cwd a push ../b --hidden | "$PYTHON" $TESTDIR/unwrap-message-id.py
1 new orphan changesets
pushing to ../b
searching for changes
no changes found
Subject: changeset abandoned
In-reply-to: <hg.81c297828fd2d5afaadf2775a6a71b74143b6451dfaac09fac939e9107a50d01@example.com>
Message-Id: <hg.d6329e9481594f0f3c8a84362b3511318bfbce50748ab1123f909eb6fbcab018@example.com>
Date: * (glob)
From: test@example.com
To: baz@example.com
This changeset has been abandoned.
1 new obsolescence markers
obsoleted 1 changesets

tests/test-hooklib-changeset_published.t

  • This file was added.
$ cat <<EOF >> $HGRCPATH
> [extensions]
> notify =
> hooklib =
>
> [phases]
> publish = False
>
> [notify]
> sources = pull
> diffstat = False
> messageidseed = example
> domain = example.com
>
> [reposubs]
> * = baz
> EOF
$ hg init a
$ hg --cwd a debugbuilddag .
$ hg init b
$ cat <<EOF >> b/.hg/hgrc
> [hooks]
> incoming.notify = python:hgext.notify.hook
> txnclose-phase.changeset_published = python:hgext.hooklib.changeset_published.hook
> EOF
$ hg --cwd b pull ../a | "$PYTHON" $TESTDIR/unwrap-message-id.py
pulling from ../a
requesting all changes
adding changesets
adding manifests
adding file changes
added 1 changesets with 0 changes to 0 files
new changesets 1ea73414a91b (1 drafts)
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Date: * (glob)
Subject: changeset in * (glob)
From: debugbuilddag@example.com
X-Hg-Notification: changeset 1ea73414a91b
Message-Id: <hg.81c297828fd2d5afaadf2775a6a71b74143b6451dfaac09fac939e9107a50d01@example.com>
To: baz@example.com
changeset 1ea73414a91b in $TESTTMP/b
details: $TESTTMP/b?cmd=changeset;node=1ea73414a91b
description:
r0
(run 'hg update' to get a working copy)
$ hg --cwd a phase --public 0
$ hg --cwd b pull ../a | "$PYTHON" $TESTDIR/unwrap-message-id.py
pulling from ../a
searching for changes
no changes found
1 local changesets published
Subject: changeset published
In-reply-to: <hg.81c297828fd2d5afaadf2775a6a71b74143b6451dfaac09fac939e9107a50d01@example.com>
Message-Id: <hg.2ec19bbddee5b542442bf5e1aed97bf706afff6aa765629883fbd1f4edd6fcb0@example.com>
Date: * (glob)
From: test@example.com
To: baz@example.com
This changeset has been published.
$ hg --cwd b phase --force --draft 0
$ cat <<EOF >> b/.hg/hgrc
> [notify_published]
> messageidseed = example2
> domain = alt.example.com
> template = Subject: changeset published
> From: hg@example.com\n
> This draft changeset has been published.\n
> EOF
$ hg --cwd b pull ../a | "$PYTHON" $TESTDIR/unwrap-message-id.py
pulling from ../a
searching for changes
no changes found
1 local changesets published
Subject: changeset published
From: hg@example.com
In-reply-to: <hg.e3381dc41c051215e50b1c166a72949d0fff99609eb373420bcb763af80ef230@alt.example.com>
Message-Id: <hg.c927f3d324e645a4245bfed20b0efb5b9582999d6be9bef45a37e7ec21208b24@alt.example.com>
Date: * (glob)
To: baz@example.com
This draft changeset has been published.

tests/test-hooklib-enforce_draft_commits.t

  • This file was added.
$ cat <<EOF >> $HGRCPATH
> [extensions]
> hooklib =
>
> [phases]
> publish = False
> EOF
$ hg init a
$ hg --cwd a debugbuilddag .
$ hg --cwd a phase --public 0
$ hg init b
$ cat <<EOF >> b/.hg/hgrc
> [hooks]
> pretxnclose-phase.enforce_draft_commits = \
> python:hgext.hooklib.enforce_draft_commits.hook
> EOF
$ hg --cwd b pull ../a
pulling from ../a
requesting all changes
adding changesets
adding manifests
adding file changes
error: pretxnclose-phase.enforce_draft_commits hook failed: New changeset 1ea73414a91b in phase 'public' rejected
transaction abort!
rollback completed
abort: New changeset 1ea73414a91b in phase 'public' rejected
[255]
$ hg --cwd a phase --force --draft 0
$ hg --cwd b pull ../a
pulling from ../a
requesting all changes
adding changesets
adding manifests
adding file changes
added 1 changesets with 0 changes to 0 files
new changesets 1ea73414a91b (1 drafts)
(run 'hg update' to get a working copy)
$ hg --cwd a phase --public 0
$ hg --cwd b pull ../a
pulling from ../a
searching for changes
no changes found
error: pretxnclose-phase.enforce_draft_commits hook failed: Phase change from 'draft' to 'public' for 1ea73414a91b rejected
abort: Phase change from 'draft' to 'public' for 1ea73414a91b rejected
[255]

tests/test-hooklib-reject_merge_commits.t

  • This file was added.
$ cat <<EOF >> $HGRCPATH
> [extensions]
> hooklib =
>
> [phases]
> publish = False
> EOF
$ hg init a
$ hg --cwd a debugbuilddag '.:parent.:childa*parent/childa<parent@otherbranch./childa'
$ hg --cwd a log -G
o changeset: 4:a9fb040caedd
|\ branch: otherbranch
| | tag: tip
| | parent: 3:af739dfc49b4
| | parent: 1:66f7d451a68b
| | user: debugbuilddag
| | date: Thu Jan 01 00:00:04 1970 +0000
| | summary: r4
| |
| o changeset: 3:af739dfc49b4
| | branch: otherbranch
| | parent: 0:1ea73414a91b
| | user: debugbuilddag
| | date: Thu Jan 01 00:00:03 1970 +0000
| | summary: r3
| |
+---o changeset: 2:a6b287721c3b
| |/ parent: 0:1ea73414a91b
| | parent: 1:66f7d451a68b
| | user: debugbuilddag
| | date: Thu Jan 01 00:00:02 1970 +0000
| | summary: r2
| |
o | changeset: 1:66f7d451a68b
|/ tag: childa
| user: debugbuilddag
| date: Thu Jan 01 00:00:01 1970 +0000
| summary: r1
|
o changeset: 0:1ea73414a91b
tag: parent
user: debugbuilddag
date: Thu Jan 01 00:00:00 1970 +0000
summary: r0
$ hg init b
$ cat <<EOF >> b/.hg/hgrc
> [hooks]
> pretxnchangegroup.reject_merge_commits = \
> python:hgext.hooklib.reject_merge_commits.hook
> EOF
$ hg --cwd b pull ../a -r a6b287721c3b
pulling from ../a
adding changesets
adding manifests
adding file changes
error: pretxnchangegroup.reject_merge_commits hook failed: a6b287721c3b rejected as merge on the same branch. Please consider rebase.
transaction abort!
rollback completed
abort: a6b287721c3b rejected as merge on the same branch. Please consider rebase.
[255]
$ hg --cwd b pull ../a -r 1ea73414a91b
pulling from ../a
adding changesets
adding manifests
adding file changes
added 1 changesets with 0 changes to 0 files
new changesets 1ea73414a91b (1 drafts)
(run 'hg update' to get a working copy)
$ hg --cwd b pull ../a -r a9fb040caedd
pulling from ../a
searching for changes
adding changesets
adding manifests
adding file changes
added 3 changesets with 0 changes to 0 files
new changesets 66f7d451a68b:a9fb040caedd (3 drafts)
(run 'hg update' to get a working copy)

tests/test-hooklib-reject_new_heads.t

  • This file was added.
$ cat <<EOF >> $HGRCPATH
> [extensions]
> hooklib =
>
> [phases]
> publish = False
> EOF
$ hg init a
$ hg --cwd a debugbuilddag '.:parent.*parent'
$ hg --cwd a log -G
o changeset: 2:fa942426a6fd
| tag: tip
| parent: 0:1ea73414a91b
| user: debugbuilddag
| date: Thu Jan 01 00:00:02 1970 +0000
| summary: r2
|
| o changeset: 1:66f7d451a68b
|/ user: debugbuilddag
| date: Thu Jan 01 00:00:01 1970 +0000
| summary: r1
|
o changeset: 0:1ea73414a91b
tag: parent
user: debugbuilddag
date: Thu Jan 01 00:00:00 1970 +0000
summary: r0
$ hg init b
$ cat <<EOF >> b/.hg/hgrc
> [hooks]
> pretxnclose.reject_new_heads = \
> python:hgext.hooklib.reject_new_heads.hook
> EOF
$ hg --cwd b pull ../a
pulling from ../a
requesting all changes
adding changesets
adding manifests
adding file changes
error: pretxnclose.reject_new_heads hook failed: Changes on branch 'default' resulted in multiple heads
transaction abort!
rollback completed
abort: Changes on branch 'default' resulted in multiple heads
[255]
$ hg --cwd b pull ../a -r 1ea73414a91b
pulling from ../a
adding changesets
adding manifests
adding file changes
added 1 changesets with 0 changes to 0 files
new changesets 1ea73414a91b (1 drafts)
(run 'hg update' to get a working copy)