This is an archive of the discontinued Mercurial Phabricator instance.

Differential D529

uncommit: move fb-extension to core which uncommits a changeset
ClosedPublic

Authored by pulkit on Aug 27 2017, 4:45 PM.
Tags
None
Subscribers
pulkit, durham, quark and 3 others

Details

Reviewers
quark
durham
durin42
Group Reviewers
hg-reviewers
Commits
rHGda2f5f19312c: uncommit: move fb-extension to core which uncommits a changeset
rHG279ec1b714a4: uncommit: move fb-extension to core which uncommits a changeset
rHGd01819c8f3c4: uncommit: move fb-extension to core which uncommits a changeset
Summary

uncommit extension in fb-hgext adds a uncommit command which by default
uncommits a changeset and move all the changes to the working directory. If
file names are passed, uncommit moves the changes from those files to the
working directory and left the changeset with remaining committed files.

The uncommit extension in fb-hgext does not creates an empty commit like the one
in evolve extension unless user has specified ui.alllowemptycommit to True.

The test file added is a combination of tests from test-uncommit.t,
test-uncommit-merge.t and test-uncommit-bookmark.t from fb-hgext.

.. feature::

A new uncommit extension which provides `hg uncommit` using which one can
uncommit part or all of the changeset. This command undoes the effect of a
local commit, returning the affected files to their uncommitted state.

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 created this revision.Aug 27 2017, 4:45 PM
pulkit added a comment.Sep 4 2017, 8:32 AM

This has been in from a week, any updates on this?

yuja added a subscriber: yuja.Sep 5 2017, 10:16 AM

Overall, the patch looks good, but I don't have expertise in this area.

The uncommit extension in fb-hgext does not creates an empty commit like the one
in evolve extension unless user has specified ui.alllowemptycommit to True.

Maybe we'll need a command option to keep the empty commit? ui.allowemptycommit
is an internal flag,

hgext/uncommit.py
50

Can't we make this share some bits with cmdutil.amend?

154

Nit: wctx.parents() should return at least one ctx even if it is null
revision.

160

Nit: amend uses old.mutable().

169

Perhaps this should exit with 1 as "hg commit" would do.

174

I'm not sure if this retractboundary() is valid. cmdutil.amend()
appears to use phases.new-commit config instead.

177

Can't we use scmutil.cleanupnodes()?

martinvonz added a subscriber: martinvonz.Sep 6 2017, 7:15 PM
martinvonz added inline comments.
hgext/uncommit.py
17

How about added files? Also, I think we use "removed" for tracked changes ("deleted" for untracked removal of file).

100

I prefer the form that doesn't depend on the order in the status tuple:

s = repo.status(...)
for f in s.added:
    ...
149

This is inconsistent with the amend extension that Jun added. We can always add support for that later.

tests/test-uncommit.t
23

s/some/part/ (I think)

quark requested changes to this revision.Sep 8 2017, 3:04 PM
quark added a subscriber: quark.

Please address other reviewers' comments.

hgext/uncommit.py
13

I'd add "(EXPERIMENTAL)". The command does not write unamend_source so it's in theory problematic with the current obsstore design.

177

Yep. That will also make _updatebookmarks unnecessary.

This revision now requires changes to proceed.Sep 8 2017, 3:04 PM
pulkit added a comment.Sep 9 2017, 1:23 PM
In D529#10887, @quark wrote:

Please address other reviewers' comments.

Yeah will do that soon, was busy for past few days.

pulkit edited edge metadata.Sep 11 2017, 10:24 AM
pulkit updated this revision to Diff 1712.
durham accepted this revision.Sep 11 2017, 12:23 PM
durham added a subscriber: durham.

Looks good, except for my ui.allowemptycommit comment.

hgext/uncommit.py
60

If I read this right, it means a user having ui.allowemptycommit set to True will cause hg uncommit to always leave a commit behind? I don't think that's what we want. ui.allowemptycommit is to let a user make an empt commit if they want. It's not really meant to say 'leave empty commits behind for me'. I don't think it should affect the uncommit command.

I see this came from our original extension. I'd still drop it.

62

Might be worth a comment that returning p1 here is special cased to not create an obsmarker later.

155

Should "amend" be "uncommit" here?

160

Same here, "amend" -> "uncommit"

pulkit edited the summary of this revision. (Show Details)Sep 11 2017, 4:51 PM
pulkit updated this revision to Diff 1729.
pulkit marked 15 inline comments as done.Sep 11 2017, 4:53 PM
pulkit added inline comments.
hgext/uncommit.py
174

The phases.new-commit config update the phase to old.phase() if --secret flag is not passed. Since we don't have that I think we can use retractboundary to directly update to old.phase()

quark accepted this revision.Sep 11 2017, 6:21 PM

Generally look good to me. I think core hg does not have very convenient commit editing APIs yet (memctx feels too low-level). I'm counting on in-memory-memory work to provide convenient APIs eventually. So I'd hold on refactoring until we have better APIs.

hgext/uncommit.py
174

I think it's still safer to follow common practice and avoid using low-level retractboundary. It seems to be a few lines change around line 85. I guess retractboundary may cause surprises cases when the uncommitted node happens to exist.

pulkit updated this revision to Diff 1823.Sep 14 2017, 2:28 PM
pulkit marked an inline comment as done.Sep 14 2017, 2:29 PM
durin42 accepted this revision.Sep 15 2017, 11:40 AM
This revision is now accepted and ready to land.Sep 15 2017, 11:40 AM
Closed by commit rHGd01819c8f3c4: uncommit: move fb-extension to core which uncommits a changeset (authored by pulkit). · Explain WhySep 15 2017, 12:03 PM
This revision was automatically updated to reflect the committed changes.
durin42 added a subscriber: pulkit.Sep 15 2017, 5:10 PM

We've had some discussion about this on irc. The evolve behavior of preserving empty commits is important in one circumstance: when the commit will briefly be empty, but should retain its identity for the purposes of evolution. Personally, I've never used uncommit and ended up with an empty commit (via the evolve version, which I currently use), so I've got a pretty strong sense that this is an edge case. Here's where we're landing for the time being:

We'll add --keep to uncommit, preserve the current default, but refuse to do no-arguments uncommit if the working directory is dirty. This prevents users from accidentally blowing their foot off in the cases that are the most likely to cause pain. This still leaves some difficulty at the margins (how do you uncommit everything if you've got a dirty working directory?), but we can probably overcome that in time. One thing we hope will help there is a generalized hg undo command, which it sounds like Facebook has at least partially drafted.

Note that this is not necessarily a final state for uncommit: it's possible that after further experimentation we'll flip the default, or otherwise adjust behavior. It's a complicated balancing act between what's intuitive (which we don't all agree on) and preserving the richest metadata possible.

Thanks,
Augie

Revision Contents
Changeset List

PathPackages
AMhgext/uncommit.py (183 lines)
AMtests/test-uncommit.t (366 lines)

Diff 1848

hgext/uncommit.py

  • This file was added.
# uncommit - undo the actions of a commit
#
# Copyright 2011 Peter Arrenbrecht <peter.arrenbrecht@gmail.com>
# Logilab SA <contact@logilab.fr>
# Pierre-Yves David <pierre-yves.david@ens-lyon.org>
# Patrick Mezard <patrick@mezard.eu>
# Copyright 2016 Facebook, Inc.
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
"""uncommit part or all of a local changeset (EXPERIMENTAL)
quarkUnsubmitted
Done

I'd add "(EXPERIMENTAL)". The command does not write unamend_source so it's in theory problematic with the current obsstore design.

quark: I'd add "(EXPERIMENTAL)". The command does not write `unamend_source` so it's in theory…
This command undoes the effect of a local commit, returning the affected
files to their uncommitted state. This means that files modified, added or
removed in the changeset will be left unchanged, and so will remain modified,
added and removed in the working directory.
martinvonzUnsubmitted
Done

How about added files? Also, I think we use "removed" for tracked changes ("deleted" for untracked removal of file).

martinvonz: How about added files? Also, I think we use "removed" for tracked changes ("deleted" for…
"""
from __future__ import absolute_import
from mercurial.i18n import _
from mercurial import (
commands,
context,
copies,
error,
node,
obsolete,
registrar,
scmutil,
)
cmdtable = {}
command = registrar.command(cmdtable)
# Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for
# extensions which SHIP WITH MERCURIAL. Non-mainline extensions should
# be specifying the version(s) of Mercurial they are tested with, or
# leave the attribute unspecified.
testedwith = 'ships-with-hg-core'
def _commitfiltered(repo, ctx, match, allowempty):
"""Recommit ctx with changed files not in match. Return the new
node identifier, or None if nothing changed.
"""
base = ctx.p1()
# ctx
initialfiles = set(ctx.files())
yujaUnsubmitted
Done

Can't we make this share some bits with cmdutil.amend?

yuja: Can't we make this share some bits with cmdutil.amend?
exclude = set(f for f in initialfiles if match(f))
# No files matched commit, so nothing excluded
if not exclude:
return None
files = (initialfiles - exclude)
# return the p1 so that we don't create an obsmarker later
if not files and not allowempty:
return ctx.parents()[0].node()
durhamUnsubmitted
Done

If I read this right, it means a user having ui.allowemptycommit set to True will cause hg uncommit to always leave a commit behind? I don't think that's what we want. ui.allowemptycommit is to let a user make an empt commit if they want. It's not really meant to say 'leave empty commits behind for me'. I don't think it should affect the uncommit command.

I see this came from our original extension. I'd still drop it.

durham: If I read this right, it means a user having ui.allowemptycommit set to True will cause hg…
# Filter copies
durhamUnsubmitted
Done

Might be worth a comment that returning p1 here is special cased to not create an obsmarker later.

durham: Might be worth a comment that returning p1 here is special cased to not create an obsmarker…
copied = copies.pathcopies(base, ctx)
copied = dict((dst, src) for dst, src in copied.iteritems()
if dst in files)
def filectxfn(repo, memctx, path, contentctx=ctx, redirect=()):
if path not in contentctx:
return None
fctx = contentctx[path]
mctx = context.memfilectx(repo, fctx.path(), fctx.data(),
fctx.islink(),
fctx.isexec(),
copied=copied.get(path))
return mctx
new = context.memctx(repo,
parents=[base.node(), node.nullid],
text=ctx.description(),
files=files,
filectxfn=filectxfn,
user=ctx.user(),
date=ctx.date(),
extra=ctx.extra())
# phase handling
commitphase = ctx.phase()
overrides = {('phases', 'new-commit'): commitphase}
with repo.ui.configoverride(overrides, 'uncommit'):
newid = repo.commitctx(new)
return newid
def _uncommitdirstate(repo, oldctx, match):
"""Fix the dirstate after switching the working directory from
oldctx to a copy of oldctx not containing changed files matched by
match.
"""
ctx = repo['.']
ds = repo.dirstate
copies = dict(ds.copies())
s = repo.status(oldctx.p1(), oldctx, match=match)
for f in s.modified:
martinvonzUnsubmitted
Done

I prefer the form that doesn't depend on the order in the status tuple:

s = repo.status(...)
for f in s.added:
    ...
martinvonz: I prefer the form that doesn't depend on the order in the status tuple: s = repo.status(...)…
if ds[f] == 'r':
# modified + removed -> removed
continue
ds.normallookup(f)
for f in s.added:
if ds[f] == 'r':
# added + removed -> unknown
ds.drop(f)
elif ds[f] != 'a':
ds.add(f)
for f in s.removed:
if ds[f] == 'a':
# removed + added -> normal
ds.normallookup(f)
elif ds[f] != 'r':
ds.remove(f)
# Merge old parent and old working dir copies
oldcopies = {}
for f in (s.modified + s.added):
src = oldctx[f].renamed()
if src:
oldcopies[f] = src[0]
oldcopies.update(copies)
copies = dict((dst, oldcopies.get(src, src))
for dst, src in oldcopies.iteritems())
# Adjust the dirstate copies
for dst, src in copies.iteritems():
if (src not in ctx or dst in ctx or ds[dst] != 'a'):
src = None
ds.copy(src, dst)
@command('uncommit',
[('', 'empty', False, _('allow an empty commit after uncommiting')),
] + commands.walkopts,
_('[OPTION]... [FILE]...'))
def uncommit(ui, repo, *pats, **opts):
"""uncommit part or all of a local changeset
This command undoes the effect of a local commit, returning the affected
files to their uncommitted state. This means that files modified or
deleted in the changeset will be left unchanged, and so will remain
modified in the working directory.
"""
with repo.wlock(), repo.lock():
wctx = repo[None]
martinvonzUnsubmitted
Done

This is inconsistent with the amend extension that Jun added. We can always add support for that later.

martinvonz: This is inconsistent with the amend extension that Jun added. We can always add support for…
if wctx.parents()[0].node() == node.nullid:
raise error.Abort(_("cannot uncommit null changeset"))
if len(wctx.parents()) > 1:
raise error.Abort(_("cannot uncommit while merging"))
yujaUnsubmitted
Done

Nit: wctx.parents() should return at least one ctx even if it is null
revision.

yuja: Nit: wctx.parents() should return at least one ctx even if it is null revision.
old = repo['.']
durhamUnsubmitted
Done

Should "amend" be "uncommit" here?

durham: Should "amend" be "uncommit" here?
if not old.mutable():
raise error.Abort(_('cannot uncommit public changesets'))
if len(old.parents()) > 1:
raise error.Abort(_("cannot uncommit merge changeset"))
allowunstable = obsolete.isenabled(repo, obsolete.allowunstableopt)
yujaUnsubmitted
Done

Nit: amend uses old.mutable().

yuja: Nit: amend uses old.mutable().
durhamUnsubmitted
Done

Same here, "amend" -> "uncommit"

durham: Same here, "amend" -> "uncommit"
if not allowunstable and old.children():
raise error.Abort(_('cannot uncommit changeset with children'))
with repo.transaction('uncommit'):
match = scmutil.match(old, pats, opts)
newid = _commitfiltered(repo, old, match, opts.get('empty'))
if newid is None:
ui.status(_("nothing to uncommit\n"))
return 1
yujaUnsubmitted
Done

Perhaps this should exit with 1 as "hg commit" would do.

yuja: Perhaps this should exit with 1 as "hg commit" would do.
mapping = {}
if newid != old.p1().node():
# Move local changes on filtered changeset
mapping[old.node()] = (newid,)
yujaUnsubmitted
Not Done

I'm not sure if this retractboundary() is valid. cmdutil.amend()
appears to use phases.new-commit config instead.

yuja: I'm not sure if this `retractboundary()` is valid. cmdutil.amend() appears to use phases.new…
pulkitAuthorUnsubmitted
Not Done

The phases.new-commit config update the phase to old.phase() if --secret flag is not passed. Since we don't have that I think we can use retractboundary to directly update to old.phase()

pulkit: The phases.new-commit config update the phase to old.phase() if `--secret` flag is not passed.
quarkUnsubmitted
Done

I think it's still safer to follow common practice and avoid using low-level retractboundary. It seems to be a few lines change around line 85. I guess retractboundary may cause surprises cases when the uncommitted node happens to exist.

quark: I think it's still safer to follow common practice and avoid using low-level `retractboundary`.
else:
# Fully removed the old commit
mapping[old.node()] = ()
yujaUnsubmitted
Done

Can't we use scmutil.cleanupnodes()?

yuja: Can't we use scmutil.cleanupnodes()?
quarkUnsubmitted
Done

Yep. That will also make _updatebookmarks unnecessary.

quark: Yep. That will also make `_updatebookmarks` unnecessary.
scmutil.cleanupnodes(repo, mapping, 'uncommit')
with repo.dirstate.parentchange():
repo.dirstate.setparents(newid, node.nullid)
_uncommitdirstate(repo, old, match)

tests/test-uncommit.t

  • This file was added.
Test uncommit - set up the config
$ cat >> $HGRCPATH <<EOF
> [experimental]
> evolution=createmarkers, allowunstable
> [extensions]
> uncommit =
> drawdag=$TESTDIR/drawdag.py
> EOF
Build up a repo
$ hg init repo
$ cd repo
$ hg bookmark foo
Help for uncommit
$ hg help uncommit
hg uncommit [OPTION]... [FILE]...
uncommit part or all of a local changeset
martinvonzUnsubmitted
Done

s/some/part/ (I think)

martinvonz: s/some/part/ (I think)
This command undoes the effect of a local commit, returning the affected
files to their uncommitted state. This means that files modified or
deleted in the changeset will be left unchanged, and so will remain
modified in the working directory.
(use 'hg help -e uncommit' to show help for the uncommit extension)
options ([+] can be repeated):
--empty allow an empty commit after uncommiting
-I --include PATTERN [+] include names matching the given patterns
-X --exclude PATTERN [+] exclude names matching the given patterns
(some details hidden, use --verbose to show complete help)
Uncommit with no commits should fail
$ hg uncommit
abort: cannot uncommit null changeset
[255]
Create some commits
$ touch files
$ hg add files
$ for i in a ab abc abcd abcde; do echo $i > files; echo $i > file-$i; hg add file-$i; hg commit -m "added file-$i"; done
$ ls
file-a
file-ab
file-abc
file-abcd
file-abcde
files
$ hg log -G -T '{rev}:{node} {desc}' --hidden
@ 4:6c4fd43ed714e7fcd8adbaa7b16c953c2e985b60 added file-abcde
|
o 3:6db330d65db434145c0b59d291853e9a84719b24 added file-abcd
|
o 2:abf2df566fc193b3ac34d946e63c1583e4d4732b added file-abc
|
o 1:69a232e754b08d568c4899475faf2eb44b857802 added file-ab
|
o 0:3004d2d9b50883c1538fc754a3aeb55f1b4084f6 added file-a
Simple uncommit off the top, also moves bookmark
$ hg bookmark
* foo 4:6c4fd43ed714
$ hg uncommit
$ hg status
M files
A file-abcde
$ hg bookmark
* foo 3:6db330d65db4
$ hg log -G -T '{rev}:{node} {desc}' --hidden
x 4:6c4fd43ed714e7fcd8adbaa7b16c953c2e985b60 added file-abcde
|
@ 3:6db330d65db434145c0b59d291853e9a84719b24 added file-abcd
|
o 2:abf2df566fc193b3ac34d946e63c1583e4d4732b added file-abc
|
o 1:69a232e754b08d568c4899475faf2eb44b857802 added file-ab
|
o 0:3004d2d9b50883c1538fc754a3aeb55f1b4084f6 added file-a
Recommit
$ hg commit -m 'new change abcde'
$ hg status
$ hg heads -T '{rev}:{node} {desc}'
5:0c07a3ccda771b25f1cb1edbd02e683723344ef1 new change abcde (no-eol)
Uncommit of non-existent and unchanged files has no effect
$ hg uncommit nothinghere
nothing to uncommit
[1]
$ hg status
$ hg uncommit file-abc
nothing to uncommit
[1]
$ hg status
Try partial uncommit, also moves bookmark
$ hg bookmark
* foo 5:0c07a3ccda77
$ hg uncommit files
$ hg status
M files
$ hg bookmark
* foo 6:3727deee06f7
$ hg heads -T '{rev}:{node} {desc}'
6:3727deee06f72f5ffa8db792ee299cf39e3e190b new change abcde (no-eol)
$ hg log -r . -p -T '{rev}:{node} {desc}'
6:3727deee06f72f5ffa8db792ee299cf39e3e190b new change abcdediff -r 6db330d65db4 -r 3727deee06f7 file-abcde
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/file-abcde Thu Jan 01 00:00:00 1970 +0000
@@ -0,0 +1,1 @@
+abcde
$ hg log -G -T '{rev}:{node} {desc}' --hidden
@ 6:3727deee06f72f5ffa8db792ee299cf39e3e190b new change abcde
|
| x 5:0c07a3ccda771b25f1cb1edbd02e683723344ef1 new change abcde
|/
| x 4:6c4fd43ed714e7fcd8adbaa7b16c953c2e985b60 added file-abcde
|/
o 3:6db330d65db434145c0b59d291853e9a84719b24 added file-abcd
|
o 2:abf2df566fc193b3ac34d946e63c1583e4d4732b added file-abc
|
o 1:69a232e754b08d568c4899475faf2eb44b857802 added file-ab
|
o 0:3004d2d9b50883c1538fc754a3aeb55f1b4084f6 added file-a
$ hg commit -m 'update files for abcde'
Uncommit with dirty state
$ echo "foo" >> files
$ cat files
abcde
foo
$ hg status
M files
$ hg uncommit files
$ cat files
abcde
foo
$ hg commit -m "files abcde + foo"
Uncommit in the middle of a stack, does not move bookmark
$ hg checkout '.^^^'
1 files updated, 0 files merged, 2 files removed, 0 files unresolved
(leaving bookmark foo)
$ hg log -r . -p -T '{rev}:{node} {desc}'
2:abf2df566fc193b3ac34d946e63c1583e4d4732b added file-abcdiff -r 69a232e754b0 -r abf2df566fc1 file-abc
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/file-abc Thu Jan 01 00:00:00 1970 +0000
@@ -0,0 +1,1 @@
+abc
diff -r 69a232e754b0 -r abf2df566fc1 files
--- a/files Thu Jan 01 00:00:00 1970 +0000
+++ b/files Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +1,1 @@
-ab
+abc
$ hg bookmark
foo 8:83815831694b
$ hg uncommit
$ hg status
M files
A file-abc
$ hg heads -T '{rev}:{node} {desc}'
8:83815831694b1271e9f207cb1b79b2b19275edcb files abcde + foo (no-eol)
$ hg bookmark
foo 8:83815831694b
$ hg commit -m 'new abc'
created new head
Partial uncommit in the middle, does not move bookmark
$ hg checkout '.^'
1 files updated, 0 files merged, 1 files removed, 0 files unresolved
$ hg log -r . -p -T '{rev}:{node} {desc}'
1:69a232e754b08d568c4899475faf2eb44b857802 added file-abdiff -r 3004d2d9b508 -r 69a232e754b0 file-ab
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/file-ab Thu Jan 01 00:00:00 1970 +0000
@@ -0,0 +1,1 @@
+ab
diff -r 3004d2d9b508 -r 69a232e754b0 files
--- a/files Thu Jan 01 00:00:00 1970 +0000
+++ b/files Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +1,1 @@
-a
+ab
$ hg bookmark
foo 8:83815831694b
$ hg uncommit file-ab
$ hg status
A file-ab
$ hg heads -T '{rev}:{node} {desc}\n'
10:8eb87968f2edb7f27f27fe676316e179de65fff6 added file-ab
9:5dc89ca4486f8a88716c5797fa9f498d13d7c2e1 new abc
8:83815831694b1271e9f207cb1b79b2b19275edcb files abcde + foo
$ hg bookmark
foo 8:83815831694b
$ hg commit -m 'update ab'
$ hg status
$ hg heads -T '{rev}:{node} {desc}\n'
11:f21039c59242b085491bb58f591afc4ed1c04c09 update ab
9:5dc89ca4486f8a88716c5797fa9f498d13d7c2e1 new abc
8:83815831694b1271e9f207cb1b79b2b19275edcb files abcde + foo
$ hg log -G -T '{rev}:{node} {desc}' --hidden
@ 11:f21039c59242b085491bb58f591afc4ed1c04c09 update ab
|
o 10:8eb87968f2edb7f27f27fe676316e179de65fff6 added file-ab
|
| o 9:5dc89ca4486f8a88716c5797fa9f498d13d7c2e1 new abc
| |
| | o 8:83815831694b1271e9f207cb1b79b2b19275edcb files abcde + foo
| | |
| | | x 7:0977fa602c2fd7d8427ed4e7ee15ea13b84c9173 update files for abcde
| | |/
| | o 6:3727deee06f72f5ffa8db792ee299cf39e3e190b new change abcde
| | |
| | | x 5:0c07a3ccda771b25f1cb1edbd02e683723344ef1 new change abcde
| | |/
| | | x 4:6c4fd43ed714e7fcd8adbaa7b16c953c2e985b60 added file-abcde
| | |/
| | o 3:6db330d65db434145c0b59d291853e9a84719b24 added file-abcd
| | |
| | x 2:abf2df566fc193b3ac34d946e63c1583e4d4732b added file-abc
| |/
| x 1:69a232e754b08d568c4899475faf2eb44b857802 added file-ab
|/
o 0:3004d2d9b50883c1538fc754a3aeb55f1b4084f6 added file-a
Uncommit with draft parent
$ hg uncommit
$ hg phase -r .
10: draft
$ hg commit -m 'update ab again'
Uncommit with public parent
$ hg phase -p "::.^"
$ hg uncommit
$ hg phase -r .
10: public
Partial uncommit with public parent
$ echo xyz > xyz
$ hg add xyz
$ hg commit -m "update ab and add xyz"
$ hg uncommit xyz
$ hg status
A xyz
$ hg phase -r .
14: draft
$ hg phase -r ".^"
10: public
Uncommit leaving an empty changeset
$ cd $TESTTMP
$ hg init repo1
$ cd repo1
$ hg debugdrawdag <<'EOS'
> Q
> |
> P
> EOS
$ hg up Q -q
$ hg uncommit --empty
$ hg log -G -T '{desc} FILES: {files}'
@ Q FILES:
|
| x Q FILES: Q
|/
o P FILES: P
$ hg status
A Q
$ cd ..
$ rm repo1 -rf
Testing uncommit while merge
$ hg init repo2
$ cd repo2
Create some history
$ touch a
$ hg add a
$ for i in 1 2 3; do echo $i > a; hg commit -m "a $i"; done
$ hg checkout 0
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ touch b
$ hg add b
$ for i in 1 2 3; do echo $i > b; hg commit -m "b $i"; done
created new head
$ hg log -G -T '{rev}:{node} {desc}' --hidden
@ 5:2cd56cdde163ded2fbb16ba2f918c96046ab0bf2 b 3
|
o 4:c3a0d5bb3b15834ffd2ef9ef603e93ec65cf2037 b 2
|
o 3:49bb009ca26078726b8870f1edb29fae8f7618f5 b 1
|
| o 2:990982b7384266e691f1bc08ca36177adcd1c8a9 a 3
| |
| o 1:24d38e3cf160c7b6f5ffe82179332229886a6d34 a 2
|/
o 0:ea4e33293d4d274a2ba73150733c2612231f398c a 1
Add and expect uncommit to fail on both merge working dir and merge changeset
$ hg merge 2
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
(branch merge, don't forget to commit)
$ hg uncommit
abort: cannot uncommit while merging
[255]
$ hg status
M a
$ hg commit -m 'merge a and b'
$ hg uncommit
abort: cannot uncommit merge changeset
[255]
$ hg status
$ hg log -G -T '{rev}:{node} {desc}' --hidden
@ 6:c03b9c37bc67bf504d4912061cfb527b47a63c6e merge a and b
|\
| o 5:2cd56cdde163ded2fbb16ba2f918c96046ab0bf2 b 3
| |
| o 4:c3a0d5bb3b15834ffd2ef9ef603e93ec65cf2037 b 2
| |
| o 3:49bb009ca26078726b8870f1edb29fae8f7618f5 b 1
| |
o | 2:990982b7384266e691f1bc08ca36177adcd1c8a9 a 3
| |
o | 1:24d38e3cf160c7b6f5ffe82179332229886a6d34 a 2
|/
o 0:ea4e33293d4d274a2ba73150733c2612231f398c a 1