# HG changeset patch
# User maf46@burn.cl.cam.ac.uk
# Node ID 57667c9b93a5a743e4629d15a0e6bd76699130c3
# Parent 4309b0a5a6010dd2e5811b77d2bc29a51acf290f
Fix zombie files in merge
Keir Fraser observed the following:
> I made a small test case that illustrates the bug in merging changesets
> with 'hg remove's in them:
>
> 1. Create a repository A containing files foo & bar.
> 2. Create clone called B.
> 3. A removes file bar, and commits this removal.
> 4. B edits file foo, and commits this edit.
>
> Now, if B:
> # hg pull ../A; hg update -m; hg commit
> Then bar remains deleted.
>
> If A:
> # hg pull ../B; hg update -m; hg commit
> Then bar is resurrected!
>
> It looks as though, when you merge across a branch, any deletions in
> your own branch are forgotten.
> ...
> Fixing this is a must, as zombie files are a real pain. :-)
Keir later patched our local copy of hg as shown below, which fixes
the problem. I've also enclosed a test which captures the test Keir
outlined...
Files deleted on a branch should not automatically reappear in a merge
Patch notes:
1. The first chunk does not change behaviour, but cleans up the code
to more closely match check of 'force' in the second chunk. I
think it makes the code clearer.
2. The second chunk fixes two bugs --
i. If we choose to keep a remotely-changed locally-deleted file,
then we need to 'get' that file. If we choose to delete it
then no action need be taken (it is already deleted in the
working manifest). Without this fix, choosing to delete would
get a Python traceback.
ii. The test for whether the file was remotely-created is
insufficient. It is only true if f is not in the common
ancestor. Otherwise the file was deleted locally, and should
remain deleted. (this is the most important fix!)
Index: hg/tests/test-merge6
===================================================================
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Add username/merge/editor to .hgrc
These take priority over the equivalent environment vars
Deprecate HGMERGE, HGUSER, and HGEDITOR in docs
Add ui section to docs
Remove undocumented HG_OPTS
Raise username code out of changelog class
Make tests ignore ~/.hgrc
manifest hash: d127ef02bc5266036b4c77a55319519e91bd475b
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCyM5NywK+sNU5EO8RAuU7AJ9zKk1TFrJXHM04jOOJJeBSp8jlTgCeP+tr
qwGFd+WaNqTepZ0wun5g9Uc=
=QEBu
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
A bunch of parsing/help updates
more explanation of how to get non-basic commands
shorten names of debug functions and add docstrings
add undo long docstring
promote anotate, export, and revert
make the global opts array global
refactor parsing
kill two unused arguments to fancyopts
update test-help
manifest hash: 459ae2273aaf54f71b4576677a681dc53ab2908c
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCyEDhywK+sNU5EO8RAr0DAJ9LTu8Fc2quLRtuwLPTQzWqlOJWKwCbBpZk
pnMkYnshsutVYljcil1P46I=
=Sleg
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Shorten help
Basic commands are now marked with "^" in the table
only basic commands are show in hg help by default
copyright banner not displayed with help by default
manifest hash: 7337678588354ca669a3b13d1fbb0ad893563947
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCx9LNywK+sNU5EO8RAokCAJ9ZK2sGEto9RLyukFX+n0nzHRoB9QCfU84I
e+Evc2g91kSS92FPrIC+Hnw=
=yUmt
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
options: kill -d for debug
You must now use --debug.
This is a preparatory step for changing the option parsing from:
hg [global-opts] <command> [command-opts]
to
hg [global-opts] <command> [global or command opts]
manifest hash: ec5ecc687340baa109a98b742f1cd8d1f4aadeed
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCx4urywK+sNU5EO8RAl90AJ0fwKU/GwHpJvlTFl5naeJbkceWewCgh4jK
kaTSSodACgPzgbFiid5RL4Y=
=siEe
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Actually warn on pulling from an unrelated repository
add some comments to findincoming
track the base nodes of the fetch set
report if the base set only contains nullid
add a test case
manifest hash: 3fc038a6041b0967a20503f5ec8876efc038841a
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCx0yLywK+sNU5EO8RAlK6AJ9J/GKPpYSMzTqmguXPWOISJ+zY5gCghd+j
ClLpn0dKZnB46dh0F8zhuuk=
=emNb
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Fix dodiff/changes
dodiff was failing to pass both nodes to changes
changes was comparing things backwards, resulting in added/deleted confusion
in dodiff
changes was deleting things from cached manifests, use copy()
changes now sorts output lists
manifest hash: 6ad972b0895b9d855e246efef49c2ebd943946b3
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCxZhZywK+sNU5EO8RApZWAJ9mYv1zc7IjXPIMwbVsfroQs8jfBACgn7R7
rTqLTTyIkM1OpL/ebnMtCOY=
=VcIr
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
[PATCH] Tests for clone command
From: Bryan O'Sullivan <bos@serpentine.com>
This patch adds two new tests. One tests clone's correct operation, the
other that clone fails when it should.
This patch depends on oserror.patch.
manifest hash: a0c9c4932baf29e7b33b2bd2003ddd2254ce65d3
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCxPX9ywK+sNU5EO8RAmSNAJ9esixtATdkszwLltu65dzpKf6P4gCcCqGT
+Z7CTbXRTFcJvmXFlvEo0xE=
=6fqz
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Add an empty line after description in verbose mode of show_changeset.
This makes it a little bit easier to see where a new changeset starts.
manifest hash: dd0dc3a135ca2fabed03b86116863c93a451670d
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFCxOVgW7P1GVgWeRoRAjtdAJ9B/ECd/Rhws2Bd+IcwRyqcYEN0ugCfbug2
Uhy2yUhV+mlrqFkoNBkcOmM=
=ttyp
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Refactor diffrevs/diffdir into changes
Add dirstate.changes to replace most of diffdir
Add localrepository.changes to replace diffrevs/diffdir
This code can now efficiently check for changes in single files, and
often without consulting the manifest. This should eventually make 'hg
diff Makefile' in a large project much faster.
This also fixes a bug where 'hg diff -r tip' failed to account for
files that had been added but not committed yet.
manifest hash: 20fde5d4b4cee49a76bcfe50f2dacf58b1f2258b
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCxMxpywK+sNU5EO8RAhzOAJ9VLQJoC+hiRYQtTSPbDhXBEJfQZwCgpDx9
GAwQ9jZHNsgXckBfXNCkJV8=
=hMuc
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
[PATCH] replace history with log
From: Radoslaw Szkodzinski <astralstorm@gorzow.mm.pl>
Merge hg history into hg log leaving the first as an alias
(tweaked by mpm: update docs, tests, use existing alias mechanism)
manifest hash: f85001a4373b8b823a6221ba5c506628059d6779
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCwSr1ywK+sNU5EO8RAmbAAJ9HlFkf8/qLOAQUtj8CWxGRPEyzVgCgtXwP
SdnAg1a2q4693+TBiOjrimY=
=w9rf
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Fix troubles with clone and exception handling
Clone deletes its directory on failure
This was deleting the lockfile out from under the lock object before
it got destroyed
This patch shuts lock up and makes the cleanup code for clone a little
cleaner.
manifest hash: f666fddcf6f3a905020a091f5e9fd2cb5d806cdd
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCwSGOywK+sNU5EO8RAkx2AKCIxPczl9YWnuUM+bMQnpVr8kv6uQCeNWld
SUxSB99PGJHhq1LWFaSJJNw=
=Frk/
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
[PATCH] umask for run-tests
From: Bryan O'Sullivan <bos@serpentine.com>
Another local factor affecting test breakage: umask. Mine is 002, but
some other people seem to use 022.
This patch makes run-tests explicitly set a umask.
manifest hash: 39aaea44746b113e992bcf0a9e93d9de3961c907
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCwOz8ywK+sNU5EO8RAvr7AJ9gRknqzQD6UmV1SpzHGqQyclcuxgCeItoD
W7yNbrtIAPyMJsfwhVECx3s=
=KhYb
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
[PATCH] Set locale before run-tests does anything
From: Bryan O'Sullivan <bos@serpentine.com>
On Mon, 2005-06-27 at 11:58 -0700, Bryan O'Sullivan wrote:
> I'm in en_us.UTF-8.
This discrepancy between my locale and Matt's broke several tests for
me, by the way. I think that the run-tests script should forcibly set
the locale to either "C" or "en_us.UTF-8" before it runs anything.
Since "C" is likely to be more portable, it's probably the right choice.
Here's the patch.
manifest hash: 1bf3291eafbde0cdeb1222bb72295c0e2974f236
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCwOyVywK+sNU5EO8RAk18AJ4pYmvvJMZfFIR39NqvquXhhWPUTQCcDhhg
Ce12AHU8o+m/aFToyRaaQL8=
=5nRs
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Add initial hook support
This adds the basic hook code as well as pre and post-commit hooks.
Argument passing is by environment variable key/value pairs so that
extra data can be passed over time. File lists will generally not be
passed to hooks as these can be extremely long (>1M).
manifest hash: 45cf9bab432782c391bc9c1c048c84cc75d52740
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCwOukywK+sNU5EO8RAsVsAJ9QipR2aKRSSvoRNo+3If6JddUDkwCgkZrM
KEmZpUOxhNHqezFVrHDRTjE=
=aedm
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Fix up test-merge1
The third test was wrong and showed up after fixing the typo in
compare code
manifest hash: 306ab9936457ad90e7c7a226d6407ca586262650
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCuPH8ywK+sNU5EO8RAql+AJsH/bxFMTMbUiaEoLNhXS6wNkzLAgCfQxDJ
3j5pZ3MzrtPD0y4wC3vIAA4=
=VKEy
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
[PATCH] Removal of a file added by merging branches
From: Michael A Fetterman <Michael.Fetterman@cl.cam.ac.uk>
Fixing a bug where removal of a file which is added when doing a merge
would cause python exception to be thrown.
manifest hash: 0bdd80e8abcee6aaa4804bb588f9939596fb3dd0
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCuNjLywK+sNU5EO8RAo44AJ4mvjsJcYOj/EmVDFrdkK3D3WHFSwCffxfV
cRl8mXc252/oRcYaMZbgrlg=
=S5x6
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
[PATCH] Don't prompt user for keep-vs-delete when the merge is about to be aborted
From: Michael A Fetterman <Michael.Fetterman@cl.cam.ac.uk>
If an "hg update" is attempted across branches, it is aborted with an
error message, basically saying that "update -m" is required. Don't
prompt the user for "keep vs delete" kinds of questions before deciding
to abort.
manifest hash: 250d3a0e866af2feac9344d499b26c254c33abc4
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCuNeDywK+sNU5EO8RAu2BAJwNWXXN2L+4TnOUGaOe9iBj6CgOYgCfRKIE
RF5czx2UQ2pSIP119xzq9mQ=
=Akj6
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
[PATCH]: Typo in localrepository.update
From: Goffredo Baroncelli <kreijack@libero.it>
I think that there is an error in the method update() of the class
localrepository. The variable 'm2n' was used instead of 'man'; so the
parent2 flags is computed instead of the ancestor flags.
manifest hash: dc11810367615245c4b39660d135ab3c0e71b85d
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCuNWpywK+sNU5EO8RAv9ZAKCxwbd242xZQ1CCrW2WDdQ4gclKtACghXYl
7j1YTRpINItwdXzak3cBS60=
=+Tf8
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
[PATCH] Merging identical changes from another branch
From: Michael A Fetterman <Michael.Fetterman@cl.cam.ac.uk>
The issue comes up when a local uncommitted *new* file (i.e. not in the
current manifest) is being merged with an identical file from a branch.
Since the file is not in the current manifest (it's either in the
current "to-be-added" list, or in the "unknown" state), there's no
(local) node from which to create a mergepoint.
manifest hash: 4e64ce654a6473524789a97bbaf8bff61b4343af
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCuM/1ywK+sNU5EO8RAn7xAKCJoH/CgzVK4h4xPJDrd2lY9XOINACgmt01
92uuMswZXcoCchQAaxew7C0=
=qRsk
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
[PATCH] Unintuive use
From: Vincent Danjean <vdanjean.ml@free.fr>
> 1) "hg pull" does not tell what and if it does something.
> 2) hg does not tell if the data have been merge or not, if we have to
> call resolve, ...
Here is a patch for 1) and 2). It also add an option to 'pull' to call
update if there is something new.
Tweaked by mpm:
- change message style
- use ui.status
- change option from resolve to update
manifest hash: a61a7af00ef2f33867f6c88f95dc8258cb1a3e51
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCuMcpywK+sNU5EO8RAucjAKC06gwAico7Wz5YOcp5u9NVxEhItwCfcIzA
VmiifZTj5ZUjTTcPhmiKh2M=
=rbLT
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
[PATCH] (3/4) Removing an added file
From: Michael A Fetterman <Michael.Fetterman@cl.cam.ac.uk>
Here's an hg changeset that deals with "hg add"ing and then "hg remove"ing
a file without ever committing it. Previously, this caused a python
exception
at commit time.
Test case included in the patch.
manifest hash: 4d012f3f8b1aa664fed9e8bc2849931b28728acd
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCuMPtywK+sNU5EO8RAuxMAKCmeMLqOUBmUfCfbQgYOAysOvKkwgCfcJai
2My0M7KqhwnJSbZXbGIOcow=
=N7N0
-----END PGP SIGNATURE-----
