Summary:
Computing cache key can be quite slow (100ms and even more). Instead let's use
a simpler cache key: size + mtime of phaseroots, obsstore and changelod and
whether changelog write was delayed.
Test Plan:
1) arc unit
2) Run test-newhiddencache.t
3) Run upstream unittest with extension enabled.
python run-tests.py --extra-config-opt extensions.newhiddencachekey=/home/stash/facebook-hg-rpms/fb-hgext/hgext3rd/newhiddencachekey.py -j20
Verify that there are no failures except for those that show that new extension were added.
Note: also need to comment out debug output in the extension
Reviewers: #mercurial, durham
Reviewed By: durham
Subscribers: durham, quark, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4604811
Signature: t1:4604811:1487958882:702ea8fdee16def674db97344832936d53f0cd9d
Summary:
Upstream mercurial doesn't run hooks for temporary amended commits
(see https://bz.mercurial-scm.org/show_bug.cgi?id=3501 and
c2ca20984e3c9f30d73f0f35c35904b64edbd692). No other hook contains info about
temporary amended commits.
Let's wrap localrepository.commit() function and add missing commits inside
this wrapper.
Test Plan: arc unit
Reviewers: #mercurial, durham
Reviewed By: durham
Subscribers: durham, mjpieters, #sourcecontrol
Differential Revision: https://phabricator.intern.facebook.com/D4634950
Tasks: 16387175
Signature: t1:4634950:1488399457:1d299c6ad3cfd418dda32a7ec20c60290f8d7778
Summary:
It makes it easier to set up a hook - just enabling the extension will enable
the hook.
Test Plan:
Run perl script
system("echo 1 >> 1");
system("hg add 1");
system("hg ci -m"."\x80");
system("hg ci -m"."\x01");
system("hg ci -m ok");
make sure only one commit was created
Reviewers: #sourcecontrol, mjpieters
Reviewed By: mjpieters
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4604980
Tasks: 16212973
Signature: t1:4604980:1487944300:8e4e7f340c746237b47773fc81759025e97b919e
Summary:
bundlerepo adds new commits that are not present in fastpartialmatch index.
Without this diff any access to the commit from bundle fails if
fastpartialmatch.raiseifinconsistent config option is set.
Let's not use fastpartialmatch in bundlerepos
Test Plan: arc unit
Reviewers: #mercurial, mjpieters
Reviewed By: mjpieters
Subscribers: mjpieters, #sourcecontrol
Differential Revision: https://phabricator.intern.facebook.com/D4635368
Tasks: 16394593
Signature: t1:4635368:1488393506:15b8f6a9fddabe681de7fd1296bbfafb9512f558
Summary:
Turned out that wrapping `changectx.__init__()` doesn't give any performance
benefits but instead it breaks initializing of changectx. For example, if
there is a commit which revision number N is a prefix of another commit A then
`hg log -r N` outputs commit A instead of commit N. Let's get rid of this
wrapping.
Test Plan: arc unit
Reviewers: #mercurial, mjpieters
Reviewed By: mjpieters
Subscribers: mjpieters, #sourcecontrol
Differential Revision: https://phabricator.intern.facebook.com/D4635311
Tasks: 16394593
Signature: t1:4635311:1488393427:eae091a0d155df336adbb122f2881a28e95ba6b6
Summary:
The sparse profile in .hg/sparse currently only can grow, removing rules from this must either be done manually (by editing the file), or one by one using the the -d/--delete command line switch.
To mirror the hg sparse --import-rules command line, hg sparse --clear-rules removes local includes/excludes rules.
Test Plan: fb-hgext/tests/test-sparse-clear.t
Reviewers: marianomartin, mjpieters
Reviewed By: mjpieters
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4609102
Tasks: 16073155
Signature: t1:4609102:1488028035:f5f7b7420d911667fb4a07ac770b7e0480a76977
Summary:
We have users who don't use inhibit, so printing this message clutters their UI.
Let's just move this debug.
Test Plan: yolo
Reviewers: kulshrax, #mercurial, quark
Reviewed By: quark
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4614036
Signature: t1:4614036:1487986024:6945b057aae62fc7a9fded9474d47ba134496e5a
Summary:
The current code checks if the filter error reports a rev number, and if it does
tries to convert it to a node. If the 'number' is actually just a short form of
a hash that happens to be all digits, then it may be a number greater than the
size of the changelog, so we need to not attempt to access it in the changelog,
otherwise we get an IndexError.
Test Plan: Added a test, it failed before with a weird error
Reviewers: #mercurial, kulshrax
Reviewed By: kulshrax
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4613988
Signature: t1:4613988:1488027938:4f5ffcb07aeecbaf59cad20fb3a592dd92ed2beb
Summary:
`mercurial.revset` was recently split into `revset + revsetlang + smartset`.
Update our code accordingly.
D4604848 has fixed the `revsetlang` part. This patch fixes the remaining
`smartset` part.
Also fixes some test failures introduced by D4547080.
Test Plan: `arc unit`
Reviewers: #mercurial
Subscribers: jeroenv, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4595417
Summary: creating a mercurial extension so that `hg record` functionality can be used via a text editor (instead of interactive command prompt or curses menu)
Test Plan: `$ run-tests.py test-edrecord.t`
Reviewers: durham, skreyen, most, #sourcecontrol, asriram, simonfar
Reviewed By: simonfar
Subscribers: quark, rmcelroy, simonfar, mjpieters, #nuclide
Differential Revision: https://phabricator.intern.facebook.com/D4547080
Tasks: 15728734
Signature: t1:4547080:1487842609:8855e35010cac3b6d2b136a80d7effa369eb1da8
Summary: The data was being sent to scuba, but then ignored altogether. Turn this into a single string instead.
Test Plan: arc unit
Reviewers: #sourcecontrol, andrasbelo
Reviewed By: andrasbelo
Subscribers: mjpieters, andrasbelo
Differential Revision: https://phabricator.intern.facebook.com/D4598196
Signature: t1:4598196:1487784582:8952e0e386325d1635029aab2b9d97e9284376ea
Summary:
Let's add a script that can be used as a commit hook to prevent bad commit
messages from happenning.
Test Plan:
Run the following perl script:
system("echo 1 >> 1");
system("hg add 1");
system("hg ci -m"."\x80");
system("hg ci -m"."\x01");
system("hg ci -m ok");
Output:
fbcode/1 already tracked!
transaction abort!
rollback completed
abort: decoding near '': 'utf8' codec can't decode byte 0x80 in position 0: invalid start byte!
non-printable characters in commit message
transaction abort!
rollback completed
abort: pretxncommit hook failed
Make sure that commit is made with "ok" message, but there are no commits with
bad commit messages
Reviewers: #sourcecontrol, tja
Reviewed By: tja
Subscribers: tja, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4597432
Tasks: 16212973
Signature: t1:4597432:1487761552:cf38eabf93374b0ec2feb653dd70033de25e6e0e
Summary:
The profiling extension is to log interactive time.
The upstream change 459366b580cf breaks the profiling extension as it could
no longer wrap `maybeprofile` in time. And @simonfar's
`ui.timeblockedsection` is going to be a superior solution. So just remove
the profiling extension and its tests.
Test Plan: Checked that the extension is not enabled in our config files.
Reviewers: simonfar, #mercurial
Subscribers: mjpieters, simonfar
Differential Revision: https://phabricator.intern.facebook.com/D4583087
Summary:
97d0be4019ac in core broke the fastlog extension for multi-path
queries. We did not have a test to catch it. It was reported as broken by a
user.
Test Plan:
```
$ ../../hg/tests/run-tests.py test-fastlog.t
```
Reviewers: #sourcecontrol
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4579133
Tasks: 16154586
Summary:
This command will allow us to expose simple api to all automation checking
working copy parent. It's meant to be overridden by our wrapper to read the
dirsate directly without shelling out to hg (see D4454504 )
Test Plan: see tests
Reviewers: #mercurial, durham, rmcelroy
Reviewed By: rmcelroy
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4454502
Signature: t1:4454502:1485441135:09902c6437a7af55dfa5c97e3165681de687231f
Summary:
fastpartialmatch index may be incorrect if fastpartialmatch extension was
disabled, then new commits were added and then fastpartialmatch was
enabled back. We need to detect this situation and rebuild the index in
this case.
One approach would be to use _handleinconsistentindex() function to mark index
as invalid. There are a couple of drawbacks to this approach:
1) It requires more significant code changes
2) The output may still be incorrect if hash with the same prefix was added while fastpartialmatch was disabled.
Another approach would be to use header version from index files
(see class _header). That's even worse because we'll need to bump index
version in the code whenever we disable fastpartialmatch.
Third approach would be to use changelog mtime + size as a validation key but
then we need to ensure partialmatch index is written after changelog.
Instead I suggest to use generation numbers. It's just a number in the config
and in the special file inside partialmatch index directory. If the
two doesn't match then index is incorrect and needs to be deleted. The biggest
drawback to this approach is that we need to bump generation number each time
the extension is enabled. I think it should be fine because the one who enables
extensions usually understands it and he or she will bump the generation number.
Test Plan: arc unit
Reviewers: #sourcecontrol, durham, mjpieters
Reviewed By: mjpieters
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4571926
Signature: t1:4571926:1487263035:53419c9ca4e344127f391ac19c1071a826fea3f1
Summary:
User reported failures because revlog.opener doesn't have `exists` function.
The reason is that sometimes opener is wrapped inside the function.
Let's use _realopener instead
Test Plan: arc unit
Reviewers: #sourcecontrol, mjpieters
Reviewed By: mjpieters
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4571659
Signature: t1:4571659:1487267797:a950ab7c88d75573df64ec1964755c054307c7ea
Summary:
Don't let errors propagate out of the gitnode() revset. Always report errors
in gitnode() as a translation failure, rather than letting exceptions propagate
up and crash mercurial. The code was previously only catching internally
generated ConduitError exceptions, but it can also throw HttpError exceptions,
and the underlying httplib code can throw its own exceptions as well as
socket.error exceptions.
This also fixes the test code to use an ephemeral TCP port rather than assuming
port 8543 will always be available. Using a fixed TCP port in test code is a
very common way to cause bogus failures if the tests are run in parallel.
(For instance, testing multiple repositories in parallel on the same build
host.)
Test Plan:
Added unit tests that check the behavior when the server returns a 500 error,
and when the server refuses the connection entirely.
Reviewers: quark, durham, rmcelroy
Reviewed By: rmcelroy
Subscribers: net-systems-diffs@fb.com, yogeshwer, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4556871
Signature: t1:4556871:1487062142:b58d770d46c975d44933bec08cfce8acb25ff16b
Summary:
Previously, smartlog use `tip` as the fallback master, and prompts:
warning: there is no master changeset locally, try pulling from server
if master is `tip`.
That makes smartlog less friendly for external users. This patch removes the
prompt and changes the fallback master to the last public revision so it
works out-of-box.
Test Plan: `arc unit`
Reviewers: #mercurial, durham
Reviewed By: durham
Subscribers: durham, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4555346
Signature: t1:4555346:1487036530:35a5fa65d5419e7324d3203fa52de9a82b8132e3
Summary:
This diff makes the following modifications to the (former) `hiddenhash` extension:
- Makes the error messages configurable, allowing us to easily change the error messages in `facebook.rc`.
- Renames the extension to `hiddenerror`, since the primary feature is now the configurable error messages rather than using a hash in the message.
- Changes the method by which the extension wraps the `changectx` constructor. The old method of wrapping would result in errors in code paths that in `context.py` that performed `isinstance()` checks on contexts.
Test Plan: To test arc patch this extension into your `facebook-hg-rpms` and enable it for a test repo. Then, attempt to access a hidden node. You should see the default error message defined in the extension. Setting `hiddenerror.message` and `hiddenerror.hint` should allow you to customize the messages. See test file for examples.
Reviewers: #mercurial, rmcelroy
Reviewed By: rmcelroy
Subscribers: rmcelroy, durham, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4548140
Tasks: 15896355
Signature: t1:4548140:1487063552:06395c1eaaa83b00ead8e80e2cd8f9d13f7d08e2
Summary: Sometimes you want to be able to justify why a line is in a sparse profile. Support empty lines and comments (starting with a hash).
Test Plan: Run the test runner
Reviewers: #mercurial, rmcelroy, durham
Reviewed By: durham
Differential Revision: https://phabricator.intern.facebook.com/D4537617
Tasks: 15993301
Signature: t1:4537617:1486687441:18a00b65df244b573bab095809fb1fbb57e21d71
Summary: If `hg debuginhibit` is run with no arguments, print out a list of inhibited nodes. This is similar to how `hg debugobsolete` works and will assist with debugging.
Test Plan: Run `hg debuginhibit` in a repo with inhibited nodes. (This can be tested by either stripping some test commits and then updating back to them, or by manually inhibiting the commits with `hg debuginhibit`.) A list of inhibited nodes should be printed. See test file for an example.
Reviewers: #mercurial, rmcelroy
Reviewed By: rmcelroy
Differential Revision: https://phabricator.intern.facebook.com/D4541514
Tasks: 16004717
Signature: t1:4541514:1486722366:10b4232830ed88931bd6ac610972114743797192
Revert Plan:
Summary:
To get rid of this DeprecationWarning:
--- /data/hgbuild/facebook-hg-rpms/fb-hgext/tests/test-hiddenhash.t
+++ /data/hgbuild/facebook-hg-rpms/fb-hgext/tests/test-hiddenhash.t.err
@@ -10,5 +10,7 @@
$ hg debugbuilddag +1
$ hg debugobsolete 1ea73414a91b0920940797d8fc6a11e447f8ea1e
$ hg log -r 0
+ $TESTTMP/hiddenhash.py:33: DeprecationWarning: BaseException.message has been deprecated as of Python 2.6
+ match = re.match(r"hidden revision '(\d+)'", e.message)
abort: hidden changeset 1ea73414a91b!
[255]
let's not access `.message` and use str(e) instead
Test Plan:
arc unit,
Will test on centos6 soon
Reviewers: #sourcecontrol, simonfar
Reviewed By: simonfar
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4521398
Signature: t1:4521398:1486470861:e39e4639346d0f1f2e4a69fc38ccf34fbc432df6
Summary:
HG_SAMPLING_FILEPATH was renamed to SCM_SAMPLING_FILEPATH long time ago.
Let's update the comment
Test Plan: Not tested
Reviewers: #sourcecontrol, quark
Reviewed By: quark
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4507708
Signature: t1:4507708:1486141731:8a96ac74006cf9d240ad4cbbff928fe49fe71cb3
Summary:
Previously, when a command tried to access a hidden revision without the `--hidden` flag, the user would get the following error:
```
abort: hidden revision '0'!
(use --hidden to access hidden revisions)
```
With this extension, the error is now:
```
abort: hidden commit b8144197c244!
```
This way, novice users aren't tempted to immediately re-run the command with `--hidden`, which may have undesired results. It also makes it much easier to find the commit in question in `hg sl --hidden` since by default smartlog does not show rev numbers.
Test Plan: Attempt to access a hidden commit, observe new error message as seen above. Note that in the test file the word "changeset" is present instead of "commit" due to the lack of the `dialect` extension in the test.
Reviewers: #mercurial, simonfar
Reviewed By: simonfar
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4506065
Tasks: 15747879
Signature: t1:4506065:1486122372:d82ebd488a00d1958c235f733a9455111954c116
Summary:
turned out that transaction can be None in some cases.
Let's just write to the cache in this case.
Test Plan: arc unit
Reviewers: #sourcecontrol, durham, andrasbelo
Reviewed By: andrasbelo
Subscribers: andrasbelo, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4507653
Tasks: 15862906
Signature: t1:4507653:1486133933:de0997ef1724ba1082c3597ea64aab760f670b47
Summary:
Recently @rmcelroy discovered that attempting to rebase a single commit in the middle of a stack results in the rest of the stack being marked as unstable. See https://fb.facebook.com/groups/sourcecontrol/permalink/1207767375939547/ for full example.
It turns out that the inhibit extension's post-transaction hook that would ordinarily inhibit the rebased (and thus obsolete) commit, and thereby hide the instability, was not firing in this case. This was similar to a previous problem with `hg next --rebase` wherein `--continue`'ing a rebase with conflicts would result in visible instability for the same reason. In both cases, the problem appears to be that `rebase` doesn't actually create a transaction, so of course the post-transaction hook doesn't fire.
The solution for `hg next --rebase` was to add a (nearly) empty transaction after the rebase call to trigger the hook. This diff just makes this the default behavior for `hg rebase` so that the hook is always called.
Test Plan:
See new test file. Basically, create a situation like:
```
o e5d56d debugbuilddag
| r3
|
| o c175ba debugbuilddag
| | r2
| |
| o 220949 debugbuilddag
|/ r1
|
o 1ad88b debugbuilddag
r0
```
Perform `hg rebase -r 1 -d 3`, which rebases commit "r1" onto "r3", which would ordinarily make "r2" unstable. Instead, the result should be as below, with no visible instability because "r1" is inhibited.
```
o 738e8e debugbuilddag
| r1
|
o e5d56d debugbuilddag
| r3
|
| o c175ba debugbuilddag
| | r2
| |
| x 220949 (Rebased as 738e8e) debugbuilddag
|/ r1
|
o 1ad88b debugbuilddag
r0
```
Reviewers: rmcelroy, #mercurial, durham
Reviewed By: durham
Subscribers: mjpieters, rmcelroy
Differential Revision: https://phabricator.intern.facebook.com/D4491870
Tasks: 15697758
Signature: t1:4491870:1486033128:45b8e9a72af1e88ab78484ddef47a788d1aca7b5
Summary: Sort the output to fix tests on different platforms
Test Plan: arc unit on mac
Reviewers: #sourcecontrol, simonfar, ikostia
Reviewed By: simonfar, ikostia
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4501449
Signature: t1:4501449:1486050577:3041dc3c3691f48e39a0e9e05b1f97954a6d88e2
Summary:
If we have too many unsorted entries then finding matching revisions is
slower. To fix it we will rebuild partial index from time to time.
_findcandidates function marks index as need-to-be-rebuilt if it notices
index file with too many unsorted entries. Then _changegrouphook rebuilds
the index.
Test Plan: arc unit
Reviewers: #sourcecontrol
Differential Revision: https://phabricator.intern.facebook.com/D4494617
Summary:
Since we have some of the nodes sorted we can now use bisect to quickly find
matching nodes.
Test Plan: arc unit
Reviewers: #sourcecontrol
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4494419
Summary:
Append header to each fast partial match index file. For now it contains
version number and number of sorted entries in the file.
File is ignored if version number from file is different.
Test Plan: arc unit
Reviewers: #sourcecontrol
Differential Revision: https://phabricator.intern.facebook.com/D4488417
Summary:
We'll add bisect algorithm soon.
Let's sort the nodes first during index rebuilding
Test Plan: arc unit
Reviewers: #sourcecontrol
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4488263
Summary:
strip may change revision numbers of many nodes at once.
It's better to rebuild index in case of strip.
Test Plan: arc unit
Reviewers: #sourcecontrol
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4474710
Summary:
This hook builds index if it's not present and it appends new entries
from the changegroup.
Test Plan: arc unit
Reviewers: #sourcecontrol
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4474704
Summary: These two functions actually use fastpartial index to resolve revision for a node.
Test Plan: arc unit
Reviewers: #sourcecontrol
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4474658
Summary:
Add commit hook to append new entries to the index.
It uses the same transactions as revlogs.
Do not append entries if index is not built (i.e.
if _partialindex does not exist). This is to ensure that we either
have complete index or none at all. Besides next diff will add
hooks to build index after pull if it doesn't exist.
Test Plan: arc unit
Reviewers: #sourcecontrol
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4474654
Summary:
Index to make node partial match fast i. e. speed up matching of partial hashes like f9bbd5.
This is an initial diff and it doesn't add any useful functionality, just debug commands.
Storage format is simple. There are a few files (at most 256).
Each file contains entries:
<20-byte node hash><4 byte encoded rev>
Each entry represents a commit (node).
Name of the file is the first two letters of the hex node hash. Nodes with the
same first two letters go to the same file. Nodes are NOT sorted inside the file
to make appends of new nodes easier.
Partial index should always be correct i.e. it should contain only nodes that
are present in the repo (regardless of whether they are visible or not) and
rev numbers for nodes should be correct too.
Test Plan: arc unit
Reviewers: #sourcecontrol
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4474642
Summary: Add a new extension to assist with debugging issues with the inhibit extension. Adds the `hg debuginhibit` and `hg debugdeinhibit` commands to manually inhibit/deinhibit commits as well as config options to print out the nodes being inhibited and truncated stack traces of each call site. I mostly made this into an extension so that it would be easy to debug inhibit issues from any user by just asking them to run their command with the appropriate `--config` flag.
Test Plan: See test file. Enabling the `debuginhibit.printnodes` and `debuginhibit.printstack` functions should result in the relevant nodes and stack traces to be printed whenever nodes are inhibited and deinhibited.
Reviewers: durham, rmcelroy, #mercurial, mitrandir, simonfar
Reviewed By: simonfar
Subscribers: mitrandir, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4485659
Signature: t1:4485659:1485985964:e6920b67450942c706c9523e9df48b4093b26d69
Summary: Although the split and fold commands usually abort on error, in some cases they simply write to stderr and return a nonzero exit status. Previously, the wrappers in fbamend would expect failures to result in an abort, and therefore didn't return early. This resulted in restack errors when this was not the case -- notably in the case of folding a commit with itself. This diff fixes the problem by returning early on error.
Test Plan: Attempting to fold a commit with itself (i.e., `hg fold --exact . .`) fails with an error message instead of crashing.
Reviewers: #mercurial, simonfar
Reviewed By: simonfar
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4493029
Tasks: 15704338
Signature: t1:4493029:1485986404:17186b47786cf79c735b33adbb19de79ac75a5be
Summary: Previously, if a user ran split or fold on commits with bookmarks on them, those bookmarks would remain on the original commit instead of the new split or folded commits. This behavior was unexpected and confusing because it meant the obsolete commit would remain visible and the user would be forced to manually move the bookmark. This change makes split and fold automatically move any bookmarks on the old commits onto the newly created commits.
Test Plan: Run `hg split` or `hg fold` on a commit or set of commits with one or more bookmarks, and observe that the bookmarks are updated to point to the newly created commits. See tests for examples.
Reviewers: #mercurial, simonfar
Reviewed By: simonfar
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4490655
Tasks: 15595449
Signature: t1:4490655:1485986009:858f099d3f95c6a45fe8b119d4f8d376b87652b7
Summary:
In certain situations, like infinitepush, the noderev cache builders would try
to save a node that didn't exist in the repo. This caused an exception. Now we
check each node before we attempt to resolve it to a rev.
Test Plan: Added a test. It failed before, and passes now.
Reviewers: #mercurial, andrasbelo
Reviewed By: andrasbelo
Subscribers: andrasbelo, mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4494848
Signature: t1:4494848:1485964835:986d4fe3e29404cac22e4acf7d88c3715fcb51b9
This is an import of a bugfix in upstream evolve by Martin.
As reported in the issue, "hg uncommit" would lose copy information of
the files that remained in the commit. The problem was simply that the
a dict of copies was iterated as "src, dst in copies.items()" where it
should have been "dst, src ...".
Summary:
Previously the noderev cache only covered the branchmap. These days remotenames
and phases are also doing a lot of node->rev look ups, so let's add them to the
noderev cache as well.
Test Plan:
Compared hg book before and after the change. Noticed a significant
perf increase.
Reviewers: #mercurial, ikostia
Reviewed By: ikostia
Subscribers: mjpieters
Differential Revision: https://phabricator.intern.facebook.com/D4487962
Signature: t1:4487962:1485871615:634d95f61e916d81440a0f440010912af5ae6a13
Summary: This is basically the same as `hg amend -i` but since allowunstable operates at the command function level we need to explicitly whitelist this command. Note that I've set the obsmarker metadata to "amend" so that commits affected will be marked as "Amended as XXX" in smartlog. I don't think any other `hg record` flags other than --amend create obsmarkers so this should be OK.
Test Plan: Run `hg record --amend` in the middle of a stack and observe that it works now.
Reviewers: durham, #mercurial, rmcelroy
Reviewed By: rmcelroy
Subscribers: mjpieters, reedriley
Differential Revision: https://phabricator.intern.facebook.com/D4466338
Tasks: 15705764
Signature: t1:4466338:1485424744:442604ce2089298f0724c27025502a3602113401
