Summary:
Ref T13098. See PHI858. If you write this at the end of a message in `arc diff`:
```
Subscribers:
#projectname
# NEW DIFFERENTIAL REVISION
# Describe the changes in this new revision.
# ...
```
...we'll currently eat the `#projectname` as an instructional comment, even if it is followed by an empty line.
Instead, stop eating stuff once we hit the first empty line. (We escape empty lines in comments already.)
After T13098 I'll maybe adjust this to use a more explicit instruction escape, like `##`, since there's no reason we're bound to `#`.
Test Plan: Added a unit test and made it pass.
Reviewers: amckinley
Reviewed By: amckinley
Maniphest Tasks: T13098
Differential Revision: https://secure.phabricator.com/D19639
Summary:
Ref T13187. See PHI838. If two hunks are separated by 7 lines of context, we can render them as either:
```lang=diff
+ Hunk A
Context 1
Context 2
Context 3
Context 4
Context 5
Context 6
Context 7
+ Hunk B
```
...or:
```lang=diff
+ Hunk A
Context 1
Context 2
Context 3
@@ +1,2 -3,4 @@
Context 5
Context 6
Context 7
+ Hunk B
```
Since we get the same number of output lines either way and the first one is more human-readable, we picked that one.
However, `diff -u` does the second one. Since human-readability is probably less important than compatibility, change the behavior to be more similar to `diff -u`.
Test Plan: Added unit tests for the edge cases with default parameters (6 context lines, 7 context lines) and made them pass.
Reviewers: amckinley
Maniphest Tasks: T13187
Differential Revision: https://secure.phabricator.com/D19603
Summary:
Ref T13130. I want to take a crack at improving performance here, but two possible approaches (inlining the actual encoding; using integers if they're big enough) aren't easy to test right now.
Restructure the tests so they can support these kinds of refactoring.
The "32bit" and "64bit" modes currently do the same thing, but I expect to introduce introduce separate encoding pathways in a future change, if the profiler says it actually helps.
(I'll hold this and everything that comes after it until I make meaningful performance improvements.)
Test Plan: Ran `arc unit`, got passes on tests.
Reviewers: amckinley
Reviewed By: amckinley
Maniphest Tasks: T13130
Differential Revision: https://secure.phabricator.com/D19407
Summary:
Fixes T8768. See PHI294. See that task for more details.
Git, Mercurial, `diff`, and `patch` have conspired to make things weird. To correctly handle files with spaces in the way everything else does and expects, we need to emit semantic trailing whitespace literals.
Test Plan:
- Created a file with spaces in it in a Mercurial repositroy, committed it, diffed it into a revision.
- Used `arc patch` to apply the change to a clean copy of the repository.
- Before patch: Mercurial incorrectly creates a file named `X`, not a file named `X Y.txt`.
- After patch: `arc patch` commit is identical to genuine commit.
- Also added test coverage. The other general behaviors here are fairly well covered already.
Reviewers: amckinley
Reviewed By: amckinley
Maniphest Tasks: T8768
Differential Revision: https://secure.phabricator.com/D18869
Summary:
The previous parser failed when only one of the old/new filenames was
quoted, as with a rename of a Unicode filename to a non-Unicode
filename, or vice versa. It also incorrectly parsed custom prefixes,
even going to far as to encode this logic in the tests: a diff of
"src/file dst/file" which is not a rename should not be recorded as
changing "src/file", but rather "file".
Switch to using the "rename from" / "rename to" as the source of truth
for old and current filenames when complex renames are done. This
matches the logic that git itself uses to parse patches; the contents
of the `diff --git` line are merely viewed as a simplest-case
prediction, with renames handled later should it not match.
The diff parser already had logic to parse "rename from" / "rename to"
lines and extract their information. As such, this diff consists
primarily of removing logic from the `splitGitDiffPaths` method, and
allowing it to quietly fail.
This resolves two ambiguity mentioned in comments and tests: in
renaming "old file old" to "file", as well as the renaming of
"old file with spaces" to "new file with spaces" when neither are
quoted.
Test Plan:
`arc unit`. Many of the existing test cases no longer
applied to the `splitGitDiffPaths` method; they were moved into a new
test method which also supplies values for "rename from" and "rename
to" lines.
As noted in the summary, this alters one of the expected values of an
existing test. Specifically, the following diff is a file addition of
`file` with custom prefixes, and not the addition of "dst/file":
```
diff --git src/file dst/file
new file mode 100644
index 0000000..1269488
--- /dev/null
+++ dst/file
@@ -0,0 +1 @@
+data
```
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: Korvin, epriestley
Differential Revision: https://secure.phabricator.com/D16405
Summary: See D14232. That didn't actually work. It looks like this does.
Test Plan:
- Ran `git commit --author ...` on build server and saw the same failure.
- Ran `git -c ... -c ... commit ...` on build server and saw it work.
Reviewers: chad
Reviewed By: chad
Differential Revision: https://secure.phabricator.com/D14233
Summary:
On `sbuild`, we currently get a failure on this test. Use an explicit `--author` so we can run the test even if `user.email` and `user.name` are not set in global Git config.
```
FAIL ArcanistBundleTestCase::testGitRepository
15 EXCEPTION (CommandException): Command failed with error #128!
16 COMMAND
17 git commit -m 'Mark koan2 +x and edit it.'
18
19 STDOUT
20 (empty)
21
22 STDERR
23
24 *** Please tell me who you are.
25
26 Run
27
28 git config --global user.email "you@example.com"
29 git config --global user.name "Your Name"
30
31 to set your account's default identity.
32 Omit --global to set the identity only in this repository.
33
34 fatal: empty ident name (for <builder@sbuild001.phacility.net>) not allowed
```
Test Plan: Ran `arc unit --everything`. Will verify in production.
Reviewers: chad
Reviewed By: chad
Differential Revision: https://secure.phabricator.com/D14232
Summary: I found a few strings that I had missed, using a mostly-broken-but-somewhat-okay custom linter ruler (https://secure.phabricator.com/differential/diff/30988/).
Test Plan: Intense eyeballing.
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: aurelijus, Korvin, epriestley
Differential Revision: https://secure.phabricator.com/D12888
Summary: Ref T7977. The `ArcanistTestCase` class is pointless and can be replaced by `ArcanistPhutilTestCase`. Furthermore, it sorta makes sense to just rename `ArcanistPhutilTestCase` to `PhutilTestCase`. Depends on D12664 and D12666.
Test Plan: `arc unit`
Reviewers: avivey, #blessed_reviewers, epriestley
Reviewed By: avivey, #blessed_reviewers, epriestley
Subscribers: aurelijus, Korvin, epriestley
Maniphest Tasks: T7977
Differential Revision: https://secure.phabricator.com/D12665
Summary:
Fixes T5870. When a filename contains spaces, Git will add a "\t" at the end of the "---" and "+++" lines. We currently consume this and think we're reading a file called "blah blah\t".
Instead, parse it out as not part of the filename.
Test Plan: Added failing unit test and made it pass.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T5870
Differential Revision: https://secure.phabricator.com/D10267
Summary:
Fixes T5555. Normally, when we `svn diff subdir/`, we use `--depth empty` to get only changes for the directory itself (usually, property changes).
However, this flag has no effect if the directory is newly added.
Adjust the diff parser so that if two sets of hunks are specified for a single file in a raw diff, we let the last one win instead of including both. This approach is a broadly more reasonable interpretation of these diffs.
Test Plan:
- Added a new file in a new subdirectory in Subversion.
- Ran `arc diff --only`.
- No double file content in resulting diff.
- Added unit test.
- There's fairly comprehensive unit test coverage for this stuff.
Reviewers: btrahan, chad
Reviewed By: chad
Subscribers: epriestley
Maniphest Tasks: T5555
Differential Revision: https://secure.phabricator.com/D9921
Summary: I'm pretty sure that `@group` annotations are useless now... I believe that they were originally used by Diviner?
Test Plan: Eye-balled it.
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: epriestley, Korvin, aurelijus
Differential Revision: https://secure.phabricator.com/D9855
Summary:
This adds support for passing range of commits for arc diff. This is useful when you want to submit code reviews for past commits without mucking around with the working copy.
This will probably require changes :)
Test Plan: Tested locally, but totally need to add tests for this
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: epriestley, Korvin
Differential Revision: https://secure.phabricator.com/D9369
Summary: Applied various linter fixes. Also make the `.editorconfig` file a bit more specific. Unfortunately, `arc lint --apply-patches` currently modifies some test data that it shouldn't, but this should be fixed after T5105.
Test Plan: Ran `arc unit` to make sure things weren't broken.
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: epriestley, Korvin, aurelijus
Differential Revision: https://secure.phabricator.com/D9440
Summary: Ran `arc lint --apply-patches --everything` over rARC, mainly to change double quotes to single quotes where appropriate. These changes also validate that the `ArcanistXHPASTLinter::LINT_DOUBLE_QUOTE` rule is working as expected.
Test Plan: Eyeballed //most// of the diff.
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: epriestley, Korvin, aurelijus
Differential Revision: https://secure.phabricator.com/D9269
Summary: Ref T4697. This is similar to the existing stuff, but `svnlook diff ...` can also produce a "Copied" header. Currently, we choke on it in Herald when running pre-commit rules.
Test Plan: Added and ran tests. In the next diff, used this in a real system.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T4697
Differential Revision: https://secure.phabricator.com/D8653
Summary:
Fixes T4063. The `git format-patch` command produces a special header
and footer which we need to detect, strip, and parse.
Test Plan:
- Added and ran unit tests.
- Submitted a diff with `git format-patch HEAD^ --stdout | arc diff --raw`.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T4063
Differential Revision: https://secure.phabricator.com/D8547
FreeBSD 9.2 comes with diff tool version 2.8.7 which behaves
a bit different from how it is expected to. Namely for diff
between two binary files it says:
Files A and B are differ
This was leading to an exception when browsing revisions with
changes in binary files.
Tweaked parse patterns in order to fix this issue. Now both
older and newer diff tools are supported.
See: <https://github.com/facebook/arcanist/pull/139>
Reviewed by: epriestley
Summary: Ref T4195. When figuring out changed content in SVN, the easiest approach is to use `svnlook diff`, but it has a slightly different header than we're used to. Adjust the parser for it and add some tests.
Test Plan: Clean unit tests.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4195
Differential Revision: https://secure.phabricator.com/D7790
Summary:
Create a new class for them, pass instance around as need.
This looks like it's mostly working, but I'd like to replace the various `new ArcanistConfigurationManager()`
calls with something more suitable.
And maybe get a better name for ArcanistConfigurationManager ("Configuration" is already taken).
Test Plan: arc unit --everything, and then some.
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley
CC: Korvin, epriestley, aran, chad
Differential Revision: https://secure.phabricator.com/D7271
Summary:
See <https://github.com/facebook/phabricator/issues/400>. Since all of `patch`, `git apply` and `hg export` either accept or emit header comments, parse them unconditionally.
This is a tiny bit messy because we already had a less-general parser for `hg export` diffs, which have a large header section.
This is less permissive than GNU `patch`, which allows comments anywhere. We could do that, but `git apply` won't read them and they seem pretty crazy.
Test Plan: Added and ran unit tests.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D7207
Summary: Added some sample rcsdiffs for adding and deleting a line from a file. Wrote some test cases to be tested by ArcanistDiffParser.
Test Plan: By making all the test cases pass.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin, AnhNhan
Differential Revision: https://secure.phabricator.com/D5324
Summary: We were incorrectly matching `$` in the regexp against a possible `\r\n`. I missed this earlier when trying to catch all of these.
Test Plan:
- Added unit test and made it pass.
- Did another search for `getLine()` to see if I could spot any more of these, but failed to identify any via inspection.
Reviewers: vrana, mbishopim3
Reviewed By: mbishopim3
CC: aran
Differential Revision: https://secure.phabricator.com/D5038
Summary:
This test currently chdir()'s into a directory which is later removed. If another test tries to run a shell script while the CWD is invalid, the shell may emit this to stderr:
shell-init: error retrieving current directory: getcwd: cannot access parent directories: No such file or directory
Among other things, this can cause the XHPAST test to fail, because it detects syntax errors by examining stderr.
Instead, retore the directory.
Test Plan: Ran "arc unit --everything", which could previously fail if XHPAST ran after Bundle.
Reviewers: btrahan, vrana
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D4738
Summary: Fixes T2112. These are fairly common now, and are used as the storage format for `hg export` and mq in most installs.
Test Plan: Ran unit tests, used `arc patch --patch`, uploaded some diffs manually.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T2112
Differential Revision: https://secure.phabricator.com/D4592
Summary:
This method is used in three cases:
# For unit tests, to set the range to 'HEAD^' or '.^' in an agnostic way.
# For "amend", to set the range to the commit to be amended (also 'HEAD^' or '.^').
# For "patch" and "upgrade" so we don't fail just because there's an invalid "base" rule somewhere in the config when doing clean-working-copy tests.
For cases (1) and (2), introduce an "arc:this" rule to mean "the current commit". For case (3), remove the call; it is no longer necessary to check the commit range in order to do tests for the working copy state after D4095.
Test Plan: Ran unit tests, "arc upgrade", "arc patch", "arc amend".
Reviewers: vrana
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D4096
Summary: See D3963. Instead, parse these diffs so they'll work with `--raw`, etc.
Test Plan:
Generated a failing diff, added it as a test case. Fixed issue. Ran test suite. Ran `arc` against it:
$ git -c diff.suppress-blank-empty=true diff HEAD | arc diff --raw --only --conduit-uri=http://local.aphront.com:8080/
Reading diff from stdin...
Created a new Differential diff:
Diff URI: http://local.aphront.com:8080/differential/diff/103/
Included changes:
M things
Reviewers: vrana, jiiix, btrahan
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D3969
Summary:
When we hit a diff which is missing file context, we try to pull it synthetically later. This works for moves and copies, but currently fails for property changes. Since it failed, we didn't have context, so we'd try to pull it again...
The general problem this creates is that when you mark a file "+x" without changing it, we can't show you the content in Differential. Not a huge deal. In some future diff, I'll build the content synthetically.
Adds commits to cover this behavior:
commit 1830a13adf764b55743f7edc6066451898d8ffa4
Author: epriestley <git@epriestley.com>
Date: Tue Nov 6 17:11:18 2012 -0800
Mark koan2 +x and edit it.
commit 8ecc728bcc9b482a9a91527ea471b04fc1a025cf
Author: epriestley <git@epriestley.com>
Date: Tue Nov 6 17:08:44 2012 -0800
Move 'text' to 'executable' and mark it +x.
commit 39c8e7dd3914edff087a6214f0cd996ad08e5b3d
Author: epriestley <git@epriestley.com>
Date: Tue Nov 6 16:36:59 2012 -0800
Mark koan as +x.
Test Plan: Ran unit tests. Previously, they looped indefinitely. Now, they pass.
Reviewers: vrana, btrahan
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D3909
Summary:
This commit doesn't change license of any file. It just makes the license implicit (inherited from LICENSE file in the root directory).
We are removing the headers for these reasons:
- It wastes space in editors, less code is visible in editor upon opening a file.
- It brings noise to diff of the first change of any file every year.
- It confuses Git file copy detection when creating small files.
- We don't have an explicit license header in other files (JS, CSS, images, documentation).
- Using license header in every file is not obligatory: http://www.apache.org/dev/apply-license.html#new.
This change is approved by Alma Chao (Lead Open Source and IP Counsel at Facebook).
Test Plan: Verified that the license survived only in unit tests and LICENSE file.
Reviewers: epriestley, btrahan, edward
Reviewed By: epriestley
CC: aran, Korvin, davidrecordon
Maniphest Tasks: T2035
Differential Revision: https://secure.phabricator.com/D3881
Summary:
See T1973. In T1675, we addressed parsing of diffs with `--no-prefix` or custom `--src-prefix` and `--dst-prefix` flags. However, this inadvetently broke diffing of files with spaces in their names, which Git does not quote. They look like this normally:
diff --git a/old file b/new file
Prior to D3744, we accidentally got this right by looking for the `a/` and `b/`. However, we no longer do, and instead produce nonsense results.
This problem is difficult because for files with spaces, `git diff --no-prefix` may generate an ambiguous line like:
diff --git a b c d e f g
From this line, we have no way to deterine if this moves "a" to "b c d e f g", or "a b c d" to "e f g", or anything in between. In some diffs we have more information later on, but in some cases we do not, e.g. for binary diffs without `--binary`.
Try to get this right in as many cases as possible:
- If there are quotes, we can unambiguously get it right. This only happens for filenames with quotes or unicode characters, however.
- If there is exactly one space, we can unambiguously get it right.
- Interpret the common case of `a/<anything> b/<anything>` in the most-likely-correct way again.
- Interpret the rare case of `<anything> <that same thing>` in the most-likely-correct way.
- Complain about any `a b c d e f g` garbage.
Test Plan: Ran unit tests. Created a diff of a file called "File With Spaces".
Reviewers: vrana, btrahan
Reviewed By: btrahan
CC: aran, ReturnZero
Maniphest Tasks: T1973
Differential Revision: https://secure.phabricator.com/D3818
Summary: We need to tweak a few patterns to accommodate the possibility that lines end in "\r\n".
Test Plan: Added failing unit test and made it pass.
Reviewers: vrana, btrahan
Reviewed By: vrana
CC: aran, mbishopim3
Maniphest Tasks: T1944
Differential Revision: https://secure.phabricator.com/D3772
Summary: The way we represent some move/copy stuff is a bit messed up, but it mostly works, so add coverage before I mess with it.
Test Plan: Ran unit tests.
Reviewers: btrahan, vrana
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T866
Differential Revision: https://secure.phabricator.com/D3752
Summary:
We have coverage for generating patches, applying them, and making sure they reproduce the original repository state. However:
- The code uses simplified patch generation which omits some flags like `-M` and `-C`, and generally produces less rich patches than we really produce. Instead, produce patches the same way `arc diff` does.
- We don't test the intermediate change representation. In theory it's not too important because if we get it wrong the output should be wrong, but in practice it makes it easier to nail down issues. We can also generate less-rich patches which still apply correctly, but would prefer not to.
- Similarly, we don't test the intermediate patch representation. This is almost entirely redundant with simply applying the patch, but easier to visualize.
Add coverage for all that stuff and fix some bugs with `-M` / `-C` patch generation that weren't caught under the simpler patch generation.
Test Plan: Ran unit tests.
Reviewers: btrahan, vrana
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T866
Differential Revision: https://secure.phabricator.com/D3751
Summary: Make this harder to get wrong. Instead of requiring a separate call for synthetic data, automatically load it if we can.
Test Plan: Unit tests; `arc diff`.
Reviewers: btrahan, vrana
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T866
Differential Revision: https://secure.phabricator.com/D3750
Summary: See T1675 for discussion. We currently strip `[abicwo12]/` from Git patches, but the user can provide arbitrary prefixes with `--src-prefix` and `--dst-prefix`, or strip prefixes entirely with `--no-prefix`. In these cases, trust they know what they're doing rather than rejecting the diff.
Test Plan: Added a bunch of tests. We have existing tests for `diff.mnemonicprefix` and normal prefixes.
Reviewers: btrahan, vrana
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1675
Differential Revision: https://secure.phabricator.com/D3744
Summary:
Adds a test which goes through a git repository commit by commit and applies them (in effect) via "arc patch", verifying that the results match the actual commit.
See D3439 for the fixture stuff.
The git repo archive has a couple of trivial commits in it:
$ ../libphutil/scripts/utils/directory_fixture.php src/parser/__tests__/bundle.git.tgz
Spawning an interactive shell. Exit when complete.
sh-3.2$ git log
commit f19fb9fa1385c01b53bdb6d8842dd154e47151ec
Author: epriestley <git@epriestley.com>
Date: Wed Sep 5 14:30:28 2012 -0700
Edit a text file.
commit 228d7be4840313ed805c25c15bba0f7b188af3e6
Author: epriestley <git@epriestley.com>
Date: Wed Sep 5 14:30:11 2012 -0700
Add a text file.
Test Plan: Ran tests.
Reviewers: edward
Reviewed By: edward
CC: aran
Maniphest Tasks: T866
Differential Revision: https://secure.phabricator.com/D3440
Summary: See T1635. I'm going to make an effort to rewrite this in a way that's safe in 32-bit PHP, i.e. without bcmath. Add test coverage to limit the chance I screw it up.
Test Plan: Ran unit tests.
Reviewers: btrahan, vrana
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1635
Differential Revision: https://secure.phabricator.com/D3249
Summary: This diff format is used by de-facto mercurial GUI called "TortoiseHG".
It is available as the only way to copy diff into clipboard (right click commit,
select "export", select "copy patch". I added this format support into arcanist
so revisions in Differential can be created from TortoiseHG via simple copy-
paste. Unit test added, manually tested.
See: https://github.com/facebook/arcanist/pull/46
Reviewed by: epriestley
Summary: When you run `hg diff -r x:y`, we get two "-r" arguments in the diff header. Currently, we parse this incorrectly.
Test Plan: Added unit test which previously failed; test now passes.
Reviewers: dschleimer, btrahan
Reviewed By: dschleimer
CC: aran, cakoose
Maniphest Tasks: T1550
Differential Revision: https://secure.phabricator.com/D3061
Summary:
New optional mode. If you set 'base' in local, project or global config or pass '--base' to 'arc diff' or 'arc which', it switches to DSL mode.
In DSL mode, lists of rules from args, local, project and global config are resolved, in that order. Rules can manipulate the rule machine or resolve into actual commits. Provides support for some 'arc' rules (mostly machine manipulation) and 'git' rules (symbolic ref and merge-base).
Test Plan:
Ran unit tests. Also:
```$ arc which --show-base --base 'arc:prompt'
Against which commit? HEAD
HEAD
$ arc which --show-base --base 'git:HEAD'
HEAD
$ arc which --show-base --base 'git:fake'
Usage Exception: None of the rules in your 'base' configuration matched a valid commit. Adjust rules or specify which commit you want to use explicitly.
$ arc which --show-base --base 'git:origin/master'
origin/master
$ arc which --show-base --base 'git:upstream'
Usage Exception: None of the rules in your 'base' configuration matched a valid commit. Adjust rules or specify which commit you want to use explicitly.
$ arc which --show-base --base 'literal:derp'
derp
$ arc which --show-base --base 'arc:halt'
Usage Exception: None of the rules in your 'base' configuration matched a valid commit. Adjust rules or specify which commit you want to use explicitly.
$ arc set-config --local base git:origin/master
Set key 'base' = 'git:origin/master' in local config.
$ arc which --show-base
origin/master
$ arc which --show-base --base 'git:HEAD^'
HEAD^
$ arc which --show-base --base 'arc:yield, git:HEAD^'
origin/master
$ arc which --show-base --base 'arc:global, git:HEAD^'
HEAD^
$ arc which --show-base --base 'arc:global, git:merge-base(origin/master)'
3f4f8992fba8d1f142974da36a82bae900e247c0```
Reviewers: dschleimer, vrana
Reviewed By: dschleimer
CC: aran
Maniphest Tasks: T1233
Differential Revision: https://secure.phabricator.com/D2748
