Summary: Git spits these out with \n at the end.
Test Plan:
```
>>> orbital ~/devtools/arcanist $ git branch --set-upstream trim master
Branch trim set up to track local branch master.
>>> orbital ~/devtools/arcanist $ arc which --base arc:upstream
RELATIVE COMMIT
If you run 'arc diff', changes between the commit:
acf7600e6e Temporarily restore apache/license linters
...and the current working copy state will be sent to Differential, because
it is the merge-base of the upstream of the current branch and HEAD, and
matched the rule 'arc:upstream' in your args 'base' configuration.
You can see the exact changes that will be sent by running this command:
$ git diff acf7600e6e728395..HEAD
These commits will be included in the diff:
3580555e4b30598f WIP
MATCHING REVISIONS
These Differential revisions match the changes in this working copy:
(No revisions match.)
Since there are no revisions in Differential which match this working copy, a
new revision will be created if you run 'arc diff'.
```
Reviewers: brennantaylor
Reviewed By: brennantaylor
CC: aran
Differential Revision: https://secure.phabricator.com/D4913
Summary:
Previously 'arc diff X' with mercurial meant to use X as the base
to diff against. Now it means use gca(X,working directory) as the base to
diff against. This matches the git behavior.
Test Plan:
Ran 'arc diff master' on a repo where master was ahead of the feature branch.
Verified that the diff result included only the diffs in the feature branch.
Reviewers: epriestley, sid0, bos, dschleimer
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D4865
Summary:
arc diff on large mercurial repos was taking 14 seconds just to get
to the commit message prompt. With these optimizations it takes 4.
- "ancestor(.) - ancestor(XYZ)" is expensive because it has to build the
entire 400000+ revision history for both. "XYZ::. - XYZ" is much cheaper
because it only looks at the revisions between XYZ and the working directory.
- "hg outgoing" has to talk to the server, which is slow. "hg log -r draft()"
gives us the same information and is much cheaper. We fall back to 'outgoing'
on older versions of mercurial.
Of the remaining 4 seconds, 2.5 are spent in 'hg status', which is a bit harder
improve.
Test Plan: Ran arc diff on our hg repo. Verified it ran faster and the diff was created.
Reviewers: epriestley, sid0, bos, dschleimer
Reviewed By: epriestley
CC: aran, Korvin, tido
Differential Revision: https://secure.phabricator.com/D4838
Summary:
Record author email information in `arc diff`, so we can recreate it in `arc patch` and elsewhere without creating any kind of email exposure issues.
In Mercurial, we currently store the whole string ("username <email@domain.com>"). Make this consistent with Git.
Test Plan: Created git and hg diffs, saw authorEmail populated.
Reviewers: btrahan, edward, vrana
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D4825
Summary:
We assume `git` is available now, but should not. Specifically, if a user runs a working copy operation like `arc list` in an SVN working copy without `git` available, they get this error: P707
We interpret git errors very narrowly; be more liberal in how we interpret them. This assumes users working with `git` will have a functional `git`, but this seems like a reasonable assumption and lets us remove some error text matching code.
Test Plan: Changed `git` to `girt`, ran `arc list`, saw a reasonable exception. Changed back to `git`, saw git detected.
Reviewers: btrahan, vrana
Reviewed By: vrana
CC: svemir, aran
Differential Revision: https://secure.phabricator.com/D4804
Summary: Fixes T2438. We currently escape everything with '@', but SVN rejects that for '.'
Test Plan:
Unit tests. Performed this commit:
$ svn st
M .
A x@123
$ arc commit --conduit-uri=http://local.aphront.com:8080/ --revision 53
Revision 'D53: asdf' has not been accepted. Commit this revision anyway?
[y/N] y
Committing 'D53: asdf'...
Sending .
Adding x@123
Transmitting file data .
Committed revision 37.
Done.
I grepped for more '@' adding but couldn't find any. It's a bit tricky to grep for though, so it's possible I missed some.
Reviewers: btrahan, vrana
Reviewed By: btrahan
CC: aran, mbishopim3
Maniphest Tasks: T2438
Differential Revision: https://secure.phabricator.com/D4703
Test Plan: Deleted file in Git, ran `arc diff`, confirmed the question, saw the file as deleted.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D4603
Summary:
A git submodule looks a lot like a normal git repo, but the .git
directory is replaced with a file that git reads to find the real
location of the git directory. When arcanist tries to write a file into
a directory inside of there, it was failing silently, and then crashing
silently when it couldn't read results back out. Instead of assuming the
git directory is a directory named .git at the toplevel of the tree, we
use the appropriate git command to get the correct git directory.
Test Plan: submit a diff from a submodule
Reviewers: epriestley, vrana
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D4482
Summary:
D4186 added an "svn status --xml x y" form to getSVNStatus(), but the parser doesn't work for multiple files, since we get multiple <target /> elements in the XML output. So, curently, `arc diff` works (one target, all files) and `arc diff x` works (one target, x) but `arc diff x y` does not (more than one target, hits the exception).
$ arc diff QUACK2 QUACK3
Exception
Expected exactly one XML status target.
Test Plan: Ran `arc diff QUACK2 QUACK3` in a working copy with modified QUACK2, QUACK3. Ran `arc diff`; `arc diff QUACK2`.
Reviewers: vrana, btrahan, codeblock, JThramer
Reviewed By: codeblock
CC: aran
Differential Revision: https://secure.phabricator.com/D4372
Summary: If you are explicit then there is no need to ask you.
Test Plan:
$ touch a
$ arc diff
$ arc diff a
$ arc diff existing
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D4186
Summary:
See D4049, D4096.
- Move commit range storage from Mercurial and Git APIs to the base API.
- Move caching up to the base level.
- Store symbolic name and resolved name separately, so we can re-resolve the correct commit from the symbolic name after dirtying caches.
- Rename `supportsRelativeLocalCommit()` to `supportsCommitRanges()` (old name wasn't very good, and not consistent with new terminology like the `--base` flag).
- Rename `getRelativeCommit()` and `setRelativeCommit()` to `getBaseCommit()` and `setBaseCommit()`.
- Introduce `reloadCommitRange()` and call it from `reloadWorkingCopy()`.
I think this fixes the problem in D4049, and provides a general solution for the class of problems we're running into here, with D4096. Specifically:
- We no longer get dirty caches, as long as you call reloadWorkingCopy() after changing the working copy (or call a method which calls it for you).
- We no longer get order-of-parsing-things problems, because setBaseCommit() reloads the appropriate caches.
- We no longer get nasty effects from calling `requireCleanWorkingCopy()` too early.
Test Plan: This is pretty far-reaching and hard to test. Unit tests; ran various arc commands. :/
Reviewers: vrana
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D4097
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 discussion in D4049.
The getWorkingCopyStatus() method gets called from requireCleanWorkingCopy() in a lot of places, which triggers resolution of the base of the commit range. This is unnecessary; we do not need to examine the base commit in order to determine whether the working copy is dirty or not. This causes problems, in D4049 and elsewhere (we currently have a lot of fluff calls to setDefaultBaseCommit() in workflows which need to call requireCleanWorkingCopy() but do not ever use commit ranges, such as `arc patch`). This is mostly an artifact of SVN, where the "commit range" and "uncommitted stuff in the working copy" are always the same.
- Split the method into two status methods: getUncommittedStatus() (uncommitted stuff in the working copy, required by requireCleanWorkingCopy()) and getCommitRangeStatus() (committed stuff in the commit range).
- Lift caching out of the implementations into the base class.
- Dirty the cache after we commit.
This doesn't do anything useful on its own and creates one caching problem (`commitRangeStatusCache` is not invalidated when the commit range changes because of `setBaseCommit()` or similar) but I wanted to break things apart here. I won't land it until there's a more complete picture.
This creates a minor performance regression in git and hg (we run less stuff in parallel than previously) but all the commands should be disk-bound anyway and the regression should be minor. It prevents a larger regression in `hg` in D4049, and lets us do less work to arrive at common error states (dirty working copy). We can examine perf at the end of this change sequence.
Test Plan: Ran unit tests, various `arc` commands.
Reviewers: vrana
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D4095
Summary:
Makes arc land support hg repositories. Both bookmarks and named branches can be landed. For the most part all the arc land options work, but there are a few caveats:
- bookmarks can only be landed on bookmarks
- branches can only be landed on branches
- landing a named branch with --merge creates a commit to close the branch before the merge.
- since mercurial doesn't start with a default master bookmark, landing a bookmark requires specifying --onto or setting arc.land.onto.default
Test Plan:
Tested arc land with all permutations of --merge, --keep-branch on both bookmark branches and named branches. Also tested --hold, --revision, --onto, --remote.
See https://secure.phabricator.com/P619
Also tested git arc land with --merge and --keep-branch.
Reviewers: dschleimer, sid0, epriestley, bos
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D4068
Summary: I plan to use it in `save_lint.php`.
Test Plan:
$api->getUnderlyingWorkingCopyRevision(); // In Git SVN repo
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D4054
Summary: `svn info` is the slowest command for discovering repository (up to 300 ms) and Subversion is probably the least used repository type with Arcanist. Let's discover it last.
Test Plan:
$ arc lint --trace
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D4039
Summary: Also delete extra newlines.
Test Plan:
$ arc diff # on top of my commit
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T2025
Differential Revision: https://secure.phabricator.com/D3996
Summary:
There's quite some logic in here:
- It automatically decides whether to create a new commit or amend.
- It partially respects 'default-relative-commit'.
- However if it points to a closed revision then it creates a new commit.
Resolves T2025.
Test Plan:
`arc diff` on:
- Clean committed repository.
- Dirty repository without commit since 'default-relative-commit'.
- Dirty repository with non-revision commit since 'default-relative-commit'.
- Dirty repository with revision commit since 'default-relative-commit'.
- `arc diff HEAD^` on dirty repository on top of closed revision.
Reviewers: epriestley, btrahan
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T2025
Differential Revision: https://secure.phabricator.com/D3967
Summary:
accommodate git's diff.suppress-blank-empty=true setting
Without this change, if you were to set diff.suppress-blank-empty=true
in your .gitconfig (as I do), then "arc diff" would always fail with the
cryptic diagnostic, "Diff Parse Exception: Found the wrong number of
hunk lines."
Test Plan:
Put this in ~/.gitconfig or .git/config
[diff]
suppress-blank-empty = true
and run "arc lint". It should pass. Without this chnage,
it would fail as described above.
Reviewers: vrana, epriestley
Reviewed By: vrana
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D3963
Summary:
Run `hg status` as a Future while getting the diff. Add a cache for
getRawDiffText().
Test Plan:
Run `arc lint --trace` on a HG repo and confirm that `diff` is only called once
and `status` is run in the background.
Reviewers: epriestley
Reviewed By: epriestley
CC: yliang, dpepper, aran, Korvin
Maniphest Tasks: T2016
Differential Revision: https://secure.phabricator.com/D3913
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:
This is horrible and git specific, but fixes a case where people are using "arc
patch --nobranch ..." when they're not currently on a branch.
The old code assumed you were on a branch and used getBranchName() to record
this, in order to return to that branch later and cherry-pick the patch.
When not on a branch, and using arc patch --nobranch, this was trying to return
to the branch '(no branch)'.
Now, I detect that we're not on a branch and just record what HEAD is instead.
Test Plan:
Checkout the SHA of master (so I'm on master, but not on a branch) then try to
patch it with a feature diff:
€ git checkout e7a3ec68159d6847372cab5ad913f2f15aa7c249
Warning: you are leaving 1 commit behind, not connected to
any of your branches:
ac1ad39 Updating a-file
If you want to keep them by creating a new branch, this may be a good time
to do so with:
git branch new_branch_name ac1ad392350a51edd10343f12b9713f5e5b3707c
HEAD is now at e7a3ec6... Fix arcconfig
€ arc patch --nobranch D7
Created and checked out branch arcpatch-D7.
OKAY Successfully committed patch.
€ git branch
* (no branch)
feature
haddock
master
€ git log --oneline | head -2
38c0235 Updating a-file
e7a3ec6 Fix arcconfig
Reviewers: vrana, epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D3743
Summary: `arc patch` currently warns about "This diff is against commit svn+ssh://... but the commit is nowhere in the working copy" in Git SVN repository.
Test Plan:
$ arc patch # in Git SVN repository
$ svn find-rev r121212112 # invalid revision
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D3539
Git commit messages must have 2 newlines between the subject and body
If used to rewrite a commit message then the commit message will be
incorrectly reformatted.
Summary:
This reduces time of `arc branch` from 13s to 7s in my repo which is faster than `git branch --verbose` (10s).
The price for this speedup is that we loose the information [ahead 1, behind 21242] but we showed it only in branches with no revision so it's not a big deal.
Test Plan:
$ arc branch
Reviewers: epriestley
Reviewed By: epriestley
CC: ahupp, aran, Korvin
Differential Revision: https://secure.phabricator.com/D3005
Summary:
- In "arc which", we recommend "--rev x --rev ." to show changes. This is not accurate if there are uncommitted changes in the working copy. Just "--rev x" shows the correct changes (implicitly, the other end of the range is the working copy state).
- When you diff only working copy changes, we currently incorrectly identify all your open revisions as belonging to the working copy. Instead, correctly identify none of them as belonging to the working copy (in theory, we could go farther than this and do path-based identification like SVN, but with --amend in hg 2.2+ this workflow should be going away in the long run).
- If you have uncommitted working copy changes, never try to amend.
Test Plan: Ran "arc which .", "arc diff ." in a working copy with dirty changes, got better results than before.
Reviewers: dschleimer, btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1507
Differential Revision: https://secure.phabricator.com/D2980
Summary: From "cmd.exe" with, e.g. SilkSVN, there are some issues getting arc to do anything useful. Resolve enough of them so that it's at least usable.
Test Plan: Created a revision from Windows / cmd.exe / arc / SVN.
Reviewers: btrahan, jungejason, vrana
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D2984
Summary: Otherwise svn diff fails when the file is binary but the "svn:mime-type = application/x-shellscript" is missing in it.
Test Plan: arc diff succeeded against deleting a binary file which doesn't have svn:mime-type = application/x-shellscript.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D2915
Summary:
- Implement "arc:amended", a base commit DSL rule which always selects HEAD (git) or `.` (hg) if it has "differential revision:" in the commit message. This is unambiguously correct in amend workflows, and can cover holes in other rules like "git:branch-unique(*)".
- Fix a bunch of Mercurial stuff:
- Our use of '.' is wrong, and based on a misunderstanding on my part of the behavior of `hg diff --rev . --rev .`, which means "ignore the second --rev flag", not ". means working directory state". As far as I know there's no explicit way to say "the working copy plus all its changes".
- The `--prune` argument to "hg log" does not support symbolic names like ".^". Use revsets instead.
- Reduce the number of times we need to run `hg branch`.
- We can safely use "." to mean "the working copy revision", and do not need to do "hg --debug id" or similar.
- Generally simplify some of the nonsense in the implementation left over from me having no idea how Mercurial works.
Test Plan:
Ran "arc which" in various scenarios in a mercurial working copy. I //think// I exercised all the changes.
Ran "arc which --base arc:amended" in hg and git working copies without "Differential Revision:" in head/. (no match) and with it (matched head/.).
Reviewers: dschleimer
Reviewed By: dschleimer
CC: Makinde, tido, phleet, aran
Maniphest Tasks: T1233
Differential Revision: https://secure.phabricator.com/D2876
Summary:
This increases the amount of information arc diff collects in
Mercurial repositories. IN particular, it collects the active
bookmark, if there is one, and svn information in hgsubversion
repositories.
The Phabricator half of this is https://secure.phabricator.com/D2897
Test Plan:
[14:06:59 Sat Jun 30 2012] dschleimer@dev4022.snc6 ~/www-hg
www-hg 2941 $ ~/devtools/arcanist/bin/arc diff --only --conduit-uri http://phabricator.dschleimer.dev4022.facebook.com
HipHop Notice: Undefined index: 1 in /data/users/dschleimer/devtools/arcanist/src/repository/api/ArcanistMercurialAPI.php on line 708
Created a new Differential diff:
Diff URI: http://phabricator.dschleimer.dev4022.facebook.com/differential/diff/126/
Included changes:
A foo
HipHop Notice: Undefined index: 1 in /data/users/dschleimer/www-hg/lib/arcanist/arcanist/FacebookArcanistConfiguration.php on line 81
mcproxy on this server is out of date
version: , expected version:
please restart (http://fburl.com/1787362)
[14:07:46 Sat Jun 30 2012] dschleimer@dev4022.snc6 ~/www-hg
www-hg 2941 $ echo '{"diff_id": 126}' | ~/devtools/arcanist/bin/arc call-conduit differential.getdiff --conduit-uri http://phabricator.dschleimer.dev4022.facebook.com | json_pretty | egrep -i 'bookmark|sourcecontrol'
"bookmark": "bar",
"sourceControlBaseRevision": "svn+ssh://tubbs/svnroot/tfb/trunk/www@583442",
"sourceControlPath": "svn+ssh://tubbs/svnroot/tfb/trunk/www",
"sourceControlSystem": "hg",
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T1331
Differential Revision: https://secure.phabricator.com/D2896
Summary:
Add a DSL command to select the first commit between HEAD and (the merge-base of some target with HEAD) that's on more than one branch.
This is similar to @csilvers' suggestion in <https://secure.phabricator.com/T1233#comment-8>. A specific problem this address is that, currently, if you use this workflow:
$ git checkout -b feature
$ git commit
$ git checkout -b subfeature
$ git commit
..i.e., develop dependent features in sub-branches, "arc diff" will try to send up (feature + subfeature). If you use the rule 'git:branch-unique(origin/master)' instead, diffing from 'subfeature' will correctly select only the commit(s) on 'subfeature'.
Test Plan: Constructed feature/subfeature branches, verified that we select the correct base commit.
Reviewers: dschleimer, csilvers, btrahan, jungejason
Reviewed By: dschleimer
CC: aran
Maniphest Tasks: T1233
Differential Revision: https://secure.phabricator.com/D2866
Summary:
This adds a new base option that Does the Right Thing (or as close to
it as I can get) for someone using Mercurial bookmarks as lightweight
branches, and where each branch should be one differential revision.
Specifically, it walks backwards through the ancestors of the working
copy until it finds a revision that is either not outgoing (ie already
in the remote) or that is bookmarked. This means that bookmarks
effectively act as delimiters, and point at the last revision that
will be included in an arc diff.
Test Plan:
[14:40:54 Tue Jun 26 2012] dschleimer@dev4022.snc6 ~/hg-dummy
hg-dummy 1226 $ hg log -r '(first(outgoing())^)::' --template='node: {node}\nbookmark: {bookmarks}\n\n' -G
@ node: c8379ef32b0d0e6cf94fe636751ea4fe1353e157
| bookmark:
|
| o node: 14f03139049cbda339190b814e52f4ec8b05c431
| | bookmark: more
| |
| o node: 6970e9263ab8c6da428420606d1f15c9980da183
| | bookmark: something
| |
| o node: 433a93023f03d5f3eddaa243fa973d32a1566aee
|/ bookmark:
|
o node: f47ccfe34267592dd2e336174a3a311b8783c024
| bookmark:
|
[14:41:00 Tue Jun 26 2012] dschleimer@dev4022.snc6 ~/hg-dummy
hg-dummy 1226 $ ~/devtools/arcanist/bin/arc which --show-base --base 'arc:bookmark'
f47ccfe34267592dd2e336174a3a311b8783c024
[14:41:05 Tue Jun 26 2012] dschleimer@dev4022.snc6 ~/hg-dummy
hg-dummy 1226 $ hg up 14f03139049cbda339190b814e52f4ec8b05c431
3 files updated, 0 files merged, 1 files removed, 0 files unresolved
[14:41:10 Tue Jun 26 2012] dschleimer@dev4022.snc6 ~/hg-dummy
hg-dummy 1226 $ ~/devtools/arcanist/bin/arc which --show-base --base 'arc:bookmark'
6970e9263ab8c6da428420606d1f15c9980da183
[14:41:14 Tue Jun 26 2012] dschleimer@dev4022.snc6 ~/hg-dummy
hg-dummy 1226 $ hg up 6970e9263ab8c6da428420606d1f15c9980da183
0 files updated, 0 files merged, 1 files removed, 0 files unresolved
[14:41:44 Tue Jun 26 2012] dschleimer@dev4022.snc6 ~/hg-dummy
hg-dummy 1227 $ ~/devtools/arcanist/bin/arc which --show-base --base 'arc:bookmark'
f47ccfe34267592dd2e336174a3a311b8783c024
Reviewers: epriestley, bos
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T1331
Differential Revision: https://secure.phabricator.com/D2863
Summary:
- Move "arc branch"-specific code to the branch workflow.
- Instead of doing "git rev-parse", just do "git branch --verbose --abbrev=40".
- Use revision owners to identify ownership, not working copy identity. Particularly with the advent of "Commandeer", you might not own commits you made.
- Do a batch lookup for commits by hash (depends on D2859, but doesn't break without it).
- Use PhutilConsole for console stuff.
- Removed color from "arc list" for the moment.
- The "--by-status" flag has a slightly different output format now.
Test Plan: Ran "arc branch" in various circumstances, verified it identifies branches in immutable history repositories.
Reviewers: btrahan, vrana, jungejason, nh, slawekbiel
Reviewed By: slawekbiel
CC: aran
Maniphest Tasks: T693
Differential Revision: https://secure.phabricator.com/D2860
Summary:
svn 1.7 got rid of the per-direcotry .svn dirs in favor of a single
.svn directory under the root of the working copy. Unfortunately, we
relied on ther being a .svn directory at the same level as .arcconfig,
which may or may not be at the root of the working copy. We now walk
up the directory tree until we find a .svn directory that we can use
for scratch files.
Test Plan:
[16:27:52 Mon Jun 25 2012] dschleimer@dev4022.snc6 ~/data/admin/facebook/scripts/db
db 21298 $ ls -la .arcconfig .svn ../../../.svn/arc/config
ls: .svn: No such file or directory
ls: ../../../.svn/arc/config: No such file or directory
-rw-r--r-- 1 dschleimer users 239 Jun 25 16:15 .arcconfig
[16:27:54 Mon Jun 25 2012] dschleimer@dev4022.snc6 ~/data/admin/facebook/scripts/db
db 21298 $ ~/devtools/arcanist/bin/arc set-config --local foo bar
Set key 'foo' = 'bar' in local config.
[16:29:40 Mon Jun 25 2012] dschleimer@dev4022.snc6 ~/data/admin/facebook/scripts/db
db 21298 $ ls -la .arcconfig .svn ../../../.svn/arc/config
ls: .svn: No such file or directory
-rw-r--r-- 1 dschleimer users 239 Jun 25 16:15 .arcconfig
-rw-r--r-- 1 dschleimer users 20 Jun 25 16:29 ../../../.svn/arc/config
[16:29:43 Mon Jun 25 2012] dschleimer@dev4022.snc6 ~/data/admin/facebook/scripts/db
db 21298 $ cat ../../../.svn/arc/config
{
"foo" : "bar"
}
[16:29:48 Mon Jun 25 2012] dschleimer@dev4022.snc6 ~/data/admin/facebook/scripts/db
db 21299 $ ~/devtools/arcanist/bin/arc get-config foo
(global) foo =
(local) foo = bar
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T1411
Differential Revision: https://secure.phabricator.com/D2856
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
Summary:
We currently use the language "relative commit" or "relative local commit" to refer to the head of a commit range. I want to move toward callign this a "base commit", since I think that makes more sense. This moves us a small step in that direction.
The DSL stuff in T1233 also needs access to this stuff, so move it up to the base. This is mostly a bite-sized piece of the change in that diff.
Test Plan: Ran "arc which" in a couple of circumstances, read explanations.
Reviewers: dschleimer, vrana
Reviewed By: dschleimer
CC: aran
Maniphest Tasks: T1233
Differential Revision: https://secure.phabricator.com/D2747
Summary:
Using .arc as the scratch and per-repository configuration directory
has some unfortunate consequenses in the real world. Among other
things, people forget to .gitignore it so it gets checked in.
Test Plan:
the only thing that seems to use this is the relative commit setting
for git. This diff consists of 2 commits, one for the .gitignore and
one for everything else.
Comment out the portion of my .git/config that defines the upstream
for the branch. Run arc diff --only with HEAD^ in
.arc/default-relative-commit. See that .gitignore is not included in
the resultant diff, that .arc no longer exists, and that .git/arc
exists and has HEAD^ in .git/arc/default-relative-commit.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T1233
Differential Revision: https://secure.phabricator.com/D2725
Summary:
This adds basic support for mutable history for arc diff and arc amend for
mercurial. This is purely opt-in (so it shouldn't affect anyone who doesn't want
this feature) by explicitly setting
"immutable_history" : false
in arc configuration.
This also fixes another instance of weird behaviour for multiple heads - the
first instance was fixed here:
https://github.com/facebook/arcanist/pull/28
Test Plan:
without "immutable_history" turned on)>
When ##arc diff## produces an update diff, it should list only commits that are
ancestors of the current revision - not ones from other heads.
Reviewers: epriestley
CC: csilvers, aran, Koolvin
Differential Revision: https://secure.phabricator.com/D2654
Summary: Mechanical changes from D2588. No "Class.php" moves yet because they aren't necessary for libraries to function.
Test Plan: See D2588.
Reviewers: vrana, btrahan, jungejason
Reviewed By: vrana
CC: aran
Maniphest Tasks: T1103
Differential Revision: https://secure.phabricator.com/D2589
Summary:
- When you run "arc diff" in Mercurial, we currently give you an empty template. Instead, prefill a likely template by parsing messages, as we do for Git.
- Unify Git and Mercurial logic for acquiring messages, since `getLocalCommitInformation()` now provides this information. This should improve Git performance, and allows us to delete some code.
- Merge messages more intelligently. Currently, we just overwrite fields. Instead, merge arrays (like ccs, reviewers, tasks) and concatenate strings (like summary and test plan). We need to special case "title", see inline.
- @csilvers: this is a precursor to implementing "--reviewers", etc.
Test Plan: See next post, test plan includes giant output.
Reviewers: btrahan, csilvers, Makinde, jungejason, vrana
Reviewed By: csilvers
CC: aran
Differential Revision: https://secure.phabricator.com/D2536
Summary:
- "git log" still includes "\n", so we're currently generating nonsense hashes like "\nabd9879bab86ad78ab...".
- Correct the log range we use in Git. See comment. When users perform merges, their expectations about what the "included commits" and what the included changes are are different. Represent them with two different ranges.
- Same deal for Mercurial
Test Plan: Ran "arc which" in various contexts.
Reviewers: btrahan, aurelijus, Makinde
Reviewed By: Makinde
CC: aran, nh, jungejason
Maniphest Tasks: T873
Differential Revision: https://secure.phabricator.com/D2460
Summary: Provide far more information about what "arc diff" intends to do.
Test Plan: Ran "arc which" in a variety of circumstances.
Reviewers: btrahan, Makinde
Reviewed By: Makinde
CC: aran
Maniphest Tasks: T1183
Differential Revision: https://secure.phabricator.com/D2454
Summary: Currently, we ship only the summary, but we need to ship the whole thing for T1189.
Test Plan: Added var_dump() + die, ran in git and hg working copies, verified 'message' included the whole message.
Reviewers: csilvers, btrahan, Makinde
Reviewed By: csilvers
CC: aran
Maniphest Tasks: T1189
Differential Revision: https://secure.phabricator.com/D2437
Summary:
This is mostly an onboarding thing, but also allows "arc upload", "arc download", and "arc paste" to work anywhere on the system.
- Try to read the Phabricator install URI from arc global config if we can't find ".arcconfig".
- Build a WorkingCopy anyway if we can't find ".arcconfig", as long as we can find ".svn", ".git", or ".hg".
- Make all the workflows handle "no project ID" at least somewhat gracefully.
Test Plan:
- Ran "arc diff" in .arcconfig-less Mercurial, Git, and Subversion working copies.
- Ran "arc upload" and "arc download" from my desktop.
- Ran "arc paste" from somewhere random.
- Cleared my config and hit the error, got useful instructions.
Reviewers: btrahan, csilvers
Reviewed By: csilvers
CC: aran
Differential Revision: https://secure.phabricator.com/D2424
Summary:
Always use the value in git.default-relative-commit when getting the relative
commit in git.
Test Plan:
Ran arc diff in a repo with git.default-relative-commit set to HEAD^ on a
branch tracking a remote (that is different from HEAD^), and checked that
the diff against HEAD^, not the remote, was published.
Reviewers: jungejason, epriestley, vrana
Reviewed By: epriestley
CC: aran, Koolvin
Differential Revision: https://secure.phabricator.com/D2409
Summary:
- We no longer need color options since we fake our way through parsing ANSI colorized diffs and use HGPLAIN (on Windows, too!). Drop 'em.
- In the case where you have nothing outgoing, we don't cache the relative commit and thus run "hg outgoing" too many times, which is fairly slow (even if you have nothing outgoing). Cache it.
Test Plan: Ran "arc diff --trace" in a mercurial working copy with nothing outgoing; verified we run "hg outgoing" only once.
Reviewers: Makinde, csilvers, btrahan
Reviewed By: Makinde
CC: aran
Differential Revision: https://secure.phabricator.com/D2399
Summary: In Windows, you can't use `X=y cmd` syntax to set variables. Use "set X=y & cmd" instead.
Test Plan:
- Ran "arc diff" in a Mercurial repo in Windows, created D2367.
- Verified this does //not// cause 'HGPLAIN' to be set in the outer shell (where you type "arc diff").
Reviewers: Makinde, tido, indiefan, btrahan
Reviewed By: tido
CC: aran
Maniphest Tasks: T1179
Differential Revision: https://secure.phabricator.com/D2368
Test Plan: Run arc diff locally, verify via git log that the commit is amended afterwards (using the mutable history paradigm)
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Koolvin
Differential Revision: https://secure.phabricator.com/D2369
Summary: Apparently my advice here was terrible, and `--no-decorate` and `--decorate=no` are both very recent additions to Git which a bunch of users don't have. Get rid of them since D2344 allows us to parse all decorate levels anyway.
Test Plan: Tried to google "git changelog", got a bunch of pages about managing changelogs with git.
Reviewers: zeeg, ehren
Reviewed By: ehren
CC: ehren, aran, NorthIsUp
Differential Revision: https://secure.phabricator.com/D2354
Summary:
"git diff -M -C" generates useful metadata (moves/copies) but (for a pure move) no diff text. Synthetically build the diff text after the fact so this information is available in Differential.
This patch is kind of nasty but I couldn't see a cleaner way to do it. :/
This also needs some UI changes in Differential: we get a full-green new file right now, but it would be better to default-hide it with "This file was moved. Show More" or similar.
Test Plan: Moved a file, ran "arc diff", got textual diff.
Reviewers: aran, tuomaspelkonen, jungejason, btrahan, vrana
Reviewed By: vrana
CC: aran, epriestley, vrana
Maniphest Tasks: T230
Differential Revision: https://secure.phabricator.com/D479
Summary:
Arcanist fails to find git's 'commit <hash>' header when log.decorate is
set in one of git's config files. git adds the named refs that point to
the commit in parentheses following the hash. This changes the regex
used by arcanist to match git commit headers to optionally match the
branch names in parentheses following the hash.
Test Plan:
Run `git config --global log.decorate short` and check that `arc diff`
runs successfully.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Koolvin
Differential Revision: https://secure.phabricator.com/D2344
Summary:
- Add branch name tab completion to "arc land".
- Default to landing the current branch.
- This is a little bit hacky but not too terrible. I'm planning to move the whole thing to PhutilArgumentParser at some point so that'll be an opportunity for a big refactor.
Test Plan: Hit tab, landed this branch.
Reviewers: zeeg, btrahan, vrana, jungejason
Reviewed By: btrahan
CC: aran, kdeggelman
Differential Revision: https://secure.phabricator.com/D2293
Summary: See rage in T1117. Don't use the <project + branch> heuristic anymore..
Test Plan: Ran "arc diff --strict HEAD^" on a commit stacked on top of this one, got no matches.
Reviewers: btrahan, vrana, simpkins, beng
Reviewed By: btrahan
CC: aran, avive
Maniphest Tasks: T1117
Differential Revision: https://secure.phabricator.com/D2221
Summary:
- Historically, "--preview" was forbidden under SVN. No reason for that now.
- The "--auto" patch moved the "--preview" / "--only" checks later than they should be.
- Fix an issue with Conduit query construction in SVN.
Test Plan: Ran "arc diff --preview" in an SVN working copy. Ran "arc diff" in an SVN working copy.
Reviewers: svemir, btrahan, vrana, jungejason
Reviewed By: svemir
CC: aran
Differential Revision: https://secure.phabricator.com/D2218
Summary:
When people create the .arc/default-relative-commit scratchfile with $EDITOR of
choice, their editor usually puts a newline at the end, which breaks arc diff.
We should trim the newline before using the contents of the scratchfile.
Test Plan:
ran arc diff in a working copy that contained a .arc/default-relative-commit
with a newline
Reviewers: epriestley, jungejason
Reviewed By: epriestley
CC: aran
Differential Revision: https://secure.phabricator.com/D2209
Summary: If you have two projects (say, libphutil and arcanist) and you prepare a patch for one of them on branch "master", run "arc diff", and then prepare a patch for the other one on the same branch, "arc diff" will try to update the first revision when you run it. Instead, make it smart enough to stay within arc projects.
Test Plan: Ran "arc which" in circumstances where it previously generated a false positive, no false positive.
Reviewers: btrahan, vrana, jungejason
Reviewed By: jungejason
CC: aran
Maniphest Tasks: T1100
Differential Revision: https://secure.phabricator.com/D2199
Summary: See discussion in D1861.
Test Plan: Ran "arc diff" on master, got an upstream-based relative commit. Ran "arc diff" on a feature branch, got a config-based relative commit. Ran "arc diff x", got an argument-based relative commit.
Reviewers: btrahan, vrana, davidreuss, elgenie
Reviewed By: davidreuss
CC: aran
Differential Revision: https://secure.phabricator.com/D2192
Summary: Allows you to set a default in .arcconfig. This default is overriden by any .arc/ setting.
Test Plan: Ran "arc diff" with a .arcconfig setting but no .arc/ setting, got a diff against the specified relative commit.
Reviewers: nh, btrahan
Reviewed By: nh
CC: aran
Differential Revision: https://secure.phabricator.com/D2093
Summary:
The default of "arc diff" to "arc diff HEAD^" in git is universally confusing to everyone not at Facebook.
Drive the default with configuration instead. Even at Facebook, "origin/trunk" (or whatever) is probably a better default than "HEAD^".
See D863 for the last attempt at this.
NOTE: This is contentious!
Test Plan: Ran "arc diff", got prompted to set a default. Ran "arc diff" from a zero-commit repo, got sensible behavior
Reviewers: btrahan, vrana, nh, jungejason, tuomaspelkonen
Reviewed By: btrahan
CC: aran, epriestley, zeeg, davidreuss
Maniphest Tasks: T651
Differential Revision: https://secure.phabricator.com/D1861
Summary:
In Mercurial, we figure out if the working copy is dirty with "hg diff", and do a somewhat expensive merge-base / outgoing operation if the relative commit isn't set. We can set the relative commit earlier and avoid some extra work.
We also may incorrectly cache this state (diff from merge-base of outgoing to tip) and pass the wrong rev and file dirty list to the linters.
Test Plan: Made commits which changed (A, B) and then (A). Ran "arc diff tip^". Before this change, observed full outgoing + merge base resolve and both "A" and "B" passed to lint. Observed execution of fewer commands and lint executing against "A" only after this change.
Reviewers: Makinde, btrahan
Reviewed By: Makinde
CC: aran, epriestley
Differential Revision: https://secure.phabricator.com/D1998
Summary: Obvious error in refactoring code around getCanonicalRevisionName() in D1954.
Test Plan: Ran "arc diff" without looping. Can you verify this fixes your case?
Reviewers: Makinde, btrahan
Reviewed By: Makinde
CC: aran, epriestley
Maniphest Tasks: T1025
Differential Revision: https://secure.phabricator.com/D1970
Summary:
See task. Allow mercurial users to diff with uncommitted changes.
- By default, commit range is merge-base of `hg outgoing` to `.` (dirstate).
- You can get JUST dirstate with `arc diff tip` or similar.
- This ended up being a giant mess various other changes to deal with empty `hg outgoing` and empty dirstate.
Test Plan: Diffed with uncommitted changes, got sensible prompts and results.
Reviewers: Makinde, btrahan
Reviewed By: btrahan
CC: aran, epriestley
Maniphest Tasks: T998
Differential Revision: https://secure.phabricator.com/D1954
Summary:
In some workflows, we don't have Conduit at all, so we'll fail with a raw exception. Test for conduit presence before trying to make the encoding call.
Also move some "instanceof" logic for updates into RepositoryAPI (factoring, windows compat).
Test Plan: Ran "arc patch --patch some.patch".
Reviewers: 20after4, davidreuss, nh, btrahan
Reviewed By: 20after4
CC: aran, epriestley
Differential Revision: https://secure.phabricator.com/D1935
Summary:
`svn rm $directory` removes the directory and its contents in svn 1.7, whereas
svn 1.6 only removes the directory's contents, so arc diff needs to not try
to read the directory.
Test Plan:
ran arc diff in a svn working copy that had a directory removed (with both
svn 1.6 and svn 1.7) and confirmed that the command did not throw an exception,
and that the removed directory was marked as directory (not a file).
Reviewers: epriestley, jungejason
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: https://secure.phabricator.com/D1889
Summary:
Resolves crazy stuff like 'HEAD' or '{2001-01-01T01:01:01}' into an
actual revision, with a provided implementation for git.
Test Plan: Tested in a hacky script.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: https://secure.phabricator.com/D1849
Summary:
svn 1.7 no longer has a .svn directory in every subdirectory of the working
copy, only one in the root of the working copy (like git, except you can still
check out subtrees). Thus, we can't check whether we're in an svn repo by
looking for a .svn directory alongside the .arcconfig file.
Test Plan: ran arc diff in an svn 1.7 working copy where it previously wasn't working
Reviewers: epriestley, btrahan, jungejason
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: https://secure.phabricator.com/D1848
Summary: good title
Test Plan: ran "arc patch DX" a bunch
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, epriestley
Maniphest Tasks: T892
Differential Revision: https://secure.phabricator.com/D1789
Summary:
On Windows, the PHP function escapeshellarg() replaces '%' with ' ' (space). This is apparently because there is no safe way to escape % inside of strings.
cmd.exe does use "^" as an escape character, so I think replacing "xyz" with "^x^y^z" might work for arbitrary strings (maybe?), or at least some subset of strings, but I don't know cmd.exe well enough to make that call without being concerned I'm introducing a security issue.
Although this patch is dumb, it's certinaly safe, and can only do something wrong if the user has environmental variables like H, P, T, or x01, in which case they're sort of asking for it.
cmd.exe also truncates output on \0, so use \1 as a delimiter instead.
Seriously it's like this thing was written in 1982 and never ever changed.
Test Plan: Created D1783 successfully after applying this patch.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran, epriestley
Maniphest Tasks: T124
Differential Revision: https://secure.phabricator.com/D1785
Summary: This use of "(cd ..)" is outside of the RepositoryAPI proper (it's when we're trying to figure out which VCS a directory uses) and explodes on Windows.
Test Plan: @koolvin applied this patch manually and got farther than before.
Reviewers: btrahan, Koolvin, jungejason
Reviewed By: jungejason
CC: aran, epriestley
Maniphest Tasks: T124
Differential Revision: https://secure.phabricator.com/D1779
Summary:
- This is simpler and reuses more code than doing "(cd %s && ...)" every time.
- If we have an issue like HGPLAIN, we have one place to fix it now.
- On Windows, the construct "(cd %s && ...)" does not mean what we'd like it to.
Test Plan: Ran "arc diff" in a git repo with this change.
Reviewers: btrahan, Makinde
Reviewed By: Makinde
CC: aran, epriestley
Maniphest Tasks: T124
Differential Revision: https://secure.phabricator.com/D1761
Summary:
If a user reverts HEAD in the index or working copy, "git diff" shows no changes
and we explode.
Instead, we should simply return no changes; the rest of the lint pipeline
accommodates this correctly.
Test Plan: Reverted HEAD in a working copy, ran "arc lint", got an error.
Applied patch, ran "arc lint", got accurate lint.
Reviewers: Koolvin, btrahan
Reviewed By: Koolvin
CC: aran, epriestley
Differential Revision: https://secure.phabricator.com/D1729
Summary:
When a user invokes the update flow from Mercurial, prefill the update message
like we do for Git.
Also add a little caching to improve performance since "hg" execs cost ~100ms.
Test Plan: Updated a Mercurial diff repeatedly, got sensible default message
fills.
Reviewers: Makinde, btrahan
Reviewed By: Makinde
CC: aran, epriestley
Maniphest Tasks: T28
Differential Revision: https://secure.phabricator.com/D1718
Summary: See comments. Skip [defaults] in .hgrc when executing commands.
Test Plan: Ran "arc diff" in a mercurial working copy.
Reviewers: Makinde, btrahan
Reviewed By: Makinde
CC: aran, epriestley
Maniphest Tasks: T922
Differential Revision: https://secure.phabricator.com/D1707
Summary:
Mark all applicable Arcanist classes as "final", except PhutilLintEngine, which
needs a little finesse.
@jungejason / @nh, does this break any Facebook stuff?
Test Plan: Linter no longer raises warnings. Ran "testEverythingImplemented" in
Phabricator.
Reviewers: btrahan, jungejason, nh
Reviewed By: btrahan
CC: aran, epriestley
Maniphest Tasks: T795
Differential Revision: https://secure.phabricator.com/D1519
ArcanistRepositoryAPI->loadWorkingCopyDifferentialRevisions()
Summary:
- See T787.
- @cpiro has an immediate use case for this, which is ##arc amend --revision
`arc which --id` --show master## for "git merge --autosquash" or similar.
- For T614, we need this to choose "--create" vs "--update".
- Other workflows should also use this to improve how often we automatically
get things right, particularly in Mercurial and SVN.
Test Plan:
Ran "arc which" in SVN, Git and HG working copies with various flags;
results seemed reasonable.
Reviewers: btrahan, cpiro, jungejason
Reviewed By: btrahan
CC: aran, epriestley
Maniphest Tasks: T787
Differential Revision: https://secure.phabricator.com/D1478
Summary:
pretty easy stuff as mercurial accepts git style patches...!
also fixed two issues where we were 1) storing the short hash and 2) storing it
with a trailing "\n". This diff makes us store the full hash AND no trailing
return character
Test Plan:
in my mercurial repo
<note repo at revision $foo>
<did something dumb>
hg commit -m "something dumb"
arc diff
<go to web and fill out stuff for DX>
hg checkout $foo
arc patch DX
<verify patch DX successfully applied!>
use conduit console to verify a few diffs were returning the correct full
revision hash with a trailing \n
Reviewers: epriestley
Reviewed By: epriestley
CC: aran
Maniphest Tasks: T630
Differential Revision: https://secure.phabricator.com/D1431
Summary:
Currently, we throw a fairly perplexing error when there are multiple valid
commit messages. Installs can also remove the "test plan" field entirely, which
is the only really strong discriminator here.
When the message to use is ambiguous, show the user all the valid messages and
prompt them to choose one.
Also add a -C flag like "git commit -C", so they can choose a message
explicitly.
Test Plan: Ran "arc diff HEAD^^^^^", "arc diff -C <rev>".
Reviewers: cpiro, btrahan, jungejason
Reviewed By: cpiro
CC: aran, cpiro
Differential Revision: https://secure.phabricator.com/D1385
revision is the correct base revision relative to the patch.
Summary: What the title says. If not correct, warn the user. This check
honors the --force flag to skip all these checks. This change also includes
moving some Differential constants into Arc so they can be used for both
projects. There is a corresponding phabricator diff (incoming) to address this
part of the change.
Test Plan:
For a project with actual diffs, a git repository tracked by phabricator, *AND*
development in master branch only, do some...
- git reset --hard HEAD^1
- arc patch DX, where X is what got us to HEAD in the first place
- verify successful patch
...then...
- git reset --hard HEAD^^
- arc patch DX, where X is what got us to HEAD in the first place
- verify warning
- verify Y versus N continues versus stops appropriately
Note if development were done outside the master branch this warning message
will fire early / often as git commit hashes are based on the commit *and* the
rest of the source code the commit is made against. This is (unfortunately) the
"typical" case so this warning is pretty active at the moment. T201 will
eventually land and when parsing a given commit update the corresponding diff.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, btrahan, epriestley
Differential Revision: https://secure.phabricator.com/D1328
Summary: Find the relative commit by finding the first non-outgoing commit, so
we don't show changes caused by merges we've performed since the last time we
pushed.
Test Plan: Checked out two hg working copies, A and B. Made a change in A. Made
a change in B. Pushed B. Merged in A. Made another change in A. Ran "arc diff"
in A. Got only changes I made in A in the diff, not the change from B.
Reviewers: Makinde, btrahan, jungejason
Reviewed By: btrahan
CC: aran, btrahan
Differential Revision: https://secure.phabricator.com/D1339
Summary: ... in case the head commit is empty. Empty commits are useful for
injecting an Arcanist commit message in a branch then sending the whole thing
off for review. As far as I know there is no situation in which an empty commit
would exist unintentionally and using ##--allow-empty## would suppress an error.
Test Plan: works fine for a branch I just sent for review
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: 1241
Summary:
Git works completely differently for commits zero and one than for 2..N so add
more special casing to handle them. See:
- {T206}
- {T596}
The getCommitRange() block is also fatal land, although I wasn't able to reach
it. I'll follow up with @s on T596.
Test Plan:
- Created a new, empty repository ("mkdir x; cd x; git init").
- Ran "arc lint", "arc unit", "arc diff" against it with no commits (the first
two work, the third fails helpfully).
- Made an initial commit.
- Ran "arc lint", "arc unit", "arc diff" against it (all work correctly).
Reviewers: btrahan, jungejason, aran
Reviewed By: aran
CC: s, aran
Differential Revision: 1142
Summary:
In D1079, I added "--color never", but this flag is provided by the "color"
extension, which is why I missed it originally, because it doesn't show up until
you enable that extension. Providing it causes installs which don't have it
enabled (disabled is the default) to fail.
Use "--config" to disable color instead. This sets a configuration setting and
works regardless of whether the color extension is present.
Test Plan: Ran "arc diff" in a mercurial working copy with the color extension
enabled and disabled.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: nh
CC: aran, nh
Differential Revision: 1092
Summary: We ostensibly now support Mercurial at least mostly, so don't
underpromise quite so much.
Test Plan: Ran "arc diff" in a Mercurial working copy without being warned.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 1085
Summary: Mercurial has a --color flag which disables the use of ANSI colors. Use
it to prevent "hg diff" from giving us colorized diff output.
Test Plan: Ran "hg diff --color never", verified it disabled ANSI color in diff
output.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: aran
CC: aran
Differential Revision: 1079
Summary:
D1061 introduced a 'text file' check, but it fails under SVN for new
directories.
- Revert D1061 (This reverts commit b2cd18252701be2093b52652fb3d1d94c5df571e.)
- Make getChangedLines() return null to indicate that the operation doesn't
make sense. I think this was the intent of the code in the lint engine.
- Fix a bug where running "arc lint" on a change in an SVN working copy from a
subdirectory would fail.
- Fix a bug where warnings with no line information were incorrectly
discarded.
Test Plan:
- Ran "arc lint" in an SVN working copy with a new directory (no failure).
- Forced FilenameLinter to always raise a warning. Added a binary file and ran
"arc lint". The warning was reported for the new binary file, a new text file,
and a new directory.
Reviewers: jungejason, andrewjcg, nh, tuomaspelkonen, aran
Reviewed By: andrewjcg
CC: aran, andrewjcg, epriestley
Differential Revision: 1076
Summary: Apparently I just never tested this or something. Make it work
correctly.
Test Plan: Ran "arc diff" in a Mercurial working copy with a binary file in
outgoing.
Reviewers: Makinde
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 1074
Summary: We incorrectly add the "Uncommitted" flag to untracked files, which
causes them to raise various prompts and not respect "--allow-untracked".
Test Plan: Ran "arc diff --allow-untracked" in a clean working copy with
untracked files, was not fatally error'd.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde, epriestley
Differential Revision: 1073
Summary:
In D962, I switched us from using "hg summary" to "hg id" to find the working
copy revision. However, "hg id" reports the revision with a trailing "+" if the
working copy is dirty, so we fail before hitting the explicit check for this
later.
Trim off the trailing '+' if it is present.
Test Plan:
Ran "arc diff" in a dirty Mercurial working copy and got a good error message
about uncommitted changes.
~/repos/hg-working-copy $ arc diff
WARNING: Mercurial support is largely imaginary right now.
Usage Exception: You have uncommitted changes in this branch. Commit (or
revert) them before proceeding.
Working copy: /INSECURE/repos/hg-working-copy/
Uncommitted changes in working copy
hello.c
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 1070
Summary:
See D945. We have this kludgy "remote_hooks_installed" mess right now, but we
have enough information on the server nowadays to figure this out without it.
Also reduce code duplication by sharing the "mark-committed" workflow.
This causes "arc merge" to effect commit marks.
Test Plan:
In Git, Mercurial and SVN working copies ran like a million
amend/merge/commit/mark-committed commands with and without --finalize in
various states of revision completion.
This change is really hard to exhaustively test because of the number of
combinations of VCS, revision state, command, command flags, repository state
and tracking state. All the reasonable tests I could come up with worked
correctly, though.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 967
Summary:
Expand 12-character hashes to 40-character hashes so other things will work
properly.
Also use "hg id" instead of "hg summary" to figure out where the working copy
is, since it's substantially simpler.
Test Plan: Ran "arc diff" and got properly 40-character hashes.
Reviewers: Makinde, aran
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 962
Summary: I renamed this at some point but missed a callsite. See T519.
Test Plan: Ran "arc patch" in an SVN working copy.
Reviewers: Girish, jungejason
Reviewed By: Girish
CC: aran, Girish
Differential Revision: 959
Summary: Move code to actually parse "hg" output into a separate class with some
tests, so I can reuse it in the import scripts. We should probably do this for
Git/SVN at some point, too.
Test Plan: Ran unit tests, used this class in Phabricator importers, grepped for
calls to removed private methods.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 942
Summary:
This should support conservative rewrite policies in git fairly well, under an
assumed workflow of:
- Develop in local branches, never rewrite history.
- Commit with "-m" or by typing a brief, non-template commit message
describing the checkpoint.
- Provide rich information in the web console (reviewers, etc.)
- Finalize with "git checkout master && arc merge branch && git push" or some
flavor thereof.
This supports Mercurial somewhat. The major problem is that "hg merge" fails if
the local is a fastforward of the remote, at which point there's nowhere we can
throw the commit message. Oh well. Just push it and we'll do our best to link
them up based on local commit info.
I am increasingly forming an opinion that Mercurial is "saftey-scissors git".
But also maybe I have no clue what I'm doing. I just don't understand why anyone
would think it's a good idea to have a trunk consisting of ~50% known-broken
revisions, random checkpoint parts, whitespace changes, typo fixes, etc. If you
use git with branching you can avoid this by making a trunk out of merges or
with rebase/amend, but there seems to be no way to have "one commit = one idea"
in any real sense in Mercurial.
Test Plan: Execute "arc merge" in git and mercurial.
Reviewers: fratrik, Makinde, aran, jungejason, tuomaspelkonen
Reviewed By: Makinde
CC: aran, epriestley, Makinde
Differential Revision: 860
Summary:
We have some git-specific logic on main pathways which should be in the API
class, move it around so "arc lint" with an engine works under Mercurial. This
resovles the error @makinde reported:
> PHP Catchable fatal error: Argument 1 passed to
ArcanistBaseWorkflow::parseGitRelativeCommit() must be an instance of
ArcanistGitAPI, instance of ArcanistMercurialAPI given, called in
/home/makinde/.arc_install/arcanist/src/workflow/lint/ArcanistLintWorkflow.php
on line 131 and defined in
/home/makinde/.arc_install/arcanist/src/workflow/base/ArcanistBaseWorkflow.php
on line 830
Test Plan: Ran "arc diff" in git and hg working copies, plus "arc lint" with a
configured "lint_engine".
Reviewers: Makinde, aran, jungejason, nh, tuomaspelkonen
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 935
Summary:
- Use "hg parents" to figure out where outgoing changes originate from, not
"~1", since that's a new feature in Mercurial.
- Add "--style default" all over the place since hg config can override this
(similar to the date config issues we saw in git).
- Cache working copy status so we don't run full hg diffs like 30 times
(similar to git/svn APIs).
- Use full "--git" flag instead of rather cryptic "-g".
Test Plan: Ran "arc diff" in my hg working copy, got the diff I expected.
Reviewers: Makinde, aran, jungejason, nh, tuomaspelkonen
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 934
Summary:
- Noncontentious Mercurial stuff from D863.
- Use "hg outgoing --branch" to make sure we aren't including outgoing
revisions on other branches.
- Use "hg summary" to figure out where the working copy is, so "hg up <older
rev> && arc diff" works like expected.
Test Plan: Ran "arc diff" from multiple branches with the working copy in
various states.
Reviewers: Makinde, aran, jungejason, nh, tuomaspelkonen
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 936
Summary: Parse some other fields which I hadn't seen yet from "hg log". Select
the right "hg log" range for local commit info.
Test Plan: Ran "arc diff" in my test mercurial repository from a branch with
tags. Looked at "local commits" in Differential.
Reviewers: Makinde, aran, jungejason, tuomaspelkonen
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 880
Summary: When a revision is created, attach relevant information about the local
commits which it came from if applicable. This supports T473, for DCVSes and
DCVS workflows with immutable history where we can't just amend commit messages.
It will also allow us to enrich the web interface.
Test Plan: Will verify this info shows up for this very diff.
Reviewers: fratrik, aran, jungejason, tuomaspelkonen
Reviewed By: fratrik
CC: aran, epriestley, fratrik
Differential Revision: 857
Summary:
- Build the manifest of file changes so unit and lint workflows work.
- Default to creating a diff between the parent of the first outgoing change
and the tip.
Test Plan:
- Ran "arc diff" in a dirty mercurial repo, got warned about
untracked/uncommitted changes.
- Ran "arc diff" in a clean mercurial repo, got a diff of everything I'd done
locally.
Reviewed By: aran
Reviewers: Makinde, aran, jungejason, tuomaspelkonen
CC: aran, epriestley
Differential Revision: 796
Summary:
There's a lot of ground left to cover but this makes "arc diff" work (on one
trivial diff) in my sandbox, at least, and supports parsing of Mercurial native
diffs (which are unified + a custom header). Piles of missing features, still.
Some of this is blocked by me not understanding the mercurial model well yet.
This is also a really good opportunity for cleanup (especially, reducing the
level of "instanceof" in the diff workflow), I'll try to do a bunch of that in
followup diffs.
Test Plan: Ran "arc diff" in a mercurial repository, got a diff out of it.
Reviewed By: aran
Reviewers: Makinde, jungejason, tuomaspelkonen, aran, codeblock
CC: aran, epriestley, codeblock, fratrik
Differential Revision: 792
Summary:
right now 'arc blame' uses the user's blame.date setting, but
it will only work when the format is iso. So it will fail for user who
has customized it.
Test Plan:
run arc cover with 'blame.date' set to relative and verified
that 'arc cover' still works
Reviewed By: epriestley
Reviewers: epriestley, codeblock
CC: hwang, aran, epriestley
Differential Revision: 745
Summary:
executes the calls to git in parallel to improve startup performance of
arc lint and possibly other commands
Gets about a 2~3x speedup when repo is in buffer cache, with 4 cores.
Test Plan:
arc linted a repo with unstaged, untracked, staged, and committed
changes, see that the same files were linted.
Reviewed By: jungejason
Reviewers: jungejason, epriestley
Commenters: epriestley
CC: aran, jungejason, epriestley
Differential Revision: 594
Summary:
If you checkout some commit you end up on "(no branch)" which currently breaks
the parser. Ignore that, and add revision IDs to the output. They aren't very
big and I hate flags so I didn't add a flag for this. You can add a flag to turn
them off if you really want.
Test Plan:
Ran "arc branch" and "arc branch --by-status" from a "(no branch)" working copy.
Reviewed By: slawekbiel
Reviewers: slawekbiel, ahupp, jungejason, tuomaspelkonen, aran, schrockn
CC: aran, slawekbiel
Differential Revision: 555
Summary:
Turns out you can hava a branch named |foo.
(Good that '&& rm -fR /usr' is not a valid name)
Test Plan:
created a branch |test, run arc branch
Reviewed By: epriestley
Reviewers: epriestley, dschafer
CC: aran, epriestley, slawekbiel
Revert Plan:
sure
Other Notes:
Differential Revision: 507
Summary:
Appending differential status, sorting, filtering and coloring git
branches.
I think it turned out rather nicely. On my repository with 70 branches
it takes 1.6s, not terrible, though 1.2s is in the conduit call - seems
like there is potential for optimization.
I didn't end up changing 'arc list', as their semmantics are slightly
different, but I'm open to ideas of consolidating them
Test Plan:
- Tested on both facebook www and arcanist repositories.
- Validated that view-all flag works
- Validated that the ordering is correct
- Validated that the statuses match the differential status.
Reviewed By: epriestley
Reviewers: epriestley
CC: aran, epriestley, slawekbiel
Revert Plan:
sure
Other Notes:
Differential Revision: 497
Summary:
The amend process used "git log HEAD^..HEAD" to get log for the
commit being amended. When run on a merge commit this can return
any number of commits from the non-first parents. Since only a
single commit was expected, arc fails here.
This diff changes the amend process to use the '--first-parent' flag
to be consistent with using '^', which references the first parent.
This should guarantee a single commit log every time.
Test Plan:
arc amend on a merge commit
Reviewed By: epriestley
Reviewers: epriestley, jungejason
CC: aran, epriestley, andrewjcg
Differential Revision: 415
Summary:
Adds "--no-textconv" to all 'git diff' commands so we don't invoke textconv. See
T178 for discussion.
Test Plan:
Added something like this to .gitattributes:
*.txt diff=uppercase
And then this to .git/config:
[diff "uppercase"]
textconv = /path/to/uppercase
...where "uppercase" is a script which takes a file and emits an uppercase
version of it.
Then I added a "wisdom.txt" text file:
The cow goes "moo".
The duck goes "quack".
Without this patch, the file appears in uppercase in Differential, i.e. textconv
runs. With this patch, it appears as the original text.
Reviewed By: tuomaspelkonen
Reviewers: tuomaspelkonen, jungejason, aran
CC: elgenie, aran, tuomaspelkonen
Differential Revision: 372
Summary:
Sometimes we need to show abbreviated hashes, passing --short is more
reliable than substring of fixed number of characters.
Test Plan:
- called with and without the parameter, got correct results
-
Reviewed By: jungejason
Reviewers: jungejason
CC: jungejason
Revert Plan:
sure
Other Notes:
Differential Revision: 170
Summary:
Modifying a new file and running 'arc cover' before committing the
modifications confused arc. The problem was that 'unstaged' status
erased 'added' status and this caused problems.
Test Plan:
Tested that arc cover ignores added files when they are modified, but
old files that are modified are still handled correctly by arc cover.
Reviewed By: epriestley
Reviewers: epriestley
CC: jungejason, epriestley
Differential Revision: 122
Summary:
I simplified this code at some point and broke it horribly
in the process. Mask in the right flag here and combine the maps
correctly. This solves the great mystery of increased developer
clowniness.
Test Plan:
Ran 'arc diff' in a working copy with staged but
uncommitted files, got an error message (previously, I didn't).
Reviewed By: aran
Reviewers: mroch, aran
CC: aran
Differential Revision: 104
Summary:
SVN has a special 'replaced' status which was not being parsed
correctly. Parse it correctly.
Test Plan:
Replaced a file, ran arc diff, got a sensible diff.
Reviewed By: aran
Reviewers: aran
CC: aran
Differential Revision: 60
Summary:
Git (the world's hardest revision control system) allows you to change
output formats by accident and/or without your direct knowledge. Protect users
from themselves.
Test Plan:
Changed "pretty" in [format] to "format:quack" so every log just
outputs the word "quack". Ran "arc diff" successfully.
Reviewed By: aran
Reviewers: aran
CC: epriestley, aran
Differential Revision: 56
Summary: Haiping is getting a pretty confusing error message when trying to
commit.
Test Plan: Created a mock repository, installed the hook, made commits against
directories with bad .arcconfigs.
Reviewers:
CC:
Summary: When running "arc diff" in a mixed-base-revision working copy, we
prevent the operation. Relax this restriction so that having a different root
revision is okay, so long as all affected files share the same revision. This
facilitates multiple similar edits without updates.
Test Plan: Reverted a file to an older revision in an SVN working copy, ran
"arc diff", was rejected. Applied patch, ran "arc diff", diff went through and
was recorded with the right SVN base revision in the database.
Reviewers:
CC:
