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:
Right now, if you specify "/api" instead of "/api/" the entire world breaks in a
horrible way. Generally, this whole thing is silly. Just use the right path.
Test Plan:
Edited .arcconfig to use the "wrong" path, so this diff demonstrates the change
works.
Reviewed By: llorca
Reviewers: jungejason, llorca, aran, tuomaspelkonen
CC: aran, llorca, epriestley
Differential Revision: 506
Summary:
This should be implemented more elegantly, but this is a mostly-reasonable
attempt at it.
Test Plan:
Ran 'arc diff --only --json' with this diff.
Reviewed By: gc3
Reviewers: gc3
CC: aran, epriestley, gc3
Differential Revision: 509
Summary:
cpojer fixed a JS instance of this in D481, but we can catch it in PHP with
XHPAST. Add a lint rule to fail if nested loops use the same iterator.
Test Plan:
Ran unit test.
Reviewed By: tomo
Reviewers: aran, pad, cpojer, jungejason, tuomaspelkonen, tomo
Commenters: cpojer
CC: aran, cpojer, epriestley, tomo
Differential Revision: 489
Summary:
This documentation isn't terribly clear and it isn't obvious how to use the
workflow. Make it more clear and provide examples.
Test Plan:
Ran "arc help call-conduit"
Reviewed By: gc3
Reviewers: gc3, aran, jungejason
CC: aran, gc3, epriestley
Differential Revision: 502
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: This once made sense but hasn't been a reasonable default for a long time, get rid of it.
Test Plan: Ran "arc list".
Reviewers: llorca, toulouse
CC:
Differential Revision: 496
Summary:
When installing a certificate, it's being written to ~/.arcrc not ~/.arcconfig
Test Plan:
Installed a certificate it said the right thing.
Reviewed By: aran
Reviewers: epriestley, aran
Commenters: epriestley
CC: aran, epriestley
Differential Revision: 491
Summary:
Provide an "install-certificate" workflow to simplify ~/.arcrc edits. See also
D460.
Test Plan:
Installed certificates via "arc install-certificate".
Reviewed By: aran
Reviewers: aran, jungejason, tuomaspelkonen
CC: aran, epriestley
Differential Revision: 465
Summary:
I looked at this quickly to see what was involved and there were only a couple
of issues so here's a patch:
- On OSX, the "-i" flag does not mean "--mime". Use "--mime" explicitly
instead. This is a minor fix which affects only OS X.
- I wasn't able to repro the "crazy executables" behavior and think it might
have been me messing something up, so nuke it until we see an issue.
- Some guid vs phid shenanigans. Differential reads "phid" but arcanist set
"guid". We should move toward "phid"; I started using "guid" before I realized
it was an overloaded term that also refers to a specific GUID implementation
(Microsoft's UUID).
Test Plan:
Uploaded an image diff, saw images in Differential.
Reviewed By: aran
Reviewers: jianfeng, jungejason, aran, tuomaspelkonen
CC: aran
Differential Revision: 466
commit template
Summary:
When users run into this, point them at the documentation explicitly.
Test Plan:
Tried to "arc diff" with a commit message of 'derp', got a live link to
documentation instead of a vague set of general instructions.
Reviewed By: aran
Reviewers: moskov, aran, jungejason, tuomaspelkonen
CC: aran
Differential Revision: 468
Summary:
We have some "@generated" files of these types now, hit them with the text
linters
Test Plan:
Ran 'arc lint' on D444
Reviewed By: aran
Reviewers: aran
CC: aran
Differential Revision: 451
Summary:
This is pretty coarse and could be refined, but I often do this when testing:
lang=diff
- if (some_complicated_condition())
+ if (true || some_complicated_condition())
aran has caught me not only doing it but sending out diffs with it like 30
times. Catch it in lint instead.
Test Plan:
Unit test, added a "true || $junk" to the code and linted it.
Reviewed By: aran
Reviewers: aran, jungejason, tuomaspelkonen, pad
CC: aran
Differential Revision: 447
unit'
Summary:
We want to handle 'arc unit' and 'arc diff' differently in our test
framework.
Test Plan:
Tested that with 'arc unit' the value was set to 'unit' and with 'arc diff'
the value was set to 'diff'.
Reviewed By: epriestley
Reviewers: slawekbiel, epriestley
CC: jungejason, grglr, aran, tuomaspelkonen, epriestley
Differential Revision: 430
Summary:
Differential showed 'okay' as the arc unit status even when there were
postponed tests.
Test Plan:
Tested that test results were pushed to differential when there were
postponed tests.
Reviewed By: jungejason
Reviewers: epriestley, jungejason
CC: slawekbiel, aran, jungejason
Differential Revision: 417
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:
We changed our Facebook implementation to echo test results while the
tests are running. We do not want to echo the test results twice.
Test Plan:
Tested that implementing the function in PhutilUnitTestEngine and
returning true showed the results and returning false didn't show the
results.
Reviewed By: epriestley
Reviewers: jungejason, epriestley, grglr, slawekbiel
Commenters: slawekbiel, aran
CC: sgrimm, slawekbiel, aran, tuomaspelkonen, epriestley
Differential Revision: 400
Summary:
I typed the wrong number of spaces into some of these.
Test Plan:
Visual inspection, ran 'arc unit'
Reviewed By: tuomaspelkonen
Reviewers: tuomaspelkonen
CC: aran, tuomaspelkonen
Differential Revision: 402
Summary:
Test Engine classes might need the differential Diff ID to be
able to attach postponed test results to diffs.
Test Plan:
Added setDifferentialDiff function to PhutilUnitTestEngine class
and made sure it was called with the correct diff ID when running
'arc diff --preview'
Reviewed By: jungejason
Reviewers: jungejason, grglr
Commenters: sgrimm
CC: epriestley, sgrimm, slawekbiel, aran, tuomaspelkonen, jungejason
Differential Revision: 395
Summary:
There are a bunch of different ways we could approach this, but practically I
think this is probably the best one even though it's kind of yucky.
A big part of my motivation here is that if we just reject UTF-8 outright, users
are going to end up in a situation they don't understand how to resolve.
UPDATE: after discussion, going with a more conservative approach until such
time as we have a more compelling case for the less strict functionality.
Test Plan:
Created a diff with a text file that contained invalid UTF-8 subsequences.
Reviewed By: tuomaspelkonen
Reviewers: tuomaspelkonen, jungejason, aran
Commenters: aran
CC: aran, epriestley, tuomaspelkonen
Differential Revision: 325
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:
This is sort of a guess. In D357, arc has gone crazy and removed a bunch of
'utils' dependencies for gc3. I think this is because his HPHP/i has them marked
as builtins.
However, I don't have an HPHP/i build (and probably don't have the fortitude to
build it outside of a Facebook environment, I spent about two hours on it at one
point and got maybe 25% of the way through the build process before running into
things I didn't know how to resolve) so I'm not sure this is the issue. No one
else's diffs have exhibited this problem eitehr, so I'm not confident this is
actually the problem or solution.
gc3, can you apply this locally to a copy of arcanist and then run 'arc lint'
(or 'arc lint --apply-patches') on your commit and see if it restores all the
libphutil/utils requirements?
Test Plan:
unable, see summary
Added 'array_keys' to this blacklist and verified the technical behavior of the
patch is correct by var_dump()'ing builtins.
Reviewed By: gc3
Reviewers: gc3, tuomaspelkonen, jungejason, aran
CC: aran, gc3, epriestley
Differential Revision: 361
Summary:
Raise an error if an array is initialized with the same key present more than
once.
Test Plan:
bin/arc unit
Also added some duplicates to ArcanistXHPASTLinter.php and verified the output
of bin/arc lint.
Reviewed By: epriestley
Reviewers: jungejason, epriestley, aran
CC: aran, epriestley
Revert Plan:
Tags:
Differential Revision: 346
Summary:
Provides a lint class as a wrapper around the external project PyLint.
This exposes some arc config variables to control the behavior:
lint.pylint.prefix - non-standard installation location of pylint
lint.pylint.logilab_astng.prefix - non-standard installation location
of logilab-astng, a dependency of pylint
lint.pylint.logilab_common.prefix - non-standard installation location
of logilab-common, a dependency of pylint
lint.pylint.codes.{error,warning,advice} - regexes matching against
PyLint message codes which should trigger arc errors/warnings/advice
lint.pylint.options - options to pass PyLint
Test Plan:
used to lint python code
Reviewed By: epriestley
Reviewers: epriestley, jungejason
CC: aran, epriestley
Differential Revision: 343
Summary:
Arc patch for image files failed, because they were treated like text
files.
Test Plan:
Tested that 'arc patch' worked for a png file.
Reviewed By: jungejason
Reviewers: epriestley, jungejason
CC: aran, jungejason
Differential Revision: 341
Summary:
I'm probably missing some edge cases but it took me almost 3 hours to get this
far and I think it only makes things work that didn't work before. Some stuff
like SVN binary patches still won't work, although they should be far easier to
implement.
Most of the magic here just comes from reading the git source code. It appears
to work correctly; I sprinkled printf() around git liberally and recompiled it
during development. Took me about 45 minutes to figure out that "Index" vs
"index" causes git to silently fail in a confusing way. :/
Git has a diff mode for binary changes but I don't think we lose much by always
using the full binaries. We can enhance it later if we want.
Test Plan:
Exported and patched binary changes (a picture of a duck) into a working copy.
Reviewed By: aran
Reviewers: tuomaspelkonen, jungejason, aran
CC: simpkins, aran, epriestley
Differential Revision: 327
Summary:
Provide a simple linter wrapper around pyflakes. This relies on finding
pyflakes via:
- lint.pyflakes.path - arcconfig setting of absolute path to pyflakes
- lint.pyflakes.prefix - arcconfig setting of the prefix that pyflakes
was installed under
- users path
Test Plan:
linted python code with PyFlakes warnings
Reviewed By: epriestley
Reviewers: jungejason, epriestley
Commenters: jungejason
CC: aran, epriestley, andrewjcg, jungejason
Differential Revision: 310
Summary:
When you run one copy of arcanist against another copy, you previously got a
meaningless error. After D311 you get a slightly more meaningful error.
Capture the new exception and raise an extremely specific error.
NOTE: Adding arcanist to .arcconfig forces the library conflict error to
trigger; otherwise there's just an implicit conflict because 'arc' reads the
running-copy library amp instead of the working-copy library map. This only
worked before because arcanist includes itself.
Test Plan:
Ran arcanist on a different copy of arcanist, got a good error message.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, jungejason, epriestley
Differential Revision: 312
Summary:
passthru() doesn't show up on --trace, and one time a while ago someone had an
issue which was harder than necessary to debug because the command wasn't
avialable in the log. Use the logged version.
Also fix a locale issue I ran into on my local machine, with "en_US.utf8" not
being a valid locale.
Test Plan:
Created an SVN repo and used "arc commit" to make commits against it.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, epriestley, jungejason
Differential Revision: 306
Summary:
This diff:
- Adds the PEP8 linter to the externals directory
- Changes the path for finding pep8.py
- Removes use of execx since pep8.py return an errors code
when it finds PEP8 violations
Test Plan:
tested linting python code
Reviewed By: epriestley
Reviewers: epriestley, jungejason
CC: aran, epriestley
Differential Revision: 309
Summary:
The --only flag implies --nounit and --nolint, so there's really no
conflict if these flags are specified together.
(My wrapper around arc diff automatically specifies --nounit in some
cases. It is useful to be able to manually add --only in some cases,
and not have it fail because it conflicts with the automatically
specified --nounit flag.)
Test Plan:
Ran "arc diff --only --nounit".
Reviewed By: epriestley
Reviewers: epriestley, jungejason
CC: aran, epriestley, simpkins
Differential Revision: 298
Summary:
Update the lint workflow to exit immediately it there is no lint engine
configured. Previously it scanned the repository to find the paths to
check before failing in this case. Scanning the repository can be
relatively slow on large repositories.
Test Plan:
Ran "arc lint" in a large repository with no lint engine configured.
Verified that it failed quickly, without scanning the repository first.
Reviewed By: jungejason
Reviewers: jungejason, epriestley
CC: aran, jungejason
Differential Revision: 297
Summary:
The story for creating and maintaining libphutil libraries and modules
is pretty terrible right now: you need to know a bunch of secret scripts and
dark magic. Provide 'arc liberate' which endeavors to always do the right thing
and put a library in the correct state.
Test Plan:
Ran liberate on libphutil, arcanist, phabricator; created new
libphutil libraries, added classes to them, liberated everything, introduced
errors etc and liberated that stuff, nothing was obviously broken in a terrible
way..?
Reviewed By: aran
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, epriestley
Differential Revision: 269
Summary:
Switch to the new PhutilServiceProfiler API.
Test Plan:
Ran "arc" with "--trace", got mostly reasonable output. I'll clean up conduit
output in a followup.
Reviewed By: aran
Reviewers: aran, jungejason, tuomaspelkonen
CC: aran
Differential Revision: 292
Summary:
This isn't necessarily a root-cause solution but this syntax is really really
bad. If we want to fix the root cause, I'd recommend making its use a lint
error?
Test Plan:
Unit tests failed before patch, passed afterward.
Reviewed By: aran
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran
Differential Revision: 270
Summary:
This is necessary for the Javelin linters. The libphutil and flib linters do it
implicitly, so there's no real tradeoff here.
Test Plan:
Ran javelin linters.
Reviewed By: tuomaspelkonen
Reviewers: tomo, aran, jungejason, tuomaspelkonen
CC: aran, tuomaspelkonen, epriestley
Differential Revision: 238
Summary:
I only dropped this because it's slightly inconvient to accommodate, but
empirically it's pretty confusing to users (who often use --diff 12345 when they
mean --revision 12345).
Test Plan:
Ran "arc patch D45", "arc patch --revision 45", "arc help patch"
Reviewed By: tuomaspelkonen
Reviewers: tuomaspelkonen, jungejason, aran
CC: aran, tuomaspelkonen
Differential Revision: 241
Summary:
Provide a formal mechanism for making conduit calls from other scripts.
Test Plan:
Called 'conduit.ping'.
Reviewed By: jungejason
Reviewers: aran, jungejason, tuomaspelkonen
CC: aran, jungejason
Differential Revision: 239
Summary:
the current command to apply a git diff is 'git apply --index',
which will fail the whole patch and does not touch the working tree when
some of the hunks do not apply. We want to allow the user to patch the
ones which apply.
Test Plan:
run 'arc apply' against a diff whcih partially applies and
verify that the hunks which apply are applied, and .rej files are
generated for the rest.
Reviewed By: simpkins
Reviewers: simpkins, epriestley
Commenters: ju
CC: aran, slawekbiel, simpkins, ju
Differential Revision: 235
Summary:
I'm making a move on Lisk stubbability and need these so I can set up
Lisk isolation unconditionally in Phabricator tests. Also document and organize
this class.
Test Plan:
Changed the expected counts in the meta-tests and got failures. Ran
all unit tests.
Reviewed By: aran
Reviewers: aran, jungejason, tuomaspelkonen
CC: aran
Differential Revision: 192
Summary:
Displaying each postponed test separately was spammy.
Now it will only display a single line with the total count
Test Plan:
~/local/arcanist/bin/arc unit --sandcastle-tests
35 tests to run.
Completed 'ParentInfoTestCase.php'
Interrupted 'MentionsUntagTestCase.php'
PASS ParentInfoTestCase
POSTPONED 34 tests
Reviewed By: sgrimm
Reviewers: jungejason, sgrimm
CC: sgrimm
Revert Plan:
sure
Other Notes:
Differential Revision: 173
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
