Presubmit Scripts
Overview
git cl
and git-cl
will check for and run presubmit scripts before you upload
and/or commit your changes. Presubmit scripts are a way of applying automatic
verification to your changes before they are reviewed and/or before they are
committed.
Presubmit scripts can perform automated checks on the files in your change and
the description of your change, and either fail your attempt to upload or
commit, show a warning that you must acknowledge before uploading/committing, or
simply show an informational message as part of the output of git cl
.
Examples of things presubmit scripts may be useful for include:
- Ensuring your change is linked to a bug via a BUG= line.
- Ensuring you have run cpplint.py or automatically running cpplint for you.
- Enforcing coding conventions.
- Preventing you from breaking dependency rules e.g. by including a header that code in your directory is not supposed to depend upon.
- Warning you if you haven't built your client and/or run unit tests within the last X hours.
Generally you want the same tests for upload and committing, but there are exceptions.
Firstly, the commit bot can autofix various problems, so these problems only need to be checked on commit (to check that they have been fixed), not on upload (another good reason to use the commit queue!). Secondly, you can skip slow tests on upload and only run them on commit, if running these every time slows development too much, but this can result in problems only being caught at the last minute, and thus requiring last-minute changes after you think they're good to go.
Bypassing tests
To skip the scripts on upload, use the --bypass-hooks
flag, as in:
git cl upload --bypass-hooks
To skip the scripts on commit, use --bypass-hooks and directly commit your change.
You should only do these if necessary, as the presubmit scripts are there for a reason, but they're not perfect.
If you have trouble with a presubmit script, it's preferable to fix it, rather than simply bypassing it. See depot_tools: sending patches for how to contribute.
If you're not sure which presubmit check is generating results you can request
double-verbose mode with git cl presubmit -v -v
. This will trigger noisy
diagnostics, which append a call stack to each presubmit result.
Design
When you run git cl upload
or git cl commit
, git cl
will look for all
files named PRESUBMIT.py
in folders enclosing the files in your change, up to
the repository root.
For each such file, it will load the file into the Python interpreter and call the appropriate Check functions, depending on what presubmit version you have selected (details below) and whether you are committing or uploading.
The same applies to git-cl upload
, git-cl dcommit
and git-cl push
.
Please note that presubmit scripts are a best-effort kind of thing; they do not
prevent users from submitting without running the scripts, since one can always
dcommit, and in fact there is a --bypass-hooks (formerly --no_presubmit
) flag
to git cl that skips presubmit checks. Further, since they use the local copy of
the PRESUBMIT.py
files, users must sync their repos before the latest
presubmit checks will run when they upload or submit.
More subtly, presubmit scripts do not guarantee invariants: even if presubmit scripts pass prior to submission to CQ, once all changes land, the scripts may fail! This is because 2 changes may individually pass the tests, and the patches both apply cleanly together, but the combined change does not pass tests. Since presubmit/precommit scripts run at upload or at start of CQ steps, if two such changes are in the CQ at the same time, they can both pass, both be enqueued, and both land, at which point the tests start failing. A common example is change 1 adding a new test, and change 2 changing existing tests. After they both land, there is a new test in the old style (from change 1), which is out of sync with the new tests (from change 2).
Another very common way that presubmit failures can land is when a PRESUBMIT or
associated settings are changed. Because presubmits generally only check the
files in the change and because presubmits and their settings affect many other
files, a disconnect can easily occur. For instance, the .eslintrc.js
files
control what warnings are enabled in .js and .ts files. However, the presubmits
only check .js and .ts files when they are modified, so a change that enables a
warning in .eslintrc.js
can pass presubmits even if hundreds of files will
trigger the warning. Ideally, developers will run git cl presubmit --files "*.js;*.ts"
to check these, but this is not enforced. In general, modifying presubmits is
the most common way to introduce latent presubmit failures. Occasionally running
git cl presubmit --all
and looking for errors is the only reliable way to find
these.
Writing tests
To create a new test, either create a new PRESUBMIT.py script or edit an existing one, adding a new function for your test.
To check your changes, first commit locally (else git cl
will complain about
the dirty tree), then:
To test the upload checks:
git cl presubmit --upload
To test the submit checks:
git cl presubmit
Version 1
In the original presubmit system the functions must match these method signatures. You do not need to define both functions if you're only interested in one type of event, and if you want to run the same tests in both events, just have them both call a single underlying function:
def CheckChangeOnUpload(input_api, output_api):
CommonChecks(input_api, output_api)
def CheckChangeOnCommit(input_api, output_api):
CommonChecks(input_api, output_api)
The input_api
parameter is an object through which you can get information
about the change. Using the output_api
you can create result objects.
Both CheckChangeOnXXX
functions must return a list or tuple of result objects,
or an empty list or tuple if there is nothing to report. The types of result
objects you may use are output_api.PresubmitError
(a critical error),
output_api.PresubmitPromptWarning
(a warning the user must acknowledge before
the command will continue) and output_api.PresubmitNotifyResult
(a message
that should be shown). Each takes a message parameter, and optional "items" and
"long_text" parameters.
Version 2
Presubmit version 2 reduces some of the overhead of managing which checks are executed on upload and/or submit. To enable presubmit version 2, in the global scope of your presubmit script, define:
PRESUBMIT_VERSION = '2.0.0'
In presubmit version 2, you are not required to use CheckChangeOnUpload
and
CheckChangeOnCommit
as your entry points. Any function that begins with the
prefix Check
will be executed, receiving input_api
and output_api
as its
parameters. CheckXXX
functions that end in Upload
will only be executed on
upload and CheckXXX
functions that end in Commit
will only be executed on
commit. CheckXXX
functions that do not end in either suffix will be executed
at both upload and commit. The format of return values of CheckChangeOnXXX
functions is the same as for CheckChangeOnUpload/CheckChangeOnCommit
. This
makes existing presubmit scripts backwards compatible with presubmit version 2
provided there are no functions in the global scope of the script that begin
with Check
. Sample functions might look like this:
def CheckSomeInvariant(input_api, output_api)
DoSomething()
def CheckSomeInvariantUpload(input_api, output_api)
DoSomething()
def CheckSomeInvariantCommit(input_api, output_api)
DoSomething()
InputApi
The input_api
parameter is an object of type presubmit.InputApi
; see class InputApi
in
presubmit_support.py
for implementation.
This object can be used to transform from local to repository paths and vice
versa, and to get information on the files in the change that are contained in
the same directory as your PRESUBMIT.py
file or subdirectories thereof.
The input_api.change
object represents the change itself. Using this object
you can retrieve the description of the change, any key-value pairs in the
description (e.g. BUG=123), and details on all of the files in the change (not
just the ones contained by the directory your PRESUBMIT.py
file resides in).
An input_api.canned_checks
object contains a set of ready-made checks that you
can easily add to your presubmit script.
The os/os.path
module is available as input_api.os_path
, so you do not need
to import this yourself.
Files in the API are represented by an AffectedFile
object through which you
can query the LocalPath()
, ServerPath()
, and the Action()
being performed
('A', 'M' or 'D' for Add, Modify or Delete).
The input_api.is_committing
attribute indicates whether the CL is being
committed or just uploaded. This is particularly useful if you wish the same
test to be run for both upload and committing, but with different behavior. A
common pattern is to prompt a warning on upload, but an error on committing,
which allows a CL to be uploaded and reviewed even if the test fails, but not
committed (without dcommit). This can be done as follows:
if input_api.is_committing:
message_type = output_api.PresubmitError
else:
message_type = output_api.PresubmitPromptWarning
Caveats
It is possible to run arbitrary Python code in the presubmit scripts. To avoid side effects and hard-to-debug errors, it is safest to run tests in subprocesses. The InputApi object provides various facilities to assist with this; see example below.
Please note that you should not use any functions or attributes on the API objects that begin with an underscore (_) as these are private functions that may change, whereas all public functions will be retained through future versions of the API.
Details and Example
The most detailed documentation for the presubmit API is in its implementation code.
The canned checks are good examples of what you can do with the presubmit API.
An example simple file might be as follows:
# Optional but recommended
PRESUBMIT_VERSION = '2.0.0'
# Mandatory: run under Python 3
USE_PYTHON3 = True
def CheckChange(input_api, output_api):
results = []
results += input_api.canned_checks.CheckDoNotSubmit(input_api, output_api)
results += input_api.canned_checks.CheckChangeHasNoTabs(input_api, output_api)
results += input_api.canned_checks.CheckLongLines(input_api, output_api)
# Require a BUG= line and a HOW_TO_TEST= line.
if not input_api.change.BUG or not input_api.change.HOW_TO_TEST:
results += [output_api.PresubmitError(
'Must provide a BUG= line and a HOW_TO_TEST line.')]
return results
However many of the canned checks, such as ChecksCommon
and CheckLongLines
,
are called from the root-level PRESUBMIT.py
(either directly or through
PanProjectChecks
) and therefore needn't be called from other Chromium
presubmit scripts.
A simple example of a custom command is:
def CheckMyTest(input_api, output_api):
test_path = input_api.os_path.join(input_api.PresubmitLocalPath(), 'my_test.py')
cmd_name = 'my_test'
cmd = [input_api.python3_executable, test_path]
test_cmd = input_api.Command(
name=cmd_name,
cmd=cmd,
kwargs={},
message=output_api.PresubmitPromptWarning)
if input_api.verbose:
print('Running ' + cmd_name)
return input_api.RunTests([test_cmd])
Alternatively, you can set cmd = [test_path]
and then pass python3=True
to
run the script under Python 3. input_api.python_executable
(runs Python 2) is
deprecated and should not be used in new presubmits.
You can look at existing scripts for examples (search: PRESUBMIT.py). Chromium's PRESUBMIT.py and Chromium-os PRESUBMIT.py are detailed examples, while the Blink bindings PRESUBMIT.py is a very simple example.