mirror of
https://github.com/zeek/zeek.git
synced 2025-10-02 22:58:20 +00:00
357341c887
- The absolute path canonifier was overzealously canonifying relevant log fields, so it's no longer generally applied to diffing all baselines. I don't think there's any logs that require local filesystem path names that aren't already tested by a unit test, but if any show up in the future, they can be canonified on a case-by-case basis. - Removed some logs from being diff'd in the diff-all script because they're either already covered by a unit test (load_scripts.log) or because of difficulty/maintainenance tradeoff (prof.log). Baselines for the external bro-testing repo still need updating. |
||
|---|---|---|
| .. | ||
| Baseline | ||
| scripts | ||
| .gitignore | ||
| Makefile | ||
| random.seed | ||
| README | ||
| subdir-btest.cfg | ||
Test Suite for Large Trace Files
================================
This test-suite runs more complex Bro configurations on larger trace
files, and compares the results to a pre-established baseline. Due to
their size, both traces and baseline are not part of the main Bro
repository but kept externally. In addition to the publically provided
files, one can also add a local set to the test-suite for running on
private traces.
Initialization
--------------
Before the test-suite can be run, one needs to download the necessary
files. Test and baselines are kept in git repositories, while any
traces are download directly. A ``Makefile`` is provided to get
everything that's needed initially:
.. console:
> make init
If you need a proxy to download the traces, enter it into a file
``.proxy`` either in the top-level directory or inside one of the
repositories.
To later update to upstream changes:
.. console:
> make pull
This updates the tests and the traces as necessary.
Running Tests
-------------
The easiest way to run all tests is simply typing ``make``. Doing so
will iterate through all git repositories found in the current
directory and run the tests in there. Output for failed tests will be
in files ``diag.log`` in the top-level repository directories.
Alternatively, one can also manually run all tests inside a single
test repository:
.. console:
> cd bro-testing
> btest
All the standard ``btest`` options can be used to run individual
tests, get diagnostic output, etc.
Updating Baseline
-----------------
To update a test's baseline, first run ``btest`` in update mode:
.. console:
> cd bro-testing
> btest -u tests/test-you-want-to-update
Then use ``git`` to commit the changes and push the changes upstream
as usual.
Adding a Local Repository
-------------------------
One can add local non-public set of tests (potentially using private
traces) by creating a git repository of a similar structure as the
public one.
If you already have such a private test repository that you want to
include into the test suite, clone it directly into ``<repo-name>``.
If you want to create a new private repository, there's a helper
script to set that up:
.. console:
> ./scripts/create-new-repo <repo-name> <repo-url>
The first argument is the local name of the repository (it will be
cloned into ``<repo-name>``); and the second is the URL of the git
repository. The repository will be initialized with a few standard
directories as well as a skeleton test in ``<name>/tests``. You can
then edit files as needed. You add trace files by editing
``Traces/traces.cfg``; see the comments in there. For each trace, you
also need to calculate a checksum with ``md5sum`` and put it into
``<url>.md5sum``. The scripts use this to decide if they need to
redownload the trace. Accordingly, if you update a trace, make sure to
also recalculate its checksum. Note that the traces will be downloaded
to ``Traces/`` but must not be added to the git repostiory; there's a
``.gitignore`` installed to prevent that.