summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* A simple test of a test functionNed Batchelder2018-11-111-0/+17
|
* One last decorator test change for 3.8Ned Batchelder2018-11-111-2/+10
|
* Update an HTML gold test due to new peephole optimizationsNed Batchelder2018-11-113-47/+37
|
* Comparison should always be expected,actualNed Batchelder2018-11-111-6/+6
|
* Canonicalize the XML outputNed Batchelder2018-11-111-5/+24
| | | | | | | | | | https://bugs.python.org/issue34160 added retaining the user's attribute order to the XML output, which removed the sorting that used to happen. This broke our XML tests, which compare against saved gold files. This adds in a rough-and-ready canonicalization to avoid the problem. Maybe the core devs will eventually support a sort_attributes option, and I can get rid of this.
* Refactor the XML comparisonNed Batchelder2018-11-111-23/+16
|
* Comparison is not symmetric, it's expected vs actualNed Batchelder2018-11-115-42/+44
|
* Always compare expected then actualNed Batchelder2018-11-111-17/+17
|
* Keep test tempdirs distinct, for saving and comparing test outputNed Batchelder2018-11-112-8/+13
|
* Python 3.8 will optimize away "while True:"Ned Batchelder2018-11-114-10/+52
|
* Better error handling in set_env.pyNed Batchelder2018-11-091-6/+22
|
* Minor cleanup in tox.iniNed Batchelder2018-11-051-3/+5
|
* Debug-time environment variables can be set with set_env.pyNed Batchelder2018-11-037-1/+103
|
* Record that a bug was fixed. #700Ned Batchelder2018-11-031-1/+4
|
* Everything should get a docstringNed Batchelder2018-11-032-0/+2
|
* Adapt to 3.8's way of tracing decorated functionsNed Batchelder2018-11-034-14/+56
|
* You can turn off contracts while debugging testsNed Batchelder2018-11-031-1/+5
|
* Make this useful for py3 alsoNed Batchelder2018-11-031-1/+2
|
* More correctNed Batchelder2018-11-011-1/+1
|
* Better version showingNed Batchelder2018-10-311-2/+1
|
* Remove use_2to3=False; it is the defaultJon Dufresne2018-10-271-7/+0
| | | | Slightly simplifies setup.py.
* Use a new event loopNed Batchelder2018-10-241-2/+2
|
* Add 3.7, and other appveyor changesNed Batchelder2018-10-241-15/+37
|
* Always include a documentation link at the end of help messagesNed Batchelder2018-10-212-14/+24
|
* Make directories for the data file if needed. #721Ned Batchelder2018-10-213-0/+14
|
* Helpers for making directoriesNed Batchelder2018-10-213-7/+20
|
* Clean __pycache__ before .pyc (a little faster?)Ned Batchelder2018-10-191-1/+1
|
* PyPy3 fixed an obscure bug long ago, we don't need this overrideNed Batchelder2018-10-191-7/+3
|
* One more error case for the command lineNed Batchelder2018-10-192-3/+21
|
* Clean up a metacov .pth file that could get left behind.Ned Batchelder2018-10-192-0/+2
|
* `[run] command_line` is the command line to use for 'coverage run'. #695Ned Batchelder2018-10-185-4/+64
|
* This was in the wrong placeNed Batchelder2018-10-181-3/+3
|
* We weren't using this argumentNed Batchelder2018-10-181-2/+2
|
* This is literally what setdefault is forNed Batchelder2018-10-181-2/+1
|
* Clean some lintNed Batchelder2018-10-172-4/+8
|
* Don't clobber sys.path[0]. #715Ned Batchelder2018-10-162-4/+4
|
* A test for 'no such table: meta' as in #716.Ned Batchelder2018-10-161-0/+12
|
* Allow later DebugOutputFile to replace earlier onesNed Batchelder2018-10-161-6/+12
| | | | | | When logging calls, get_one() is called with no filters. It would be the_one, so a later get_one with filters for pids wouldn't take effect. Now the earlier is only interim, and the later one wins.
* Put this helper in the right placeNed Batchelder2018-10-161-6/+7
|
* Note a bug we closedNed Batchelder2018-10-161-1/+3
|
* Do we need more than 10 tries?Ned Batchelder2018-10-151-1/+1
|
* Use one transaction to speed combiningNed Batchelder2018-10-152-40/+50
|
* Michał suggested this would make the information easier to find.Ned Batchelder2018-10-142-3/+4
|
* Defer using the database when calling set_context #716Ned Batchelder2018-10-143-14/+19
| | | | | | | | | | | The collector calls set_context() before any code is run. If we touch the database there, it will get created *very* early. This causes problems with pytest-cov, which will delete those early-created files when erasing data. By deferring the database access until add_lines is called, the data file stays off the disk until the collection is done (or until the context switches), which avoids the problem.
* Use --source to ensure tests are showing what we want.Ned Batchelder2018-10-143-8/+10
| | | | | | | | | | | Travis started having these extra files in coverage reports from our tests: site-packages/pkg_resources/_vendor/six.py site-packages/pkg_resources/extern/__init__.py I don't know why they are being pulled in, but the tests aren't interested in them, and it is throwing off the results.
* Have to rethink how to use contexts ourselvesNed Batchelder2018-10-142-4/+0
| | | | | | Contexts fail when using PYTRACER, and we want to measure coverage under PYTRACER (at least for PyPy), so we need a more sophisticated strategy for how to use contexts.
* Missed one file when updating license linksNed Batchelder2018-10-141-1/+1
|
* Not sure how I went for so long without one of these creeping inNed Batchelder2018-10-141-0/+3
|
* Oops, we were always measuring coverage on TravisNed Batchelder2018-10-131-1/+1
|
* Fewer conditionals for debug outputNed Batchelder2018-10-133-23/+30
|
View Lorry Controller Status