adding code coverage to testing
dummy commit to get buildbot to fire
get paths right
listening to error messages helps when debugging - I accidentally double-added the attribute plugin
a hopeful fix to get buildbot to properly ignore integration and performance tests
flagging performance tests as performance tests, raising exceptions on tests that take a long time (so we know to fix them) and flagging integration tests. Also, I had the logic reversed on buildBotMode, it's now off by default (which was the original plan) a few tests get broken by this commit, I'll clean it up once buildbot runs over it
tell buildbot to ignore performance tests by default
Output xunit test results.
test messaging
add bin to the PATH via the env call
change testrunner to print out the Module.Class.Method of the currently running test
mark/skip integration tests
Add in WMCORE_ROOT, WMCOREBASE and WTBASE
Add get_relative_path() back which Melo accidentally (I assume) deleted...
Adding in command line option/machinery to let you completely delete your database after each test. (AKA, giving you enough rope to hang yourself)
Add COUCHURL to the env
bomb out on tests better
trying to find out where my debug statements go
trigger cvs mail
fixing failingness
trying with UTC + 60 secs
testing timezone bump .. trying to figure out what time the CVS server is on
backing out a change. buildbot was using central, not UTC time, so the CVS checkouts weren't working properly. undoing what I just did to see if it actually worked
throw exit codes on failures, return doesn't work that way
further smoothing this out
documenting
more testing. testing a testing framework is a very testing experience. test.
dummy commit to trigger mailsource
hunting a weird glitch with test runner
removing old code
removing debug statement
producing proper exit codes for unittest
Correctly print the env if setup is called from another directory
Ignore configuration files when linting
working on output
figure out return values from testrunner
we don't need an environment variable called WMCORE_ROOT since we have had one called WMCOREBASE for quite some time now
testing a new test runner which will keep make the output *much* nicer and simplify the (crufty) test-finding code I wrote. should be a temporary fix while I make sure buildbot doesn't choke
If lint isn't on the path don't create a LinterRun class
Turn on detailed lint reports by adding a -r flag
Allow for passing in a module for linting
better formatting, less syntax errors
better formatting
fixing output
cleaning up test output to make it less ugly
removing some debug statements
catch all exceptions on module import. syntax errors shouldn't break the whole build system
committing code that works, I don't care how much dave will try to shame me
pushing where I am real quick so others can look at it.
getting closer, have different bugs now
provide a sensible default for the WMCOREBASE environment variable. not sure if we even want to keep it though...
catch all errors around the import. don't want to ruin the fun.
handle exceptions better when setting up unittests. report if a test module fails to load (usually due to missing externals), but still run the others
replace relpath, which doesn't exist in python24
debugging
working on pythonhome
trying another way of getting the unittests out of a directory
making this work with buildbot
fix dumb bug
try to get more verbosity
there is, apparently, a difference between addTest() and addTests(). A LOT of tests weren't running because of this.
make setup.py more verbose when running tests. Also, this commit should interrupt buildbot from doing a full test. let's see.
dummy commit. last test of the CVS-mail-slurping bits on buildbot
trying to do loadTestsByName fails if a previous import already imported the module, give a 2nd try with loadTestsByModule in that case
adding more verbosity to tests
dummy commit, testing buildbot
turning back on stdout/stderr
fixing the difference between pass and fail
I'm a moron. backing out my changes. but only half of them. I can't even consistently fail.
bugfix, trying to trigger buildbot
further modifications
modifying test output to match what buildbot wants
cleaning up dir stuff
do paths intelligently
Give some polite messages when trying to run lint/coverage without the modules being installed.
Minor doc addition.
Add in EnvCommand to make setting up a sane environment easy. Use via eval `python setup.py -q env`.
fixed a bug in the coverage tests, should work now
adding code coverage, which works except for the fact that it doesn't. need to hunt down why the command-line version of this works fine, but the API version reports 0 lines executed
give a better explanation if someone tries to run lint without having lint installed.
same munging, adding right file this time
munging unittests for buildbot output.
trivial formatting change
Improve dumb coverage test
Add in crude test coverage command
add some minor comments
Further ditching of lint output (stdout) when running report
Ditch the lint output (stderr) when running report
Pick up how the lint score is calculated from the rc file
Add report and coverage commands for later use...
didn't seem to like the space for the -r arg
Change formatting of lint to work with buildbot, hopefully....
Trying to get some easier to read ouptut from the lint...
Failed unit tests now return exit code 1
add Alerts.Oracle in setup
Ballsed up the imports (sorry Steve, though everyone should have the lint module... ;))
Correctly set the python path (sys.path) to import the needed modules
Not everyone has a pylint module.
Make lint give a parseable output
Better handling of lint fails (seems exceptions are silently thrown by lint in py2.4), also, only lint .py files
Add test/python and src/python directories to PYTHONPATH in case they aren't there already.
Correct some documentation
Add some docs and a todo
Add a lint command which runs pylint with our standards file on all code in src.
Add some docs, auto generate the list of packages
Add test and clean commands, can now automate the testing in buildbot. Currently not doing so well: "FAILED (failures=5, errors=468)" oops ;) TODO: Pull database URL's from environment, and skip tests where database URL is not present (e.g. for a slave without Oracle connection)
Add DBSBuffer.
Add a todo to derive and run tests from setup.py (mainly to trigger a build...)
There's no JSM on this branch...
Port this over from HEAD.
Add WebTools and HTTPFrontEnd.
Add WebTools.
Add twitter.
Add WMCore.JobStateMachine.
ErrorHandler.Handler no longer exists
Add TreeSort algorithm
Add the WMBS Oracle port and the Job mask DAO objects.
Add SiteScreening
Add the Alerts code.
Removed non comitted dir.
Added missing package.
Updated packages in setup file.
add new services
Added WMComponent as additional lib directory.
Updated setup file.
Removed package WMCore.JobFactory which does not seem to exist.
-added ThreadPool to setup.py -started on trigger class
Added scripts for building wmcore
This form allows you to request diffs between any two revisions of this file. For each of the two "sides" of the diff, select a symbolic revision name using the selection box, or choose 'Use Text Field' and enter a numeric revision.