Zach Mullen
c32088cfee
Local test timeouts will now always take precedence over the global --timeout option.
2009-12-08 10:26:43 -05:00
Zach Mullen
b30f627190
CMake global timeout (--timeout option) should prevail over individual test timeouts if it is lower than the individual timeout.
2009-12-07 13:25:54 -05:00
David Cole
0b38bb4c53
Fix issue #2336 - honor the -C arg to ctest. Honor it for all stages of running -D dashboards from the command line and running ctest_configure, ctest_build and ctest_test commands in -S scripts. Also, allow a script to change it by setting the CTEST_CONFIGURATION_TYPE variable: allows for multiple configuration build/test cycles within one script. Add a new signature for the cmake command build_command that accepts CONFIGURATION as one argument. The original build_command signature is still there, but now marked as deprecated in the documentation. Of course... also add CTestConfig tests to verify that -C is honored for -D dashboards and -S scripts.
2009-12-04 12:09:01 -05:00
Bill Hoffman
22c37b2b19
Hanle the case where a test can not be run because it is a bad executable.
2009-12-02 16:37:43 -05:00
Zach Mullen
c17ce46c73
Align test output for timeout condition.
2009-11-30 16:09:00 -05:00
Zach Mullen
a52c0118d4
Added the --timeout option to ctest command line. This sets a global timeout on all tests if no more specific timeout is set on them.
2009-11-30 16:08:11 -05:00
Brad King
e1548142fb
CTest: Move initial checkout to ctest_start()
...
In CTest command-driven script mode we support starting without a source
tree. Previously the ctest_start() command would do some initialization
but could not do anything that required CTestConfig.cmake from the input
source tree. Later, ctest_update() would run CTEST_CHECKOUT_COMMAND to
create the source tree, and then re-initialize everything. This
delayed-initialization approach led to many complicated cases of which
only some worked. For example, the second initialization only worked
correctly in Nightly mode and simply failed for Experimental and
Continuous builds.
A simpler solution is to run CTEST_CHECKOUT_COMMAND during ctest_start()
and then have a single initialization path. In principle this change in
behavior could break scripts that set the checkout command after
ctest_start() but before ctest_update(). However, the convention we've
always followed has been to set all variables before ctest_start().
See issue #9450 .
2009-11-24 08:58:59 -05:00
Zach Mullen
2921fa96ed
Handle multiple carriage return issue on windows VS9. Also make sure that running ctest in showonly mode does not kill our cost store, since many ctest tests do this and corrupt our cost data.
2009-11-16 10:58:43 -05:00
Bill Hoffman
1b48d361d8
fix number of callers to work with newer versions of valgrind
2009-11-11 08:43:15 -05:00
Zach Mullen
c9c0ee4056
Change logic of ctest subdirs command to allow for absolute paths. Also added test coverage for passing absolute paths to subdirs.
2009-11-10 10:40:24 -05:00
Zach Mullen
f7899c7445
Fixed a bug where it was possible for a test to be started twice if a lower-indexed test depended on it.
2009-11-09 16:07:47 -05:00
Zach Mullen
17afd5257b
Bug 9090: CTest does not handle absolute paths in CTestTestfile SUBDIR( ) entries.
...
The ctest subdirs command now checks the relative path first, and if that does not exist, also checks if the given path was absolute. Thanks vodall for the patch.
2009-11-09 14:07:36 -05:00
Zach Mullen
9f105a3f6d
Fix style warning in cmsystemtools, cosmetic change to ctest code
2009-11-05 11:37:47 -05:00
Zach Mullen
9aa4e134b5
Fixed infinite loop condition when a test's dependent test executable could not be found.
2009-11-05 11:03:07 -05:00
Zach Mullen
85912a3777
Added TARGET, CONFIGURATION, PROJECT_NAME, and FLAGS arguments to ctest_build command
2009-11-03 16:46:30 -05:00
Zach Mullen
d47ada823a
Added logic to check for cycles in the test dependency graph before any tests are run. Previously a cycle resulted in a segfault from stack overflow.
2009-11-02 12:52:50 -05:00
Zach Mullen
8612aa10b6
Hook for scheduling tests in a random order
...
This may help statistically detect implicit dependencies among unit
tests while running in parallel.
2009-10-29 15:30:12 -04:00
Bill Hoffman
d7a3277fb3
Fix bug#9767 catch missing leak.
2009-10-29 08:38:04 -04:00
Zach Mullen
e2b63e473d
Fix working dir issue for ctest show only mode (-N)
2009-10-27 10:16:01 -04:00
Zach Mullen
99697308f3
Add a ctest test to cover cmCTestBatchTestHandler. Since its behavior isn't fully defined yet, we just run it as is and don't care much about the result.
2009-10-26 09:44:08 -04:00
Zach Mullen
f2010871ba
Ctest was broken for subdirs. Restored working directory state for tests so that their executables could be found.
2009-10-20 10:54:02 -04:00
Zach Mullen
727fa7bcb6
Fix ctest output alignment for cases where total tests run is not the same width as max test index.
2009-10-15 12:02:22 -04:00
David Cole
85feea2d0d
Fix floating point comparison warnings. Thanks to Alex Neundorf for the patch.
2009-10-13 16:39:48 -04:00
David Cole
4439f17b20
Fix issue #5668 - use CollapseFullPath when determining if covered file is within source or binary tree. Allows gcc/gcov coverage analysis using MinGW on Windows.
2009-10-12 14:51:54 -04:00
Zach Mullen
7fd3d7cf80
BUG: #0009648 Change "The following tests FAILED" message to print on stdout rather than stderr
2009-10-05 12:47:09 -04:00
Zach Mullen
b304186bc5
Match width of ctest "Start xx: " line to line up with the end test line
2009-10-05 10:20:52 -04:00
Alexander Neundorf
b3040beb41
fix two more warnings from icpc
...
asfGlob and abfGlob were there since rev 1.3 of this file (Oct 17th 2004) and unused since
then, so remove them
Alex
2009-10-04 11:40:02 -04:00
David Cole
ccb0cf1306
Fix warnings in CMake source code.
2009-10-02 15:30:01 -04:00
David Cole
44bcba7461
Fix warnings in CMake source code. Suppress rampant warnings emanating from Qt files.
2009-10-01 16:47:08 -04:00
Zach Mullen
a5be445ca9
BUG: CTest should honor test timeouts.
2009-09-29 14:31:58 -04:00
Zach Mullen
202bb45695
BUG: 0009612: --output-on-failure option doesn't work with the new parallel CTest handler
2009-09-29 09:45:43 -04:00
Brad King
96afb12087
Convert CMake to OSI-approved BSD License
...
This converts the CMake license to a pure 3-clause OSI-approved BSD
License. We drop the previous license clause requiring modified
versions to be plainly marked. We also update the CMake copyright to
cover the full development time range.
2009-09-28 11:43:28 -04:00
Zach Mullen
5e06f0a557
Fixed uninitialized memory issue in ctest show-only mode
2009-09-25 11:45:19 -04:00
Zach Mullen
5a2f24856e
Output a message in ctest when each test is started
2009-09-24 13:49:20 -04:00
Zach Mullen
65c418e56f
Reformat ctest -N output. Removed the "Start processing tests" message as well.
2009-09-24 09:49:46 -04:00
Zach Mullen
84f3f6cad6
Tests which are not run should be added to the failed test list.
2009-09-23 15:37:50 -04:00
Zach Mullen
4342e37d63
More SLURM experimentation (ctest batch mode)
2009-09-21 15:29:34 -04:00
Zach Mullen
a4c6f178fe
Fixed a slurm batch argument identifier.
2009-09-21 14:58:30 -04:00
Zach Mullen
87b4c66ca8
Need to quote args when generating batch scripts from ctest
2009-09-21 14:21:41 -04:00
Zach Mullen
e4293b4b38
Re-enabled failing tests; fixed ctest_build output to be consistent in the error condition across different make programs so that these tests would pass.
2009-09-21 13:40:40 -04:00
Bill Hoffman
932dac93b7
Fix uninitialized errors.
2009-09-16 11:49:09 -04:00
Zach Mullen
243bbae954
Removed fork-and-continue option from ctest generated batch script entries
2009-09-14 11:23:20 -04:00
Brad King
bcbb4626a3
Avoid shadowing std::vector member
...
The cmProcess::Buffer class derives from std::vector. We were using
local variable 'data' in the GetLine method but this name shadowed a
member of vector with GNU. This renames it to 'text'.
2009-09-12 10:20:00 -04:00
Bill Hoffman
63a7d85249
Fix for bug#9442, ctest crash if CTEST_SOURCE_DIRECTORY was not set.
2009-09-11 17:15:02 -04:00
Brad King
8c24c1dcc3
Fix new CTest output handling for no newline
...
When we clear the buffer for an output pipe after returning the last
partial line (without a newline) we need to set the partial line range
to empty. Otherwise the buffer object is left in an inconsistent state.
2009-09-11 16:20:24 -04:00
Bill Hoffman
8a690289c2
Add label summary times to ctest default output. Also, remove parallel time output. Add flag to disable label summary.
2009-09-11 13:34:35 -04:00
Brad King
6a7eae7184
Rewrite CTest child output handling
...
This commit fixes cmCTestRunTest and cmProcess to more efficiently
handle child output. We now use the buffer for each child output pipe
to hold at most a partial line plus one new block of data at a time.
All complete lines are scanned in-place, and then only the partial line
at the end of the buffer is moved back to the beginning before appending
new data.
We also simplify the cmProcess interface by making GetNextOutputLine the
only method that needs to be called while the process is running. This
simplifies cmCTestRunTest so that CheckOutput can be called until it
returns false when the process is done.
2009-09-11 12:26:41 -04:00
Brad King
b6c26cded2
Initialize cmCTestRunTest instances robustly
...
All instances of this class need a cmCTestTestHandler, so we now require
one to construct it. The instance also provides the cmCTest instance
too.
2009-09-11 10:09:48 -04:00
Brad King
b41a548d86
Add parentheses around '&&' between '||' for gcc
...
The GNU compiler warns about possible operator precedence mistakes and
asks for explicit parentheses (-Wparentheses). We add the parentheses
to silence the warning. This also fixes one real logic error in the
find_package() implementation by correcting expression evaluation order.
2009-09-11 08:18:15 -04:00
Zach Mullen
c57fb18920
Added some ctest batch capabilities
2009-09-10 11:18:05 -04:00
Zach Mullen
4b4e801eba
BUG: Fixed segfault and bad reporting if a ctest executable could not be found. Also added some batch testing code that is not yet complete.
2009-09-10 11:16:08 -04:00
Zach Mullen
a516040579
ENH: ctest now writes time cost data to a file after a test set is run, and uses these time costs to schedule the processes the next time ctest is run in that build tree.
2009-09-08 17:10:35 -04:00
Zach Mullen
6727834a8c
BUG: Fixed extraneous newlines from ctest process output
2009-09-08 14:48:23 -04:00
Zach Mullen
39e5f9d963
ENH: Replaced the EXPENSIVE test property with a COST test property taking a floating point value. Tests are now started in descending order of their cost, which defaults to 0 if none is specified.
2009-09-08 13:39:13 -04:00
Zach Mullen
a02ef56401
BUG: Fixed issue where ctest would hang if a process terminated with output in its buffers but no newline
2009-09-08 10:16:16 -04:00
Zach Mullen
384f4d1f3f
Fixed warning
2009-09-08 09:12:44 -04:00
Zach Mullen
5fb958fde9
ENH: Added ctest test options PROCESSORS and RUN_SERIAL. These allow specification of resource allocation for given tests running with the ctest -j N option. RUN_SERIAL ensures that a given test does not run in parallel with any other test. Also forced appending of "..." to the longest test name in ctest.
2009-09-07 10:26:17 -04:00
Bill Hoffman
cdb966532b
Fix memory and process leak in ctest_run_script.
2009-09-04 17:01:47 -04:00
David Cole
68b8332088
Increase curl submit timeout. A submit will timeout if there are 120 seconds of very little activity. 30 seconds was too short.
2009-09-04 14:00:26 -04:00
Zach Mullen
5517e17bf9
Fixed ctest output processing. Should now display output as it occurs, as well as be able to consume multiple lines if they exist within the timeout.
2009-09-04 13:50:06 -04:00
Bill Hoffman
7d190a65ca
Change run_ctest_script in ctest to not stop processing when there is an error in the script being run. Also, add a RETURN_VALUE option so that you can find out if the script failed
2009-09-04 13:24:25 -04:00
Zach Mullen
7f9a0f508f
Fixed output as-it-happens issue. Now displays output as it receives each newline.
2009-09-04 11:23:38 -04:00
Zach Mullen
c6e5dd21fd
Added the test property EXPENSIVE, which denotes that the given test(s) should be started prior to tests that are not marked as such. Also fixed test dependencies, and a few uninitialized variables in cmProcess.
2009-09-04 10:16:06 -04:00
Zach Mullen
7f6ff73396
Fixed 2 unused variable warnings
2009-09-03 15:50:47 -04:00
Zach Mullen
d4adde13d7
Allowed tests to pull more than one line of output in their quantum. Fixed uninitialized variables in the case that the test process could not start.
2009-09-03 15:33:44 -04:00
Zach Mullen
20713ab89d
Fixed warnings
2009-09-03 11:14:13 -04:00
Zach Mullen
659171d1f7
ENH: Added PARALLEL_LEVEL option for ctest_memcheck(). Added PROCESSORS option to set_tests_properties (implementation to come).
2009-09-03 10:47:14 -04:00
Zach Mullen
c5cf9b0e5f
STYLE: line length
2009-09-02 16:32:43 -04:00
Zach Mullen
7e20db0224
ENH: Added PARALLEL_LEVEL option to ctest_test() command.
2009-09-02 12:35:42 -04:00
Zach Mullen
69fd641adb
Fixed ctest output where max test index is not the same width as the total number of tests. Also some preliminary changes for batching ctest jobs
2009-09-02 10:08:40 -04:00
David Cole
d741a6a57f
Add curl timeout options to the SubmitUsingHTTP method. They were only in the SubmitUsingFTP method.
2009-09-01 17:23:50 -04:00
Zach Mullen
90cc5c5e04
ENH: Improved test reporting output
2009-09-01 11:58:04 -04:00
Zach Mullen
6d0b6f2b9e
Fixed ctest_memcheck docs ( http://www.cmake.org/Bug/view.php?id=9242 )
2009-08-31 11:32:39 -04:00
Zach Mullen
59b34a6a1a
Fixed Dart time recording for ctest
2009-08-31 10:28:39 -04:00
Zach Mullen
5a5cc52230
Fixed conversion warning on 64 bit machines
2009-08-31 09:50:35 -04:00
Zach Mullen
387ba0c646
Fixed line length issue
2009-08-30 10:57:30 -04:00
Zach Mullen
b0b5ffcf5c
MemCheck should now work again in ctest
2009-08-28 15:08:03 -04:00
Zach Mullen
996bb51a1d
Replaced std::stringstream with cmOStringStream
2009-08-28 11:40:34 -04:00
Zach Mullen
b427d1985e
Added ctest -N test. Fixed ctest working directory bug. MemCheck fix coming soon...
2009-08-28 11:08:39 -04:00
Zach Mullen
177edc5ed1
Fixed ctest -N segfault issue. Further refactored ctest. Enabled failover for ctest
2009-08-27 10:37:30 -04:00
Zach Mullen
8ffd8d0a03
ENH: refactored ctest. All testing is now parallel. If no -j option is specified, defaults to a MP level of 1 (non parallel)
2009-08-26 12:09:06 -04:00
Zach Mullen
8c6aa445a2
Fixed line length over 80
2009-08-20 09:46:43 -04:00
Zach Mullen
a2ef34d344
Fixed overwriting of a previous change set
2009-08-19 09:24:55 -04:00
Zach Mullen
b9daa192af
ENH: Refactored CTest test execution code into an object
2009-08-19 08:58:36 -04:00
Bill Hoffman
c02c8c287c
Add test times to log file as well as the stdout.
2009-08-18 14:03:50 -04:00
Bill Hoffman
3c34d0440c
If labels are found on the tests, then print a time summary for all the tests run with each label.
2009-08-18 13:34:05 -04:00
Brad King
dd2cf7388b
Fix classification of externals in svn status
...
CTest runs 'svn status' to identify modified and conflicting files in
the working directory. This commit fixes the interpretation of the 'X'
status, which corresponds to svn eXternals. This status should be
ignored rather than treated as a local modification.
2009-08-12 08:06:57 -04:00
Bill Hoffman
a696fe8183
Output total time when using -j N
2009-08-11 22:02:49 -04:00
Bill Hoffman
a86df0251f
Fix failing test on release build for VS 10 cmSystemTools::GetLineFromStream crashes if the stream is not open in that case.
2009-08-11 16:25:56 -04:00
Brad King
a48f29c904
Set current directory variables in CTest scripts
...
The commit "Fix get_filename_component ABSOLUTE mode" broke the code
get_filename_component(cwd . ABSOLUTE)
because CTest scripts did not make cmMakefile::GetCurrentDirectory()
available. This commit fixes the problem by setting the proper
information on CTest script instances of cmMakefile.
This also makes CMAKE_CURRENT_SOURCE_DIR and CMAKE_CURRENT_BINARY_DIR
available to CTest scripts. They are set to the working directory at
script startup.
2009-07-31 09:19:19 -04:00
Brad King
e9d7ebb3ec
BUG: Do not double-initialize local generators
...
All global generator CreateLocalGenerator methods automatically
initialize the local generator instances with SetGlobalGenerator. In
several places we were calling SetGlobalGenerator again after receiving
the return value from CreateLocalGenerator. The double-initializations
leaked the resources allocated by the first call to SetGlobalGenerator.
This fix removes the unnecessary calls.
2009-07-28 10:47:02 -04:00
Bill Hoffman
40850e7f1f
ENH: add test times and a total time to the output of command line ctest
2009-07-27 14:17:38 -04:00
Bill Hoffman
a5b06dcbb6
ENH: fix warning
2009-07-19 13:40:38 -04:00
Bill Hoffman
eb5d0a7f61
ENH: fix warning
2009-07-17 16:15:11 -04:00
Bill Hoffman
3250cb3d3b
ENH: compute the max test name width based on the length of the tests
2009-07-17 10:06:38 -04:00
Alexander Neundorf
7048373296
STYLE: don't load CMakeDetermineSystem and CMakeSystemSpecific directly from
...
cmCTestScriptHandler, but have it load the new script CTestScriptMode.cmake
-> that makes it more flexible, also add a simple test that the system name
has been determined correctly
Alex
2009-07-12 04:51:51 -04:00
Alexander Neundorf
46f99c6565
STYLE: move the code for writing the initial cache into its own separate
...
function, makes the long ProcessHandler() a little bit shorter
Alex
2009-07-11 16:30:18 -04:00
Alexander Neundorf
a605bd5225
STYLE: rename InitCache to InitialCache, since it contains the contents for
...
the initial cache and is not e.g. a flag which shows whether the cache
should be initialized
Alex
2009-07-11 16:27:28 -04:00
Brad King
ea00bb990b
COMP: Fix cmCTestHG for old HP compiler
...
The compiler does not have a fully compliant std::string.
2009-07-10 13:08:54 -04:00
Brad King
d4d467dbd5
ENH: Teach CTest to handle Mercurial repositories
...
This creates cmCTestHG to drive CTest Update handling on hg-based work
trees. Currently we always update to the head of the remote tracking
branch (hg pull), so the nightly start time is ignored for Nightly
builds. A later change will address this.
See issue #7879 . Patch from Emmanuel Christophe. I modified the patch
slightly for code style, to finish up some parsing details, and to fix
the test.
2009-07-10 11:08:05 -04:00