Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
|
|
|
|
modules.
|
|
|
|
|
|
process, just use os:type().
|
|
|
|
|
|
|
|
* peppe/common_test/dev/OTP-8805:
Fix problem with {fail,Result} from end_tc not being recognized
Fix problem with prepared_tests tuple not being recognized
Add verification terms to test suite
Fix various test suite, makefile, and doc related problems
Fix error in the repeat_until_* group property handling
|
|
* peppe/common_test/dev/OTP-8753:
Fix problem with prepared_tests tuple not being recognized
Add verification terms to test suite
Fix various test suite, makefile, and doc related problems
Fix error in the repeat_until_* group property handling
Add test suite for groups with repeat property
Add verification events to sequence group test suite
Fix error in handling of sequence groups
Fix error in test suite
Fix problems with sequences and sub-groups
Add test cases to sequence suite
Fix error with group term in ct:run_test/1
Add test suite to examplify problem with nested groups in a sequence
|
|
|
|
|
|
|
|
|
|
Getting sequences to work properly when a sub-group fails in init_per_group has been unsuccessful (see tests in ct_sequence_1_SUITE). Skipping parent group tests from the auto_skip conf clause fails because of problems with where/how the group status is stored.
Work in progress...
|
|
Only real program files should be installed with INSTALL_PROGRAM while
scripts should be installed with INSTALL_SCRIPT and data files with
INSTALL_DATA.
|
|
|
|
ts_lib:erlang_type/0 used to dectect whether the system was
run in Clearcase, in an uninstalled source tree or in an
installed system.
There now only remain two cases: running in a source
tree/git repository or in an installed system. Update
ts_lib:erlang_type/0 and all callers accordingly.
|
|
|
|
Still a few things missing:
* Documentation
* More tests
* Fix old broken test cases
|
|
The overview html log files now have a textured background. This default new look can be switched off with the 'basic_html' option.
|
|
It's possible now to specify execution of test case groups (and test cases within groups) by means of run_test options and test specification terms. From test specifications, groups (and test cases) may also be skipped.
ct_run has been improved to report errors better.
New test suites have been introduced.
|
|
This is to enable execution of specific groups, and/or test cases within groups, by means of run_test options or test specifications.
|
|
|
|
|
|
To be consistent with the behaviour of the run_test repeat flag/option, the repeat* group property has been changed to specify absolute number of test runs. Previously {repeat,N} meant "execute the group 1 time + N repeats". Now it means "execute the group N times".
|
|
Directories added with run_test -pa or -pz in the pre erl_args part of the command line will be converted from relative to absolute, this to avoid confusion when Common Test switches working directory during the test run.
Also fixed in this commit: Problem with timeouts during init_tc or end_tc (in the Test Server framework).
|
|
|
|
|
|
Also fixed bug in test_server that calls end_per_testcase after test case timeout or abortion.
|
|
|
|
Also fixed bug: return value {fail,Reason} from end_tc(init_per_suite) was ignored.
|
|
|
|
|
|
ct_run:script_start() interface
The possibility to pass start arguments to ct_run:start_script/0 by means of an application environment variable has been implemented. This will be used by ct_test_support for automatic testing of all common_test start interfaces.
|
|
Documentation still missing.
|
|
|
|
Some editors get confused by illegal break character.
|
|
* bg/opt-receive:
Test that gen_server:call/2,3 are fast even with a huge message queue
erts: Add tests for the receive optimization
Update primary bootstrap
erts: Implement recv_mark/1 and recv_set/1 for real
compiler tests: Cover the error handling code in beam_receive
compiler test: Test optimization of receive statements
Optimize selective receives in the presence of a large message queue
Introduce the new recv_mark/1 and recv_mark/1 instructions
Compile tests that communicate with R12 nodes with the r12 option
Move p_run/2 to test_lib
gen: Inline wait_resp_mon/2 to help the compiler optimize
OTP-8623 bg/opt-receive
reveive statements that can only read out a newly created reference are now
specially optimized so that it will execute in constant time regardless of
the number of messages in the receive queue for the process. That
optimization will benefit calls to gen_server:call(). (See gen:do_call/4
for an example of a receive statement that will be optimized.)
|
|
|
|
R12 nodes cannot load code that use the optimized receive that
we are about to implement.
|
|
* bg/remove-stray-ose-support:
configure: Remove stray OSE/Delta support
Makefiles: Remove stray OSE/Delta support
kernel tests: Remove stray OSE/Delta support
system tests: Remove stray OSE/Delta support
erl_interface tests: Remove stray OSE/Delta support
epmd: Remove stray OSE/Delta support
epmd: #ifdef out start_epmd() for other platforms than VxWorks
emulator tests: Remove stray OSE/Delta support
emulator: Remove stray OSE/Delta support
emulator: Eliminate #ifdef for sys_tty_reset()
test_server: Remove stray support for OSE/Delta
OTP-8585 bg/remove-stray-ose-support
|
|
Remove code supporting testing on OSE/Delta.
Some cross-testing support only used by OSE/Delta is kept
(see the mention of OSE in test_server_internal.hrl), because
it could presumably be useful in the future if we are to test
some other embedded system.
|
|
* bg/ts_run:
ts_run: Don't run make:all/1 without a good reason
OTP-8556 bg/ts_run
|
|
ts:run() runs make:all/1 to facilitate interactive development
of a test suite - when you have changed your test suite, just
run ts:run() and it will rebuild the test suite for you.
Some test suites, such as for the public_key application,
are meant to be built once and for all by "make release_tests".
Re-building it later with make:all/1 will fail because
it references an internal .hrl file in the source directory
for public_key.
Running make:all/1 for the public_key test suite would not
cause any harm in most circumstances, because make:all/1 would
consider the *.beam files to be up-to-date and therefore not
actually try to re-build anything.
In 2c6d9b57ce4557a431bcf02565c3634a0ed7ca61, however, the
public_key test suite started to include "test_server.hrl"
using -include_lib() instead of -include(). That means that
"test_server.hrl" will be included from the Erlang/OTP
system being tested, and if that system has been built
*after* the test suite was built, the *.beam files in the
test suite will be out-of-date and make:all/1 will try to
re-build them (and fail).
To avoid this problem, modify ts:run() as follows:
* If the test suite has an Emakefile (such as the emulator test
suite), it is clearly meant to be (re-)built by make:all/1,
so we run make:all/1.
* If there is no Emakefile and no .beam files (such as in the
erl_interface test suite), it also seems clear that the
test suite is meant to be built by make:all/1. To make sure
that the test suite will also be *re-built*, create an
Emakefile that will re-build all *.erl files in the
directory for the test suite.
* If there is no Emakefile and there are *.beam files (such
as for the public_key test suite), we assume that the test
suite are not meant to re-built. However, there may be
generated *_SUITE_make.erl files (if there are *_SUITE_data
directories containg Makefile.src files, as in the xmerl
application). If there are any *_SUITE_make.erl files,
generate an Emakefile for compiling only those files
and run make:all/1. Otherwise, don't run make:all/1.
|
|
* bg/compiler-remove-r11-support:
compiler: Don't support the no_binaries option
erts: Don't support the put_string/3 instruction
compiler: Don't support the no_constant_pool option
compiler: Don't support the r11 option
test_server: Don't support communication with R11 nodes
binary_SUITE: Don't test bit-level binary roundtrips with R11 nodes
erts: Test compatibility of funs with R12 instead of R11
OTP-8531 bg/compiler-remove-r11-support
|