summaryrefslogtreecommitdiff
path: root/selftest/selftest.pl
AgeCommit message (Collapse)AuthorFilesLines
2010-11-21selftest: Even if setting up an environment failed, the next tests mayJelmer Vernooij1-1/+0
still work.
2010-11-01s4-test: we don't need to set LDB_MODULES_PATH any moreAndrew Tridgell1-1/+0
2010-10-30s4-test: don't override modules dirAndrew Tridgell1-1/+0
it is set correctly during the build
2010-10-30s4-modules: remove LD_SAMBA_MODULE_PATHAndrew Tridgell1-1/+1
this isn't needed now that we automatically get the module path right in our build and install trees Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
2010-10-19selftest: terminate selftest if we can't setup the environmentAndrew Tridgell1-2/+3
2010-10-18selftest: correct the import of Time::HighResMatthieu Patou1-2/+3
The "use" is always evaluated first even if enclosed into a if. So on plateform without Time::Hires (ie. sun8 in the build farm). This leads to an error. Instead we use the synthax require Module + Module->import("function")
2010-10-02s4-selftest: support 'make testenv SCREEN=1'Andrew Tridgell1-2/+3
this can be used to start a test envioronment in screen windows. Use: make testenv SCREEN=1 SELFTEST_TESTENV=dc to launch just one environment (in this case, "dc")
2010-10-01selftest: Let selftest provide the tempdir, rather than creating it as ↵Jelmer Vernooij1-0/+4
sideeffect of tests.py.
2010-09-28selftest: Abort early on SIGPIPE.Jelmer Vernooij1-0/+8
2010-09-25selftest: If a testsuite is killed by a signal, let selftest die.Jelmer Vernooij1-2/+2
2010-09-24selftest Don't run 'speed' tests for very longAndrew Bartlett1-0/+2
'make test' is too long, and the main thing we need with these tests is to ensure they don't segfault - there is no need to benchmark every box in the build farm, and we have no 'fail' metric in any case. Andrew Bartlett
2010-09-23selftest: Pass prefix to setup_env function.Jelmer Vernooij1-4/+4
2010-09-23selftest: Use variable for clientdir.Jelmer Vernooij1-18/+20
2010-09-22selftest: Fix testsuite counts when running only a limited number ofJelmer Vernooij1-8/+15
testsuites.
2010-09-22selftest: Fix idlist running.Jelmer Vernooij1-2/+2
2010-09-22selftest: Fix use of --load-list.Jelmer Vernooij1-1/+1
2010-09-22selftest: Remove unused --analyse argument.Jelmer Vernooij1-6/+0
2010-09-22selftest: Let tests add prefix to tests by themselve.Jelmer Vernooij1-23/+12
2010-09-21selftest: Use idlist testing for python tests.Jelmer Vernooij1-2/+3
2010-09-21selftest: Support running individual tests using idlists, for testsuites ↵Jelmer Vernooij1-7/+12
that support them.
2010-09-21selftest: Fix generation of test id file.Jelmer Vernooij1-4/+4
2010-09-14selftest: If setting up environment fails, mark testsuites that use it asJelmer Vernooij1-12/+5
errorring, don't skip it.
2010-09-14subunit.pm: Fold Subunit::Filter into Subunit, trim further.Jelmer Vernooij1-5/+2
2010-09-14subunit.pm: Simplify subunit handling in perl.Jelmer Vernooij1-12/+4
2010-09-14subunit.pm: Pass through milliseconds in time reports.Jelmer Vernooij1-0/+7
2010-09-05selftest: Cope with parentheses in testnames in --load-list.Jelmer Vernooij1-5/+8
2010-09-05selftest: Print out unmatched tests when using --load-list.Jelmer Vernooij1-7/+12
2010-09-04testr: Fix configuration to use idfile, in case more than a couple ofJelmer Vernooij1-1/+1
hundred tests fail.
2010-09-04selftest: Don't make printing the log output a side-effect of check_env(), ↵Jelmer Vernooij1-0/+1
do it manually.
2010-09-04selftest: Remove stop() method, not used anywhere.Jelmer Vernooij1-3/+1
2010-09-04selftest: run client tests with log level 1Andrew Tridgell1-0/+1
this PIDL print functions to display
2010-06-25selftest: Clarify generation of idlist option.Jelmer Vernooij1-0/+2
2010-06-17selftest: use 127.0.0.11-16 for the clientStefan Metzmacher1-7/+7
This gives more room for servers. Note: socket_wrapper currently only supports 127.0.0.1-16 see MAX_WRAPPED_INTERFACES. metze
2010-06-16s4:selftest Add vampire_dc test environmentAndrew Bartlett1-2/+8
This starts a domain controller, after running 'net vampire' to populate it. We don't use it as the 'all' environment yet, as the rest of the code isn't quite ready to handle it. Andrew Bartlett
2010-06-02s3-selftest: Fix blackbox smbclient s3 tests during make selftest.Günther Deschner1-1/+3
Guenther
2010-04-12selftest: Cope with strict.Jelmer Vernooij1-1/+1
2010-04-10selftest: Support passing --load-list to test runners.Jelmer Vernooij1-5/+15
2010-04-08selftest: Support --load-list option.Jelmer Vernooij1-7/+39
2010-03-31smbtorture: Report number of available tests in a testsuite using subunit.Jelmer Vernooij1-1/+1
2010-03-31selftest: Use standard subunit command for progress reporting.Jelmer Vernooij1-5/+4
2010-03-31selftest: Use standard subunit commands for progress reporting.Jelmer Vernooij1-0/+5
2010-03-12selftest: Remove dns_host_file every time we startAndrew Bartlett1-0/+3
2010-03-11selftest: let the client also use the dns_host_fileStefan Metzmacher1-1/+2
metze
2010-02-19s4:selftest Add infrastructure for testing against an RPC proxyAndrew Bartlett1-1/+13
This also changes the 'testenv' code to use a new environment 'all' (we may wish to make other complex tests depend on this in future), and exports more names in more namespaces. Andrew Bartlett
2010-01-19s3-selftest: use absolute path when setting SOCKET_WRAPPER_DIR.Günther Deschner1-1/+1
Guenther
2010-01-11s4-selftest: when a command fails show both normal and expanded commandAndrew Tridgell1-0/+12
It is sometimes hard to tell which varient of something like $SMB_CONF_PATH or $USERNAME is being used in a test. By giving both the expanded command ($command with environment variables expanded) and non-expanded command it is easier to reproduce bugs outside the test environment.
2010-01-11s4-test: fixed make test without having done make installAndrew Tridgell1-0/+1
client.conf didn't specify "setup directory"
2009-11-17selftest: make sure we don't buffer any outputStefan Metzmacher1-0/+3
metze
2009-10-22selftest: make python run unbufferedAndrew Tridgell1-0/+3
This makes some output in make test easier to follow
2009-10-19selftest: try to get the valgrind errors showing again in the build farmAndrew Tridgell1-6/+3