Unit-tests are only ever run in debug builds by
desing (there is check in test/run_unit.sh to
build with --enable-debug to run the tests).
In our release buids, we have seen cases where
the process exists with an error code because
the unit-test logic thinks it ran a test and failed.
It also generates error log and other log entries
when initializing and uninitializing the unit-
test logic.
This excludes unit-test logs and the failure
check in non-debug builds to avoid such errors
and reduce noise.
Change-Id: I9f472172b72770551a491ff6e19ab9b997fcf365
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
It appears that both cypress and the C++ WOPI tests are extremely
timing sensitive, and fixing this should not hold up merging the
logging improvements. So for now don't disable logging in these modes.
Signed-off-by: Michael Meeks <michael.meeks@collabora.com>
Change-Id: I43e8397263e6960b668a29e7ad67f45394c52b52
Reduce the uses of MOBILEAPP conditionals by using the isMobileApp
function.
Signed-off-by: Jaume Pujantell <jaume.pujantell@collabora.com>
Change-Id: If541307fbc457b342674cc560b6c53454f3904cf
We don't need a SocketPoll and its wake-pipes before we
run a test - obviously, and we should close it when we
fork a new test.
Signed-off-by: Michael Meeks <michael.meeks@collabora.com>
Change-Id: I56875b66ce2ba141baacdd70c973e9e9af02d1d0
We use [[maybe_unused]] to reduce explicit supression
especially under compile-time directives.
Change-Id: Ic5f3f3227a80efe52097cb35520d05b9cdaacb42
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Replace a few cases of Util::toString(...), too.
Signed-off-by: Aron Budea <aron.budea@collabora.com>
Change-Id: I53c5dbbdf2f60710e94add9dd137aa9becb7bd3e
Removed all uses of KIT_IN_PROCESS, used Utils::isKitInProcess()
when necessary. Also removed the now unused parameter 'limit' from
forkLibreOfficeKit.
Signed-off-by: Jaume Pujantell <jaume.pujantell@collabora.com>
Change-Id: I068d3f55ab49076590f111847c87b3188f4d25d0
We now exit immediately with error code when
we are given --unitlib argument but fail to
load the given library.
Change-Id: I4eb5f3a37b9838ecf83bdceef4d09f888afcca83
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Seems a useful thing to capture the internal
state when a test fails, as it might help
understand what was going on.
Change-Id: I591a88662ab86ab4dc7f39dafcc4d7d279a4d9c4
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Make sure that the logging subsystem is
already initialized before starting
unit-tests.
Change-Id: I2ce3ffdb2d3b0094ae7ed496d7cacfc02af89c21
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This prepares to start the next test
when the DocBroker of the current one
is destroyed.
Change-Id: I3a9672d8d03ad497f15a9bca262dfe1d4f197e41
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
We now upload 3 times to succeed.
This simulates a user modifying and saving
twice before giving up and closing the
document, which succeeds in uploading.
Change-Id: Ida40f79a76fc94436c0d4b4fc1ad3e7b7acc997b
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
The BaseUnit test didn't belong to UnitTimeout
and, more important, it needed access to
private members to both validate their state
and to reset them (since the test is artificially
initializing both WSD and Kit tests and cannot
uninitialize them, lest we unload ourselves).
As such, the self-test is now internal to
BaseUnit, with the added bonus that it
is called on all tests and not just UnitTimeout.
Also, more assertions have been added.
Change-Id: Ieaf60594f39e978a7250407262bd8bbc9b642c43
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
There is no reason to print the test name
explicitly when each log entry is already
prefixed by the test name.
Change-Id: I72c0a0f2989a65f477902b97e6be816446fba5c3
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Since UnitBase (and children) are invoked from
different threads, UnitBase::exitTest could
have a race. This ultimately results in
invoking std:🧵:join() concurrently,
which is unsafe. This can result in the
following deadlock.
In this case websrv_poll had invoked
UnitBase::exitTest concurrently and
evidently finished, leaving the main
coolwsd thread deadlocked.
(gdb) info thread
Id Target Id Frame
* 1 Thread 0x7f67528f0840 (LWP 4058508) "coolwsd" __futex_abstimed_wait_common64 (private=128, cancel=true, abstime=0x0, op=265, expected=4058510, futex_word=0x7f675285c910)
at ./nptl/futex-internal.c:57
(gdb) t a a bt
Thread 1 (Thread 0x7f67528f0840 (LWP 4058508) "coolwsd"):
#0 __futex_abstimed_wait_common64 (private=128, cancel=true, abstime=0x0, op=265, expected=4058510, futex_word=0x7f675285c910) at ./nptl/futex-internal.c:57
#1 __futex_abstimed_wait_common (cancel=true, private=128, abstime=0x0, clockid=0, expected=4058510, futex_word=0x7f675285c910) at ./nptl/futex-internal.c:87
#2 __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7f675285c910, expected=4058510, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=128) at ./nptl/futex-internal.c:139
#3 0x00007f67529cd6a4 in __pthread_clockjoin_ex (threadid=140081742857792, thread_return=0x0, clockid=0, abstime=0x0, block=<optimized out>) at ./nptl/pthread_join_common.c:105
#4 0x00007f6752d42337 in std:🧵:join() () from /lib/x86_64-linux-gnu/libstdc++.so.6
#5 0x000055775c5fa8e0 in SocketPoll::joinThread (this=0x55775cc9b090) at net/Socket.cpp:282
#6 0x000055775c5c8121 in UnitBase::endTest (this=this@entry=0x55775cc0fe80, reason="") at common/Unit.cpp:545
#7 0x00007f67528df61f in UnitWSDClient::endTest (this=0x55775cc0fe80, reason="") at ./UnitWSDClient.hpp:118
#8 0x000055775c5cc4d1 in UnitBase::exitTest (this=0x55775cc0fe80, result=result@entry=UnitBase::TestResult::Ok, reason="") at common/Unit.cpp:486
#9 0x00007f67528d1fa3 in UnitWOPIFileUrl::invokeWSDTest (this=0x55775cc0fe80) at UnitWOPIFileUrl.cpp:269
#10 0x000055775c4accdb in UnitWSD::invokeTest (this=0x55775cc0fe80) at ./common/Unit.hpp:318
#11 0x000055775c4a38cd in COOLWSD::innerMain (this=0x7ffd34248650) at wsd/COOLWSD.cpp:5459
#12 0x000055775c4a59db in COOLWSD::main (this=<optimized out>) at wsd/COOLWSD.cpp:5706
#13 0x00007f6753ab0877 in Poco::Util::Application::run() () from /lib/libPocoUtil.so.80
#14 0x000055775c48316c in main (argc=14, argv=0x7ffd34248898) at wsd/COOLWSD.cpp:5845
Change-Id: If6162690462719f14a35241604f8a59b38a806c0
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
These only served to delete the Unit singletons,
but since we have GlobalArray, deleting them
isn't needed.
Change-Id: Ia2e53a1e3b43707196f45c06c38c6bb664c18e29
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This is needed so we stop the DocBroker of
the old test, which is now finished.
Change-Id: I05df85fbbeca254928cd9a140198e7f4ae89fec8
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Apply the test filter on all tests,
not just the first.
Change-Id: Ib34e90e883f5de2ddcf04c710f433cb54677575d
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Now that we may have multiple tests in a
single unit-test module, we should call
UnitWSD::configure for each one, not just
once at the start. Each test can modify
the configuration settings to meet its
test conditions.
Change-Id: I08de2d998edee5b4cbade08f2496196b596a8600
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
When invoking exitTest it's useful to know
something about the reason. Whether it's
because the test passed or timed out,
it's useful to log this information.
Change-Id: I8e8ae525b7c2ef1de2e394d08107341cdc87c008
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Now that multiple tests and filtering are
supported, it's useful to know what tests
exist in each unit-test module.
Change-Id: Ifaa7eccabab6b23a57edf2786c0cd726ab0265db
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Using the envar COOL_TEST_OPTIONS, which is
provisioned to support other options than
just filtering, it is now possible to
only run the tests with names partially
matching the filter's case-insensitive
string.
Change-Id: I6168887677fd0fddd72bac2538837615b10b9fae
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Since we can have multiple tests in a single
unit-test .so, we need to merge the results
of each test into a single result that is
used when we exit the process and flag for
failure or success.
Change-Id: Ia49f5542efd6dd998b61c17965fd3fe3245159ba
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Since we no longer need to set the handle,
this function is an initialization of the
given test, before running it.
Change-Id: Ia6fb16b06dd548977fcfcc35a81ac37d979be29d
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This adds support for multiple tests
within a single unit-test .so.
We did have many, many tests within
the converted "old style" tests already.
Except, they couldn't support the proper
event-driven tests, nor were they running
as fast as possible. They also blocked the
main thread in COOLWSD.
With this support, we can now avoid
creating a new .so for each test case.
This will reduce the effort to add new
tests and the build time. In addition,
the execution of each unit-test is
dominated by the initialization, which
takes in the order of 5 seconds. Running
multiple tests after this high load time
reduces the per-test average time.
Later we will support listing tests,
running particular tests (i.e. filtering)
and other debugging and troubleshooting
facilities.
Change-Id: I7be43994265ebabdb5d379f7726bb9bd7f78795a
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
We now properly cleanup unit-tests .so libs.
This helps having multiple tests by making
the dlhandle an independent variable of
all the tests.
Change-Id: Ifc8db400aebfe8ea41dcc4df2d11856c25b36074
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This adds support to having multiple tests in
a single unit-test .so.
Change-Id: I37526484e8a692b25393f659dedd846eedc22733
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Normally, we should exit a test with a single result:
either a success or a failure. It is, however,
possible to have an early success that forgoes
proper cleanups and cause subsequent failures.
Here, we at least log these cases, because it
could unintentionally hide real failures.
Change-Id: I17e6365bd54b3f7336976fb51437776811733eb8
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
