The old code worked for "gcc -fsanitize=address", but the sanitizers
tinderbox builds with "clang -fsanitize=address".
Follow-up to commit f67b8901dd (test:
double the timeout in isDocumentLoaded() for sanitizers, 2021-11-03).
Signed-off-by: Miklos Vajna <vmiklos@collabora.com>
Change-Id: I4b802ec0d3d20516e3d58cc1c65d7e1422632b2b
The sanitizers tinderbox regularly times out, and it's good to know if
the timeout value is an agressive one or the load is really that slow.
Signed-off-by: Miklos Vajna <vmiklos@collabora.com>
Change-Id: I2903481ea9efde68aca3f4ec00bd6e7e2d3b840a
The .uno:Copy command has to get executed before
we can read the clipboard. Since .uno commands
are executed asynchronously, it possible to
invoke getClipboard before having the data
copied into it, thereby failing the test.
Here we wait for the first sign that the
.uno commands have been executed, although
we can't know if they have finished or not.
This improves the test reliability significantly.
Change-Id: I7da1e4a32092c05dfb533571ea50d5edbc07f554
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
TileCacheTests now completely uses
async WebSockets. It is both more
stable and faster (by about 20-25%).
Change-Id: I6da0fd12ddf671a7ad666561249b6fb1c5d32a01
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
SocketPoll is captured as weak_ptr into
WebSocketSession and a much better shutdown
support is now available. The new logic
can do async-shutdown after flushing and
will do sync-shutdown if SocketPoll is
no longer around.
Change-Id: Ia206cab58a13f20f7aeb3a6d8c57afee731c8231
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
ProductVersion used to be "6.4" with core.git distro/collabora/cp-6.4,
but core.git distro/collabora/co-2021 sets it to just "2021".
Signed-off-by: Miklos Vajna <vmiklos@collabora.com>
Change-Id: I08c1204c42bba21c8036975138e3d999b7357bdf
./test/helpers.hpp:211:1: error: non-void function does not return a value in all control paths [-Werror,-Wreturn-type]
Perhaps it could go further and even call LOK_ASSERT_FAIL().
Signed-off-by: Miklos Vajna <vmiklos@collabora.com>
Change-Id: I4f68f6d290bcb0a832ea71153d5f699d5366def9
With this, tests use SSL connections not just based
on the build settings (and whether or not config.h
is included in the test translation-unit), but also
based on the actual config, which can be changed
by the test in question.
Change-Id: I6cce5949e7785cee481fdc805a1804df6fdd8b8d
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This reduces the execution time of some
long-running tests and improves stability.
A full debug build of Core and Online
now finishes on my machine in under 3 minutes,
down from slightly over 4 minutes. A significant
saving when running the tests frequently and
repeatedly.
Change-Id: I66c1b06ec3c9e8ab90a30f3a15c03f8d560ed3c6
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
'bytes' can be smaller than the size of the buffer, so go with that size
instead.
Change-Id: I025b56379ee449fe33710b660d845fa2dd0865ee
Signed-off-by: Miklos Vajna <vmiklos@collabora.com>
The text message in the WS Close Frame is optional.
Here we actually copy the remaining data after parsing
the status code, not assuming it exists or that the
status code was exactly 2 bytes (although in a well-
behaving WS implementation it always will be 2 bytes).
Also, no good reason to wait a full second between
socket reads in getErrorCode. However a small sleep
seems sensible to avoid logging "got 0 bytes" or so.
Change-Id: I162dfb0dd3ebf615a3e8e237006883388b9ff494
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This moves the test log macros into a new home,
test/testlog.hpp, to avoid cycling dependencies.
Change-Id: Iacb80e813a64ff830fa18f63ec4de2535ee702b7
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
These helpers are for testing only. They should
ideally be moved to the test helpers, but because
of dependency on the FileDeleter in FileUtil they
remain in FileUtil.
Change-Id: I93c7e08823edec8f6a53419f0a6596f3255f23f9
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
The tokenizer(s) are more generic than the protocol
logic, and are used from contexts that don't involve
the protocol as such.
Change-Id: Ie8c256bf11a91e466bff794021f41603c9596a7f
More readable and typically more efficient.
Change-Id: I9bd5bfc91f4ac255bb8ae0987708fb8b56b398f8
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/95285
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Tested-by: Jenkins
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
LibreOffice core uses that, too, and we support an even more
restricted set of compilers.
Change-Id: I0d0e2c8608e323eb5ef0f35ee8c46d02ab49a745
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/92467
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Tor Lillqvist <tml@collabora.com>
This test sometimes failed on me just because the load was aborted due
to the timeout, doubling the limit makes the problem go away.
Change-Id: Ided724d67a513391b86836065d4e627581857a92
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/90771
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Because the new-style tests are intrustive,
the exception that CppUnit throws on assertion
failures is caught and processed with the
application logic, which is far from ideal,
because it's very difficult to find the
cause of failure.
What we'd like is a way to control what happens
when an test assertion fails, such that we can
properly log/print the failure, and even break
in the debugger.
The new macros allow us to control the behavior
at compile-time and have added flexibility.
For now, they log an assertion failure before
invoking the CPPUNIT macro, and support a
compile-time directive to assert, which is
useful for breaking in the debugger.
Change-Id: If464ba246e3ec747f31496a4215cb73ef735dfaf
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/87625
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
The bulk of this commit just changes std::vector<std::string> to
StringVector when we deal with tokens from a websocket message.
The less boring part of it is the new StringVector class, which is a
wrapper around std::vector<std::string>, and provides the same API,
except that operator[] returns a string, not a string&, and this allows
returning an empty string in case that prevents reading past the end of
the underlying array.
This means in case client code forgets to check size() before invoking
operator[], we don't crash. (See the ~3 previous commits which fixed
such crashes.)
Later the ctor could be changed to take a single underlying string to
avoid lots of tiny allocations, that's not yet done in this commit.
Change-Id: I8a6082143a8ac0b65824f574b32104d7889c184f
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/89687
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Logs from test are now more atomic (less interleaving).
And each integration-test (old-style tests running from
new-style tests) will now print clearly when it begins
and when it fails, making it easier to find failures.
Misc other minor cleanups in test logging.
Change-Id: Iff664e42a04d1c6dbf1332b2884c35183ef85e21
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/89309
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
