| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
| |
Should hopefully reduce random failures of timing-dependent tests which
rely on subsecond precision for accuracy.
(cherry picked from commit c845b2e39fffeb560983a301d810616a6495469d)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Ideally timers would be perfectly accurate and we could sleep for 1.0 seconds,
but on OSX in particular they could fire 50-100 ms late (I've seen 4 ms in
practice) and we don't want the tests to fail because of this, because we don't
really need that kind of precision in practice. So we reduce the timer
intervals by 100ms to be safe.
Increase logging level and add timestamps in testcommon. Should help to debug
frequent test failures such as Condition [t2.HasExpired()] on OSX hosts on
Travis.
Add a SettingsGuard to the Console logger for its specific settings, such as
showing times and microseconds.
(cherry picked from commit a5e6cff1d435329b0121417ed9509e315ce0edd5)
(cherry picked from commit 8d02eebce553ed822e0fcd60d6e319384e15ba4b)
|
| |
|
|
|
|
|
|
| |
Also increase test verbosity to help debug the test if it fails again.
(cherry picked from commit 00c2127e9832591c248fa6eea05dfaf785b8380d)
|
| |
|
|
|
|
| |
Hopefully this will fix intermittent test failures on AppVeyor.
|
|
|
|
|
|
| |
Deleting random files tends to cause random test failures. The results are not
predictable and depend on the order in which files are uploaded and directories
created.
|
|
|
|
|
| |
Detect when housekeeping has run and account is unlocked, and stop the loop
early, so that we don't have to wait for 32 iterations of the loop.
|
| |
|
|
|
|
|
| |
Fixes test failures on IPv6 enabled hosts where localhost can resolve to an
IPv6 address, but the server is only bound to the IPv4 localhost address.
|
|
|
|
|
|
| |
Define WIN32 properly when building with CMake, and in this case, use native
commands such as attrib.exe instead of Cygwin chmod, so that users don't have
to install Cygwin or msys-git.
|
| |
|
| |
|
| |
|
|
|
|
|
| |
On Windows there are no symlinks, so there should be 3 fewer files in the directory,
and thus 6 fewer keepalives sent.
|
|
|
|
|
| |
Should fix some problems with running the perl scripts, causing test failures
on Windows.
|
|
|
|
|
|
|
|
| |
Previously, files were uploaded using bbackupd, which did it in the order that
it read the directories, which was OS-dependent. So destroying certain fixed
RaidFiles would have an unpredictable effect on the store. Now, we write a manifest
when creating the test files, and upload them in exactly that order, so that the
ID of each file will always be the same.
|
| |
|
|
|
|
|
|
| |
pure C++.
Allows it to work with Visual Studio and no Cygwin/MSYS shell.
|
| |
|
| |
|
| |
|
|
|
|
|
| |
Avoid using the rm command, and use the correct name for the test executables
compiled by our cmake alternate build system.
|
|
|
|
|
| |
Perl on Windows tends to use IPv6 for localhost if possible, but the server binds
to IPv4, breaking the test.
|
|
|
|
|
| |
StopServer() does a hard terminate on Windows, which can leave temporary files
in place, causing problems later in the test.
|
|
|
|
|
| |
Con is a special filename on Windows and you just can't create a file with
that name.
|
|
|
|
|
| |
Otherwise when the counts differ later, we won't know whether we did the setup
properly in the first place or not.
|
|
|
|
|
| |
I think something is on the PATH in Appveyor that makes rmdir /s/q fail,
especially when run through Perl.
|
|
|
|
| |
Break dependendency of test/bbackupd on individual files from other modules.
|
|
|
|
| |
Fixes test/backupstorefix.
|
|
|
|
| |
wrong
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|\
| |
| |
| | |
Merge branch 'master' of https://github.com/agrandville/boxbackup into appveyor
|
| |\ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This test has always been fragile, and usually fails on Travis for reasons
unknown, so replace the log parsing and intercepts with simple mocking code.
This doesn't cover 100% of what the old test did, including diff timer aborts
and the number of blocks used by the uploaded files, but it should be
completely robust.
|
| | |
| | |
| | |
| | |
| | | |
"context" might easily be used in a test function, and doesn't follow our
naming convention anyway, so renamed to sTlsContext.
|
| |/
| |
| |
| |
| |
| |
| |
| |
| | |
See discussion of incompatible struct packing on ARM processors:
http://lists.boxbackup.org/pipermail/boxbackup/2010-November/005818.html
http://lists.boxbackup.org/pipermail/boxbackup/2011-February/005978.html
Thanks to Leif Linderstam for identifying the problem and proposing a
solution. This is just a test for the problem, not a fix in itself.
|
|/ |
|
|
|
|
| |
Switch httpserver daemon start/stop to use standard functions.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a new exception code to represent an object being completely missing (not
found on the store at all), separate from not being found in a particular
directory.
Improve mapping of server-side exceptions to protocol error messages returned
to the client.
Add handling for missing exceptions, such as
BackupStoreException::PatchChainInfoBadInDirectory, and the new
BackupStoreException::ObjectDoesNotExist.
Fix mapping for BackupStoreException::CouldNotFindEntryInDirectory to make it
distinguistable from BackupStoreException::ObjectDoesNotExist.
|
| |
|
|
|
|
|
| |
Stop the client while waiting for housekeeping, to ensure that it doesn't
prevent housekeeping from running and cause the test to fail.
|
| |
|
|
|
|
|
| |
Also checks the returned error code, and gives more useful diagnostics on
failure.
|