| Commit message (Collapse) | Author | Age |
|
|
|
|
| |
Allows tests to depend on lib/bbackupd instead of bin/bbackupd, which was
always a hack, and really doesn't work with CMake.
|
|
|
|
| |
Was due to accessing the exception's string after it had gone out of scope.
|
| |
|
|
|
|
| |
Allows it to be used by S3Store, or in fact anything that uses Configuration.
|
|
|
|
| |
S3Stores don't have account numbers.
|
|
|
|
|
|
| |
Make some keys not required in BackupDaemonConfigVerify, and verify them
afterwards instead, when the rest of the configuration file has been read
(and in future, the store type would be known.)
|
|
|
|
|
|
|
|
|
| |
Thanks to Kai Liebenau for finding and fixing this bug.
Also use VssFreeSnapshotProperties instead of rolling our own, also thanks to
Kai.
This has not been compile tested on Windows.
|
|
|
|
| |
Thanks to Achim for the patch! https://www.boxbackup.org/ticket/67
|
| |
|
| |
|
| |
|
|
|
|
|
| |
And compile errors on recent MinGW.
|
|
|
|
|
|
|
|
| |
Helps to debug tests that depend on subsecond timing.
Log time to wait on command socket in fractional seconds instead of raw
box_time_t.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
Add a LogFileOverwrite option to allow configuring whether LogFile is
overwritten or appended to. Overwriting happens at startup, and the start
of each backup in bbackupd (via ResetLogFile()).
|
|
|
|
| |
Prevents backup log files from becoming too large.
|
|
|
|
|
| |
Some tests need to do this. I don't think it happens in normal use,
because the daemon's response to HUP takes a different path, I think.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The std::auto_ptr is now returned unclosed. If the caller of RunSyncNow()
does nothing with the return value, then it will be released, closing the
connection. However the caller can save the std::auto_ptr locally and
poke around in it if necessary. This interface is designed for tests only!
Merged back changes from the test refactor branch to reduce diffs.
Fix failure to record client store marker, after BackupClientContext refactor.
This caused the daemon not to abort safely if the CSM changed under our feet,
meaning that our cached data is invalid.
Fix for earlier patch to allow tests to access the BackupClientContext after
backup completes, because a new CSM wasn't generated in time for bbackupd to
record it, because the connection wasn't closed yet.
|
|
|
|
|
|
|
|
|
|
| |
Allows it not to throw an exception if timers weren't initialised when
cleanup was requested. Normally we want an exception thrown, but not while
we're cleaning up a test that might have failed with timers uninitialised.
More timers fixes after cleanup no-exception option.
Merged back changes from the test refactor branch to reduce diffs.
|
|
|
|
|
|
| |
Hopefully all of these changes are inconsequential.
Merged back changes from the test refactor branch to reduce diffs.
|
|
|
|
|
|
|
| |
For some reason the timing changed and the test was no longer waiting for the
same length of time as bbackupd, causing it to fail.
Merged back changes from the test refactor branch to reduce diffs.
|
|
|
|
| |
Avoid reporting an error if the file is missing or empty.
|
| |
|
|
|
|
|
| |
Allows us to remove code to handle cleanups after exceptions and on Location
destruction.
|
|
|
|
|
|
|
|
| |
Conflicts:
bin/bbackupd/BackupDaemon.cpp
bin/bbackupd/BackupDaemon.h
test/bbackupd/testbbackupd.cpp
|
|
|
|
| |
Was accidentally removed by the refactoring of the backup idle loop.
|
| |
|
| |
|
|
|
|
|
| |
Otherwise it interrupts us needlessly while we're sleeping waiting for
the next backup.
|
| |
|
| |
|
|
|
|
| |
immediately after the end of the current one.
|
|
|
|
|
|
|
|
| |
Add helpful debugging messages about why the decision was taken to sleep
or start a backup, and how long the system will wait until the next
backup.
Flatten the code to avoid the double loop.
|
|
|
|
| |
Caused an assertion failure and crash.
|
|
|
|
| |
Otherwise all callers need to initialise them, or get false positives.
|
|
|
|
|
| |
Makes the daemon responsive to polling and commands, even during a file
upload operation.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Passing raw pointers is bad C++ style, and dangerous, because Protocol
will free the passed-in pointers after uploading them, so we should not
keep using them.
Reduce code duplication in BackupClientDirectoryRecord patch/normal upload.
Return a std::auto_ptr<BackupStoreFileEncodeStream> instead of a
std::auto_ptr<IOStream> from BackupStoreFile::EncodeFile* functions.
|
|
|
|
|
| |
to a std::vector::iterator.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
about it but start backing it up if it appears subsequently.
|
|
|
|
|
|
| |
afresh with a new one, should solve most corrupt database errors such as
the ones reported by Achim.
|