summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAge
...
* Catch exceptions during restore and report them, rather than abortingChris Wilson2007-03-08
| | | | | bbackupquery (refs #3)
* Count a difference when a local directory does not exist, but it doesChris Wilson2007-03-08
| | | | | exist on the server (refs #3)
* Win32 compile fixChris Wilson2007-03-05
| | | | | (refs #3, merges [1306])
* Add logging in remaining places where Common OSFileOpenError can be thrown.Chris Wilson2007-03-05
| | | | | (refs #3, merges part of [1099], and [1370])
* Add logging in remaining places where Common OSFileOpenError can be thrown.Chris Wilson2007-03-05
| | | | | | | Fix compile warning (signed vs unsigned comparison) (refs #3, merges part of [1099], and [1368])
* Update to match new recommended installation of pcreposix.h as regex.h.Chris Wilson2007-03-04
| | | | | (refs #3, merges [1281], reverting [1233])
* Add logging in remaining places where Common OSFileOpenError can be thrown.Chris Wilson2007-03-04
| | | | | (refs #3, merges part of [1099] and [1359])
* Watch out for our leak tracking data being destroyed and don't crash whenChris Wilson2007-03-04
| | | | | subsequent objects are destroyed. (refs #3, merges [1341])
* Declare that we have O_BINARY but not <sys/types.h> for MSVCChris Wilson2007-03-04
| | | | | (refs #3, merges [760] and [763])
* Fix make error when optional file doesn't exist (refs #3, merges [1098])Chris Wilson2007-03-04
|
* #include Test.h to fix compile (needed for TEST_THAT)Chris Wilson2007-03-04
| | | | | | | | | Improved error message Improved comments (refs #3, merges parts of [712], [1289] and [1337])
* Update build instructions for Cygwin/Win32 (refs #3)Chris Wilson2007-03-04
|
* Simplify check for PERLChris Wilson2007-03-04
| | | | | | | | | Define substitution TARGET_PERL and preprocessor PERL_EXECUTABLE to a native Perl (not Cygwin) on Win32, since Cygwin perl will not run inside bbackupd for unit tests (e.g. SyncAllowScript). (refs #3)
* Typo fixChris Wilson2007-03-04
|
* Fix compile warningChris Wilson2007-02-09
|
* Log failure to delete files and directories (refs #3)Chris Wilson2007-02-09
|
* Catch exceptions while writing files, attributes, resume info and checkingChris Wilson2007-02-09
| | | | | file existence (refs #3)
* Don't throw an exception if we fail to delete a file in the way of restore,Chris Wilson2007-02-09
| | | | | just log the error and return an error code (refs #3)
* Spelling fix (refs #3)Chris Wilson2007-02-09
|
* Make BackupClientRestoreDir return a result code compatible withChris Wilson2007-02-09
| | | | | BackupClientRestore (refs #3)
* Check whether the restore target parent directory exists before tryingChris Wilson2007-02-09
| | | | | to create the target directory or save restore info in the parent (refs #3)
* Catch errors during restore (refs #3)Chris Wilson2007-02-09
|
* Stop locations from being deleted from the store just because their localChris Wilson2007-01-25
| | | | | directory can't be found, as suggested by Gary (refs #3)
* Improve error message when server fails to start (refs #3)Chris Wilson2007-01-21
|
* Buffer store file integrity checks (refs #3)Chris Wilson2007-01-17
|
* Buffer directory reads (refs #3)Chris Wilson2007-01-17
|
* When comparing, don't recurse into excluded directories, as their contentsChris Wilson2007-01-17
| | | | | | | | will not be on the store. Fix display of filenames during compare when backing up the root directory (double slash) (refs #3)
* Moved MakeFullPath into its own library file so that we can share itChris Wilson2007-01-17
| | | | | (BackupQueries needs it too) (refs #3)
* Use BufferedStream to speed up housekeeping by about ten times. (refs #3)Chris Wilson2007-01-17
|
* Added a BufferedStream class that can be wrapped around an IOStream to Chris Wilson2007-01-17
| | | | | | improve read performance when many small reads will be performed, e.g. while reading directories and during housekeeping. (refs #3)
* Allow suppressing warnings about unsupported file types by excluding thoseChris Wilson2007-01-16
| | | | | files by name, using ExcludeFile (refs #3)
* Automatically initialise logging at startup, with a static object, to logChris Wilson2007-01-16
| | | | | | | | to console and syslog. All logging objects automatically register themselves with the global logging system. (refs #3)
* Check for exclude entries that end in a path separator, and log a warningChris Wilson2007-01-16
| | | | | (refs #3)
* Compile fix for RHEL4 ([NICK]) (refs #3)Chris Wilson2007-01-16
|
* Cosmetic whitespace fix (refs #3)Chris Wilson2007-01-16
|
* Log AccessDenied errors in a more concise form than general exceptions Chris Wilson2007-01-16
| | | | | (refs #3)
* Add option parsing with getopt()Chris Wilson2007-01-15
| | | | | | | | | | | Add "-D" flag as SINGLEPROCESS equivalent Add "-q" and "-v" options to control master logging level Log fatal errors as FATAL rather than ERROR Log daemon start and stop as NOTICE rather than INFO (refs #3)
* Throw CommonException AccessDenied if we didn't get access to the file,Chris Wilson2007-01-15
| | | | | for better error description (refs #3)
* Allow Loggers to log everything, as is their wont. Master level will beChris Wilson2007-01-15
| | | | | set in Daemon, patch to follow. (refs #3)
* Log important output at NOTICE level, instead of INFO (refs #3)Chris Wilson2007-01-15
|
* Add a new logging level, NOTICE, between INFO and WARNING (justification:Chris Wilson2007-01-15
| | | | | | | | | | we need two levels of output for LogAllFileAccess, neither of which are warnings, one is very verbose, but must not be compiled out like TRACE). Make Loggers default to logging everything. Make the global log level filter work. (refs #3)
* Skip mount points and NTFS reparse points inside a location, even ifChris Wilson2007-01-15
| | | | | not excluded (refs #3)
* Detect reparse points and change the device number (refs #3)Chris Wilson2007-01-15
|
* Spacing (refs #3)Chris Wilson2007-01-14
|
* Add missing files to manifest, required by configure (refs #3)Chris Wilson2007-01-14
|
* Fix building from distribution tarballs again (refs #3)Chris Wilson2007-01-14
|
* Don't do things with essential side effects inside ASSERT() macros Chris Wilson2007-01-14
| | | | | (refs #3, refs #9)
* Be nice to rpm that doesn't like empty macros (refs #3)Chris Wilson2007-01-14
|
* Don't die on unreadable files (refs #3)Chris Wilson2007-01-14
|
* - Make sure timer is stopped before removing signal handler, otherwiseChris Wilson2007-01-14
| | | | | SIGALRM will kill us.