Fixes#6942 and fixes#4443
- Return number of deleted entries
- Fix minor memory leak
- FdoSecrets: make all prompt truly async per spec and update tests
* the waited signal may already be emitted before calling spy.wait(),
causing the test to fail. This commit checks the count before waiting.
* check unlock result after waiting for signal
- FdoSecrets: implement unlockBeforeSearch option
- FdoSecrets: make search always work regardless of entry group searching settings, fixes#6942
- FdoSecrets: cleanup gracefully even if some test failed
- FdoSecrets: make it safe to call prompts concurrently
- FdoSecrets: make sure in unit test we click on the correct dialog
Note on the unit tests: objects are not deleted (due to deleteLater event not handled).
So there may be multiple AccessControlDialog. But only one of
it is visible and is the correctly one to click on.
Before this change, a random one may be clicked on, causing the
completed signal never be sent.
* Closes#6335
* Modify application settings presentation to allow for alternative saving strategies
* Transition Database::save calls to using flags to control saving behavior. Reduces boolean flags on function call.
* Made direct write save option a local setting to prevent unintentional carry over between platforms.
* Introduced in #6438, modified signal is not blocked at the Database level when emitting is blocked. This causes infinite saving to occur when Always Save After Every Change is enabled.
This allows one to directly use Diceware-compatible wordlists without having to convert the file to the plain wordlist format.
The accepted formats are described in the Diceware documentation:
https://diceware.readthedocs.io/en/stable/wordlists.html
* Fixes#6459
Improves the overall handling of FdoSecrets showing client executable paths to the user. It does the following:
* Check executable file existence as described in [RFC] fdosecrets: add optional confirmation to secret access (#4733)
* Show application PID and dbus address in the client list
* When the executable file is inaccessible, depending on where the client name is shown:
* when shown inline, e.g. in notification text, where space is limited, clearly say that the path is invalid
* when shown in auth dialog, show warning and print detailed info about the client
* when shown in the client list, draw a warning icon
Co-authored-by: Jonathan White <support@dmapps.us>
* Support NFC readers for hardware tokens using PC/SC
This requires a new library dependency: PCSC.
The PCSC library provides methods to access smartcards. On Linux, the third-party pcsc-lite package is used. On Windows, the native Windows API (Winscard.dll) is used. On Mac OSX, the native OSX API (framework-PCSC) is used.
* Split hardware key access into multiple classes to handle different methods of communicating with the keys.
* Since the Yubikey can now be a wireless token as well, the verb "plug in" was replaced with a more
generic "interface with". This shall indicate that the user has to present their token to the reader, or plug it in via USB.
* Add PC/SC interface for YubiKey challenge-response
This new interface uses the PC/SC protocol and API
instead of the USB protocol via ykpers. Many YubiKeys expose their functionality as a CCID device, which can be interfaced with using PC/SC. This is especially useful for NFC-only or NFC-capable Yubikeys, when they are used together with a PC/SC compliant NFC reader device.
Although many (not all) Yubikeys expose their CCID functionality over their own USB connection as well, the HMAC-SHA1 functionality is often locked in this mode, as it requires eg. a touch on the gold button. When accessing the CCID functionality wirelessly via NFC (like this code can do using a reader), then the user interaction is to present the key to the reader.
This implementation has been tested on Linux using pcsc-lite, Windows using the native Winscard.dll library, and Mac OSX using the native PCSC-framework library.
* Remove PC/SC ATR whitelist, instead scan for AIDs
Before, a whitelist of ATR codes (answer to reset, hardware-specific)
was used to scan for compatible (Yubi)Keys.
Now, every connected smartcard is scanned for AIDs (applet identifier),
which are known to implement the HMAC-SHA1 protocol.
This enables the support of currently unknown or unreleased hardware.
Co-authored-by: Jonathan White <support@dmapps.us>
This works:
```
$ keepassxc test1.kdbx test2.kdbx --pw-stdin
Database password: <manual input 1234>
Database password: <manual input 4321>
```
But this doesn't (only `test1.kdbx` is unlocked):
```
$ printf '%s\n' 1234 4321 | keepassxc test1.kdbx test2.kdbx --pw-stdin
Database password:
Database password:
```
The problem is that `Utils::setDefaultTextStreams()` is called multiple times
when unlocking multiple databases with `--pw-stdin`, which appears to break the
pipe. Simply call it once to avoid the problem.
Fixes: #5012 (as far as I can tell by simulating the script in Linux)
Signed-off-by: Joan Bruguera <joanbrugueram@gmail.com>
The HEX button automatically unselects `selectBoxLower` therefore there is no need to also add `ghijklmnopqrstuvwxyz` in the excluded characters text field.
* Allow for multiple vendor ID's to be checked at once. This allows for the use of one tracking index, streamlining KPXC code.
* Remove support for libusb 0.x on Linux
* Better handling of USB errors during initial key query. Output warnings to console.
* The main window doesn't hide properly during unlock sequence if it is in the background (ie, not minimized and not hidden to tray). This change makes sure the window hides after interaction on all platforms.
* Closes#2322
* Show locked databases in tabbed interface in unlock dialog for browser and auto-type workflows.
* Make the DatabaseOpenDialog window Application-Modal so that it blocks input to the main UI when the dialog is open. This reduces corner cases by avoiding the possibility of databases getting closed or unlocked
behind the open dialog.
* Introduce search CLI command to replace locate command. Search can provide the same functionality but in a more fine-grained fashion
* Replace use of Group::locate in code: Use EntrySearcher in clip cli command best-match option. This removes the matching against group hierarchy of an entry which is kind of nonsense as clip expects exactly one match. Matching against groups can be done using search command.
* Remove obsolete Group::locate method
Externally opened attachments are now lifecycle-managed properly.
The temporary files are created with stricter permissions and entirely
random names (except for the file extension) to prevent meta data leakage.
When the database is closed, the files are overwritten with random
data and are also more reliably deleted than before.
Changes to the temporary files are monitored and the user is asked
if they want to save the changes back to the database (fixes#3130).
KeePassXC does not keep a lock on any of the temporary files, resolving
long-standing issues with applications such as Adobe Acrobat on Windows
(fixes#5950, fixes#5839).
Internally, attachments are copied less. The EntryAttachmentsWidget
now only references EntryAttachments instead of owning a separate copy
(which used to not be cleared properly under certain circumstances).
* Use hash based storage for last used directories instead of a key/value for each type
* Explicitly declare certain operations as sensitive and follow the "Remember Last Database" setting for those.
* Introduce database backup directory location (close#6619)
unset, wl-copy will try to guess the MIME-Type based on the data.
For some reason this did not work on my machine and i was unable
to paste passwords in Firefox.
