Releasing lock modifiers during Auto Type does not work reliably
on X11 and can cause some modifiers to get stuck and more precisely
layout switching can get stuck.
Instead of trying to find out what modifiers to release we can just
abort when the user is holding modifiers that may affect the typing
sequence.
Fixes#6350
* Shortcut to toggle search all entries
* Select first match only when we have a window match
When we default to full database search it's possible the user would select the first match without by accident.
In this case when our query is empty we don't select anything and you need to either type something or press down to select the first item.
* Added username, password, and TOTP keyboard shortcuts and a help tip
* Closes#6176
Co-authored-by: Jonathan White <support@dmapps.us>
AutoTypeActions are required to return a result object with
information if they can be retried or not. An error string is
also provided to show a user friendly message why said action did
not succeed if even after retries it keeps failing.
This is a prerequisite for waiting for modifier keys to be released
on X11.
Completely rewritten XCB Auto-Type keymap system.
- supports multiple simultaneous layouts
- prefers current layout if it has all keysyms available
- removed hardcoded KeySymMap
- removed clunky custom KeySym emulation
Biggest breaking change is removing KeySym emulation for keys that
do not exist in any of the layouts currently in use. It would be
possible to make it work but if you are trying to type syms that
are not available in any of your layouts you are abusing it. It
also adds unnecessary complexity and opens up timing issues when
the keymap is modified on-the-fly. Now we are just reading it.
This also workarounds a Qt related issue where QX11Info::display()
returns a connection to X server that fails to receive updated
keymap data when client settings change. We use our own connection
now to get it working.
* Fix#5958 - Modal dialogs do not have WM_STATE set even though they are a valid top-level window with a valid name. In this case, we need to poll for WM_TRANSIENT_FOR which returns the top level window the dialog is a child of.
* Close#1825 - Add full support for T-CONV and T-REPLACE-RX placeholders. Exception is support for the "raw" type in T-CONV.
* Close#5333 - Allow comment syntax to be present in the Auto-Type sequence
* Closes#725
Support Auto-Type {PICKCHARS} placeholder. Open a dialog that lets you pick characters of an entry's password by their position. Supports typing {TAB} in between characters to move between fields (if necessary). Also supports using arrow keys to quickly navigate around the choice grid.
* Close#2603 - Add support for modifier syntax (+, ^, and %)
* Fix#2633 - Allow reference syntax {REF:...} in Auto-Type sequences
* Close#5334 - Tell the user which part of the Auto-Type sequence is invalid for easy correction
* Fix#2401 - Select the right window on macOS prior to starting Auto-Type
* Allow for nested placeholders
Significant improvements to the Auto-Type select dialog. Reduce stale and unnecessary code paths.
* Close select dialog when databases are locked.
* Close open modal dialogs prior to showing the Auto-Type select dialog to prevent interference.
* Never perform Auto-Type on the KeePassXC window.
* Only filter match list based on Group, Title, and Username column data (ie, ignore sequence column)
* Always show the sequence column (revert feature)
* Show selection dialog if there are no matches to allow for a database search
* Close#3630 - Allow typing {USERNAME} and {PASSWORD} from selection dialog (right-click menu).
* Close#429 - Ability to search open databases for an entry from the Auto-Type selection dialog.
* Fix#5361 - Default size of selection dialog doesn't cut off matches
Since Homebrew moved all its stuff to /opt/homebrew, our hard-coded
install_name_tool patch magic stopped working. This patch uses
macdeployqt for all executables to prevent this kind of behaviour.
Fixes#6042
Fixes#4444
Some websites these days do not present both the "username" and the "password"
input box on the same webpage (e.g. Google, Amazon). So no custom sequence is
possible to enter both the said attributes in one go.
So, two new context menu actions have been added:
1. Perform Auto-Type of just the username
2. Perform Auto-Type of just the password
These context menu actions are analogous to "Copy username" and "Copy
password", except it avoids sending all characters via clipboard.
* Create a sub-menu in the Context Menu of Entry.
* The sub-menu offers the following sequences:
- {USERNAME}
- {USERNAME}{ENTER}
- {PASSWORD}
- {PASSWORD}{ENTER}
Replaces all string configuration options with enum types
that can be checked by the compiler. This prevents spelling
errors, in-place configuration definitions, and inconsistent
default values. The default value config getter signature was
removed in favour of consistently and centrally default-initialised
configuration values.
Individual default values were adjusted for better security,
such as the default password length, which was increased from
16 characters to 32.
The already existing config option deprecation map was extended
by a general migration procedure using configuration versioning.
Settings were split into Roaming and Local settings, which
go to their respective AppData locations on Windows.
Fixes#2574Fixes#2193
Moves MacUtils into a separate sub folder and adds
WinUtils, NixUtils, and OSUtils for platform-native
code on Windows, Unix-like, and generic/all systems.
* Fix#3351
On macOS, entry specific Auto-Type (Cmd+Shift+V) now raises the last active window before lowering the KeePassXC window. This improves accuracy of the window that is actually typed into.
This change adds a right-click context menu to the
AutoType dialog, which allows the user to copy
either the username or password. The dialog then
automatically closes.
* Store the currently active window right when the global keyboard shortcut is triggered
* Eliminate unnecessary window raise/lower and delays on macOS
* Remove duplicate addition of macutils symbols from mac Auto-Type plugin
* Fix tests to fake trigger a global autotype sequence
Many lines were not conformant with the project's formatting rules.
This patch should fix all formatting and whitespace issues in the code
base.
A clang-format directive was put around the connect() calls containing
SIGNALs and SLOTs whose signatures would be denormalized because of the
formatting rules.
The Database, DatabaseWidget, and DatabaseTabWidget classes share many responsibilities in inconsistent ways resulting in impenetrable and unmaintainable code and a diverse set of bugs and architecture restrictions. This patch reworks the architecture, responsibilities of, and dependencies between these classes.
The core changes are:
* Move loading and saving logic from widgets into the Database class
* Get rid of the DatabaseManagerStruct and move all the information contained in it into the Database
* Let database objects keep track of modifications and dirty/clean state instead of handing this to external widgets
* Move GUI interactions for loading and saving from the DatabaseTabWidget into the DatabaseWidget (resolves#2494 as a side-effect)
* Heavily clean up DatabaseTabWidget and degrade it to a slightly glorified QTabWidget
* Use QSharedPointers for all Database objects
* Remove the modifiedImmediate signal and replace it with a markAsModified() method
* Implement proper tabName() method instead of reading back titles from GUI widgets (resolves#1389 and its duplicates #2146#855)
* Fix unwanted AES-KDF downgrade if database uses Argon2 and has CustomData
* Improve code
This patch is also the first major step towards solving issues #476 and #2322.
Use nullptr instead of 0 or NULL to initialize a null pointer. In some
cases, readability was enhanced by replacing 0 with more meaningful
values according to the type of the pointer being initialized.
The sole purpose of a few objects was calling a static member of the
class they belonged to. This is not needed, as you can access a static
member with the <class_name>::<member_name> notation.
AutoTypeAction class had no out-of-line definitions and because of this
the compiler would place its vtable everywhere the class is used.
By simply defining the virtual destructor in the .cpp file, the issue
goes away.
Also, a few classes derived from AutoTypeAction had missing 'override'
qualifiers, which have now been added.
Performing a dev build against the latest version of Qt failed
because of some deprecated members. They have been replaced
according to the Qt documentation.
Further, Q_OS_MACOS is now the only macro available to identify a
machine running macOS, the others are now deprecated.
See https://doc.qt.io/qt-5/qtglobal.html#Q_OS_OSX and
https://doc.qt.io/qt-5/qtglobal.html#Q_OS_MAC.
* Create history-based merging that keeps older data in history instead of discarding or deleting it
* Extract merge logic into the Merger class
* Allows special merge behavior
* Improve handling of deletion and changes on groups
* Enable basic change tracking while merging
* Prevent unintended timestamp changes while merging
* Handle differences in timestamp precision
* Introduce comparison operators to allow for more sophisticated comparisons (ignore special properties, ...)
* Introduce Clock class to handle datetime across the app
Merge Strategies:
* Default (use inherited/fallback method)
* Duplicate (duplicate conflicting nodes, apply all deletions)
* KeepLocal (use local values, but apply all deletions)
* KeepRemote (use remote values, but apply all deletions)
* KeepNewer (merge history only)
* Synchronize (merge history, newest value stays on top, apply all deletions)
