We always using Unicode functions to interact with UI. We convert UTF8 string to UTF16 and vis-versa.
Overwrite input string instead of using resize that caused old test to remain.
Fix case of readOnly by using correct message.
change position of OK/cancel button to match other dialogs.
Activate translation on this dialog.
This started to cause issues after latest changes to disable privileges when they are no more needed.
Because of the bug, the privileges could not be enabled again because they were wrongly removed.
Cause seems to be IME as documented in KeePass project (https://keepass.info/help/kb/sec_desk.html#ime). We use the same approach as KeePass to disable IME in Secure Desktop.
This commit also add few changes:
- we switch to secure desktop only if SetThreadDesktop succeeds
- we call SwitchDesktop to switch to original desktop only if we actually succeeded in displaying secure desktop
If vulnerability detected, a warning message is displayed during mount or backup/restore header, and changing the password is disallowed since it will not change the master key.
* Implement Language selection into settings
Initial commit to create a new tab in PreferencesNotebook for
Language selection. By default, if nothing is chosen, it uses the
current behaviour of using the language from system environment
variables. If another language is chosen from the settings, it is
saved into the Configuration.xml and this is used instead.
* Fix SetStringSelection() assert issue on macOS
* Add header include to fix build
* Add current language pack, authors and way to use literal strings
* Translations also for FreeBSD
* Minimal GTK3 WX build on FreeBSD requires wxGraphicsContext
* Get Preferences properly instead of workaround function
* Use WrapSizer instead of BoxSizer for author line
This forces long author lists to be put on a new line, reducing
the need to increase window width.
* Update Finnish translation
* Borrow translation from IDM_LANGUAGE where it makes sense
* Remove colon and thus unneeded function
* Simplify Language tab layout
* Reintroduce macOS specific fixes to Forms.cpp
* cleanup
* EMVCard.cpp: ArrayToHexWideString: prohibit conversion of a string constant
../Common/EMVCard.cpp: In function 'std::wstring VeraCrypt::ArrayToHexWideString(con
st unsigned char*, size_t)':
../Common/EMVCard.cpp:28:43: warning: ISO C++ forbids converting a string constant
to 'wchar_t*' [-Wwrite-strings]
28 | static wchar_t* hexChar = L"0123456789ABCDEF";
| ^~~~~~~~~~~~~~~~~~~
Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
* EMVCard.cpp: ArrayToHexWideString: fix of the comparison of different types
../Common/EMVCard.cpp: In function 'std::wstring VeraCrypt::ArrayToHexWideString(con
st unsigned char*, size_t)':
../Common/EMVCard.cpp:32:43: warning: comparison of integer expressions of different
signedness: 'int' and 'size_t' {aka 'long unsigned int'} [-Wsign-compare]
32 | for (int i = 0; i < cbData; i++)
| ~~^~~~~~~~
Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
* SecurityTokenKeyfilesDialog.cpp: removed initialization of an unused variable
Forms/SecurityTokenKeyfilesDialog.cpp:58:24: warning: unused variable 'i' [-Wunused-
variable]
58 | size_t i = 0;
| ^
Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
* Core/Unix: throwing an exception instead of ignoring the error
Fixes: 5a6b445f ("fix warnings and UB (#1164)")
Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
---------
Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
Co-authored-by: Vasiliy Kovalev <kovalev@altlinux.org>
We use only alphanumeric characters for the name in addition to the '_' character
We ensure the random desktop name doesn't already exist
We create the random desktop on the Secure Desktop thread and we look over SwitchDesktop until it works.
We call SetThreadDesktop (hOriginalDesk) in main thread once we finish
- use a more standard-looking memory value tag instead of fully random one that will look suspicious and outstanding
- If we fail to allocate 1MiB for derivation area, we device the size by two in a loop until we succeed. This is better than falling back directly to 8KiB size
- Better method to derive actual encryption key: instead of simply duplicating 128bit key value, we combine a xor addition and self-encryption to build a 256-bit key
- use both encID and pbKeyDerivationArea to derive IV for encryption and not only endID
