Release build executables should have 3 component version numbers with
the version tag appended, e.g. keymandesktop-14.0.155-alpha-local.exe.
Coming later, rename keymandesktop-version.exe to keyman-version.exe.
Fixes#3665.
On some systems, subkeys of HKCU\Software\Keyman could have incorrect
permissions and these would not be corrected with earlier fixes such as
in #2316. This fix resolves the problem by recursively correcting
permissions on all subkeys, rather than just the top two levels.
This issue caused Metro-style applications (Windows Store apps) to fail
to accept Keyman keyboard input.
Addresses Coverity reports for compiler.cpp and other files.
Note that whitespace was reformatted on this PR. You may want to compare
with ignore-whitespace.
Fixes#3561.
Resolves issues with upgrading disabled keyboards, and fixes a secondary
issue where Windows was giving us invalid LANGIDs for locales, which was
causing trouble with transient LANGIDs.
Relates to #3561.
If a keyboard has languages installed under one of the four transient
language profiles, then we need to jump through some additional hoops to
upgrade it to the 14.0 model.
This does not yet handle all scenarios around installed but non-loaded
keyboards.
Relates to #3561.
If a keyboard is installed for a transitional profile but another
keyboard is already installed for that profile, then Keyman would
crash, expecting the transitional profile to be missing.
Fixes#3633.
Memory allocation was half its required size due to missing
`sizeof(Char)` calculation.
Also added extra tags to Sentry events for command line and executable,
which made tracing this issue so much easier.
Fixes#1843.
If there are so many keyboards installed that the menu would grow larger
than 3/4 of the screen height, then the menu will become scrollable.
Fixes#1659.
If there are too many keyboards to fit in the OSK toolbar, the toolbar
will now show a dropdown menu button and excess keyboards will be
visible there.
Fixes#3084.
This does two things:
1. Cleans up a bunch of places where we used to use `(int)` typecasts
for pointer math, which was problematic. We now use `(INT_PTR)` per
MSDN https://docs.microsoft.com/en-us/windows/win32/winprog64/rules-for-using-pointers
and then cast that down to `(int)` where necessary, e.g. when storing
string lengths which are never going to be more than a few hundred
characters! Doing this explicitly helps to clarify that we are aware
of the typecast and believe it to be safe.
2. Adds in some build infrastructure for future use of Coverity Scan
https://scan.coverity.com/ which we plan to use for further code
quality updates. I have submitted the project to Coverity and are
now waiting for approval so we can check results. Once we have
approval, I do plan to add this to the nightly build (we need to
keep submissions under 3 builds/day).
Note: I have not yet added Keyman Core (Windows) to this project,
nor are we currently building Keyman Core (macOS) or Keyman for
Linux, but we should consider adding those in future.
