There is a race condition where two processes may both attempt to open
the same project file. This appears to happen, for example, if an author
double-clicks on a single-click shortcut, launching Keyman Developer
twice in quick succession.
This fix adds a .lock file which is deleted automatically on close of
the project or on normal or abnormal process termination. If a second
process encounters the .lock file, it will simply open the welcome view
instead of attempting to open the project.
Fixes: #11584
Fixes: KEYMAN-DEVELOPER-1PR
Previously, we used U+FFFC as a marker in the debugger (for deadkeys and
LDML keyboard markers), which displayed as the letters OBJ in a small
dotted square. However, we switched the underlying component to
RichEdit, in order to better support rendering of many scripts (EDIT has
some problems). The RichEdit control silently converts U+FFFC to U+0020.
I switched to U+0001 as this is handled better, but has an unfortunate
missing-glyph marker.
In the future, it would be better to use a non-character marker
(leveraging RichEdit), but this is a much more complex change, and this
fix at least matches previous functionality.
Fixes: #13293
The debugger memo takes input with \n but internally represents this as
\r\n, so we need to strip out \r in order to track caret position
correctly.
Fixes: #13292
Fixes: #13278
The fallback default keyboard corresponding to an uninitialized cookie is `Keyboard_us:en`. This should only occur if the site actually registers its stub before initialization - and we weren't checking/filtering for this properly.
The error itself only occurred if there _were_ stubs pre-registered before init - if none existed, the error does not result.
The cached context comparison for v10 kmw compiler includes the `nul`
statement, unlike the comparison in the earlier compiler. This
difference was missed in earlier iterations including the recent patch
in #13003. This patch fixes the problem for both `context(n)` and
`index(store,n)` offset calculations.
Fixes: #13306Fixes: #13307
Follows: #13003
This opens up the Delphi IDE with the correct KEYMAN_ROOT environment,
which allows for debugging. This is very helpful when working with
multiple worktrees, and also helps to avoid editing a project in the
wrong tree.
Requested by @DavidLRowe. Default was either CWD or My Documents,
neither of which is helpful.
The initial dir will now be folder of the .kps file, and then will
remember last used folder for each additional use of the dialog for that
.kps file.
Passing this parameter causes docker image builds on mac to fail (#13295).
Since `amd64` is the default anyways (and we don't support other values),
we can simply remove that parameter.
Fixes: #13295
Two separate bugs addressed, with `index()` references and with
`context()` references -- both have the same root cause, of not taking
`nul` at the start of the context into account (as `nul` is not included
in the `m_miniContext` member, being a non-character). We already fixed
this issue for `if()` quite a long time ago, and some of the same
patterns can be with `m_miniContextIfLen` for example.
Fixes: #13304Fixes: #13316
Add 4 test keyboards to validate `nul` and `if` used in conjunction with
`index` and `context` and corresponding references in core unit tests.
Add a script to rebuild baseline keyboards using a copy of kmcomp.exe
16.0.138; this is setup and tested only on Windows (YMMV on WINE, etc).
Add the 4 additional baseline test keyboards to kmcmplib unit tests for
build consistency between kmcomp 16 and kmcmplib 18 (all pass).
Note: two of the new Core tests currently fail. This is expected, see
issue #13304.
Fixes: #13303
Add validity checks to kmcmplib to verify that offsets provided to
`context(n)` are in range for the context, and do not point to
non-character elements (`nul` and `if`) in the context string.
Fixes: #13301Fixes: #13302
Relates-to: #13299
Relates-to: #13276
