Steven R. Loomis

Results 747 comments of Steven R. Loomis

@mcdurdin looks like it was an out of date build that failed. I'll try rerunning…

it actually probably could use some kind of test because it affects all of the process_event paths. but that could be done later, when further changes are made

fix commit [56aeee3](https://github.com/keymanapp/keyman/pull/7066/commits/56aeee35c76de489bf8439b96cdfe6943a458897) was accidentally added to #7071 … it will come out in the ~wash~ (rebased that PR, all ok now).

uf. https://build.palaso.org/buildConfiguration/KeymanDesktop_TestPullRequests/332386?buildTab=log&logView=flowAware&focusLine=2398&linesState=329 > C:\BuildAgent\work\7ac43416c45637e9\keyman\windows\src\engine\keyman32\kmprocess.cpp(113,27): error C2660: 'km_kbp_process_event': function does not take 4 arguments [C:\BuildAgent\work\7ac43416c45637e9\keyman\windows\src\engine\keyman32\keyman32.vcxproj] but yet looking at the build log: https://github.com/keymanapp/keyman/commit/e8dff5cf4e015ce1e28e9b31ba0977112bc001b9#diff-8762b89391fb2aafa43506a684aedaae668bfff3551b906ac03ede86e7eb9141R115 https://github.com/keymanapp/keyman/blob/e8dff5cf4e015ce1e28e9b31ba0977112bc001b9/windows/src/engine/keyman32/kmprocess.cpp#L115 it's being called with 5 arguments not...

> `), (uint8_t)_td->state.isDown), KM_KBP_EVENT_FLAG_DEFAULT)` should be > `), (uint8_t)_td->state.isDown, KM_KBP_EVENT_FLAG_DEFAULT))`

Discussion with @rc-swag β€” - [ ] will move the `is_key_down` parameter into a flag. So `process_event` will go from 4 parameters to 4.

I think this pr already hits all call sites. I'll merge and we can consider followon, it's still the feature branch.

> However, if we change the meaning of an API parameter without changing the shape of the call, we have to (a) manually audit all usages, and (b) update every...

Per slack discussion: - see #5052 PR with capslock tests some tests around this. The updated versions and test-keyboards are here in the windows platform acceptance test on the wiki...

@mcdurdin updated Β§C5015.3 above w/ pseudocode for keyboard processor. I know we said we don't need a `vkey` table originally because they will be mapped by the compiler, *but* for...