fix(core): implement ldml_processor::get_key_list()

[srl295]

• ldml::vkeys class updated to keep a set<> of keys
• fix the test case to not leak!

also

• move some test utils to statics
• add a new test action type and @@key-list keyword
• print warning on unhanded @-commands

Fixes: #12298

@keymanapp-test-bot skip

[keymanapp-test-bot[bot]]

User Test Results

Test specification and instructions

User tests are not required

Test Artifacts

• Developer
  • Keyman Developer
  • Compiler Regression Tests
  • kmcomp.zip
• Keyboards
  • Test Keyboards
• Linux
  • Keyman for Linux: focal
  • Keyman for Linux: jammy
  • Keyman for Linux: noble
  • Keyman for Linux: oracular
• macOS
  • Keyman for macOS
• Web
  • KeymanWeb Test Home
• Windows
  • Keyman for Windows
  • FirstVoices Keyboards for Windows
  • Text Editor (32 bit)
  • Text Editor (64 bit)

[srl295]

@rc-swag @mcdurdin would this need user testing?

also note that gap keys / unassigned keys are NOT included in the list.

[mcdurdin]

@rc-swag @mcdurdin would this need user testing?

Yes, I think it should have testing (all desktop platforms)

also note that gap keys / unassigned keys are NOT included in the list.

They probably need to be, otherwise they will fall back to default output (think Shift modifier for example). The rule we decided on for LDML keyboards was that if a layer is defined at all for a given modifier set, it is defined for all keys on that modifier set (even if it is no-op).

[srl295]

@rc-swag @mcdurdin would this need user testing?

Yes, I think it should have testing (all desktop platforms)

• [ ] OK. Will ponder the plan.

also note that gap keys / unassigned keys are NOT included in the list.

They probably need to be, otherwise they will fall back to default output (think Shift modifier for example). The rule we decided on for LDML keyboards was that if a layer is defined at all for a given modifier set, it is defined for all keys on that modifier set (even if it is no-op).

• [ ] OK. So the set will be more fixed, more based on the modifier set than the exact key layout. I can probably just generate the list then.

[rc-swag]

@rc-swag @mcdurdin would this need user testing?

Yes, I think it should have testing (all desktop platforms)

• [ ] OK. Will ponder the plan.

Following the test the found the issue in the first place will exercise the code. Not sure for the other platforms if there is a ready made user test.

[srl295]

great, yes bringing in the test case from https://github.com/keymanapp/keyman/issues/12298#issue-2488657340 makes sense.

i think i could do this with a set of the modifier combinations, and then just expand on the 'vkey map' X modifier combinations.