Add secondary joystick hard drop and deliver build 05
This commit is contained in:
1 parent
a07c41555a
commit
49bc78158a
21 files changed
+230
-35
No files matched your search
@@ -0,0 +1,80 @@
|
||||
# Secondary joystick hard drop — November 11, 2026
|
||||
|
||||
Brian requested Fire 2 hard drop, equivalent to Space. The game uses the selected
|
||||
right/left joystick's named standard CMOC secondary mask (right 0x04, left 0x08)
|
||||
from the same stable button snapshot used to filter keyboard rows. In gameplay,
|
||||
it joins the existing Q_HARD held action before edge detection: one drop on press,
|
||||
no repeat while held, and no rotation on release. Space and Fire 2 share the same
|
||||
logical edge, preventing a second drop while either alias stays down.
|
||||
|
||||
The secondary button is ignored in menus and with Joystick Off. A held button
|
||||
across scene transitions must be released before input is rearmed. Joystick Up
|
||||
hard drop, Down soft drop and primary short/long-press actions remain. Both port
|
||||
mappings are covered by hardware-mock adapter tests with real q_input_step/q_drop;
|
||||
held samples leave exactly four settled cells, and another released/repressed
|
||||
button locks a second piece. Wrong port, Off, menu and transition cases pass.
|
||||
All four shared keyboard rows continue to be filtered. The Controls page now
|
||||
lists Up/Fire 2 hard drop and Fire 1 rotation/actions.
|
||||
|
||||
Build05 was first built using official VCC Plus 3.6.1 (53), then rebuilt and
|
||||
qualified with installed 3.6.1 (54), CMOC 0.1.100 Release. The two builds produce
|
||||
identical binary/disk bytes. The game-side calibration-pointer compiler workaround
|
||||
remains despite the general compiler fix now included in the installed app. Code/const
|
||||
end $7AFE, static end $13FD, original 1024-byte stack reserve. No persistent state,
|
||||
score formula or save schema changed. Native qualification follows; host simulation alone does not prove physical
|
||||
joystick behavior.
|
||||
|
||||
|
||||
## Final installed-app check
|
||||
|
||||
Brian authorized the final installed VCC Plus check after Core installed build54.
|
||||
The fresh native project grant identified project
|
||||
273E195B-8BFC-495F-A1F8-E1469759B910 revision0 and advertised joystick injection
|
||||
and diagnostics. Source hashes matched main.c, keyboard.c/h, menu asset and
|
||||
Makefile. The official Release build CC51E81E-B5E2-4670-B8FA-15785BAFEB51
|
||||
succeeded without diagnostics; artifact F137D778-646A-43AE-9CF1-92A658464FF3.
|
||||
See BUILDS.md for binary/disk hashes and exact memory bounds.
|
||||
|
||||
Disposable run C3EAFAD1-3CB6-435F-9B40-03E55C6CB534 used MC6809 and 512 KB,
|
||||
booted from that exact artifact. Score loading and initial drawing were allowed
|
||||
to finish before controls were evaluated. Public playtest_input steps used
|
||||
standard right-port X/Y snapshots (0..16383, center8191) and named fire1/fire2;
|
||||
get_playtest_input and input_trace confirmed the actual states and release.
|
||||
No emulator internals or guest memory/register writes were used.
|
||||
|
||||
- Fire 2 on the title did not start a game.
|
||||
- Empty-board baseline had no settled cells. Fire 2 was held 180 fields, then
|
||||
released 60 fields: four green settled cells and the next purple piece still
|
||||
falling. There was no repeat or release rotation.
|
||||
- Fire 2 pressed again for 12 fields, then released 120: eight settled cells
|
||||
(green + purple), with the yellow piece falling. Exactly one additional lock.
|
||||
- Actual analog X=0 and X=16383 moved the yellow piece to the respective walls.
|
||||
- Primary held 8 fields/released 90 rotated the yellow piece vertically and
|
||||
left it rotated. Up (Y=0) still hard-dropped it.
|
||||
- Keyboard P/Enter pause/resume followed by actual joystick LEFT moved the next
|
||||
gray piece to the left wall. Keyboard confirmation did not disable the stick.
|
||||
- A 120-field idle audio capture had PCM peak0 in both channels. This isolated
|
||||
silence does not establish that Brian's normal-window buzzing is fixed.
|
||||
|
||||
The run stopped normally at field5806, with zero disk writes. Its writable test
|
||||
disk was retained with the same pristine hash. The normal paused session,
|
||||
DualSense assignments, stopped tape and personal score media were preserved.
|
||||
|
||||
Both secondary port masks pass host adapter/gameplay fixtures; this guest run
|
||||
qualifies the right port only. Actual left port, physical controller/CoCo,
|
||||
HD6309 on this candidate and 128 KB remain unqualified. No save schema or scoring
|
||||
formula changed. Host input/store/keyboard, rules and video suites pass with
|
||||
AddressSanitizer/UndefinedBehaviorSanitizer.
|
||||
|
||||

|
||||

|
||||

|
||||
|
||||
## Brian's acceptance test
|
||||
|
||||
Use a writable copy of QUADPSL05.DSK, type DOS to load it, and select Right
|
||||
joystick (the default). Confirm left/right movement; primary tap/release rotation
|
||||
that stays rotated; secondary hard drop once per press even while held; another
|
||||
drop after release/repress; and joystick movement after keyboard pause/resume.
|
||||
Up remains another hard drop and Down soft drop. Old disks keep their scores;
|
||||
build05 starts fresh. Note machine RAM and whether buzzing occurs at the title.
|
||||
Reference in new issue
Block a user