Repository navigation
precap-v0.1.3
tagged this
02 Nov 01:17
v0.1.0 passed T1-T9. Three reports came back from daily use; each was fixed
on its own and re-verified on the camera before the next was attempted.
1. The 10 s half-press timeout was too short: now 30 s.
2. Releasing the half press wrote one photo anyway. The release sends [7]
and the firmware turns that into a shot. P_SWALLOW is set on release;
the next frame is dropped and the flag clears itself. The next half
press also clears it.
3. Deleting a photo in playback froze the camera. The ring holds job ids,
and playback rebuilds that database, so a saved id can belong to somebody
else. drop_one and commit_one now bail out when F_JOBDB returns 0 and
when job+0x10FC does not match the id they were given.
Known limit, accepted: P_SWALLOW is consumed by the next frame, so if the
firmware produces none (AF never locked) the flag survives until the next half
press. A full press with no half press before it would then lose one frame.
One frame, not a freeze.
21 tests pass; rebuilding is byte-identical to the shipped BIN.
PreCap_v0.1.3.BIN, 2,780 B, sha256 90003218...