AJ B

TZX files with CSW blocks fail to load

Both are re-recorded and replaced in the zip. Boovie2_31.rzx 2008-08-17 (was EmuZWin 2.7 now Spectaculator 8.0.3092); FutureGames_End.rzx 2020-10-29 (was SPIN 0.5 now Spectaculator 8.0.3092). Originals are on archive.org. Boovie2_31 was truly malformed, FutureGames_End same category failure as chernobil. I guess the value for testing is not in these individual files and sourcing a full new dump is a separate bigger task

Two marked as failing in the regression database are replaced in the archive and now play. Their sha256 no longer matches as the CSV till describes the old ones: boovie2/Boovie2_31.rzx futuregames/FutureGames_End.rzx

Please close this ticket as closed-invalid. This issue is invalid and the patch will break things . I tested with 2 different patches in my tree, the poluted build & test environment led to invalid findings.

Please close this ticket. This fix and patch is no good. I tested with 2 different patches in my tree, unclean build tree led to invalid findings

RZX 'repeat last frame' resolves to the wrong frame -> libspectrum_rzx_playback error

The 8 missing INs are the ROM keyboard scan: frame 6986 ends with EI, so z80_interrupt() defers the interrupt via z80_interrupt_event — which z80_interrupt_event_fn() ignores under RZX playback, dropping it. An RZX frame boundary is where the recorder accepted an interrupt, so the EI delay should not apply during playback. --- a/z80/z80.c +++ b/z80/z80.c @@ -195,7 +195,7 @@ /* If interrupts have just been enabled, don't accept the interrupt now, but check after the next instruction has been executed...

audio2tape -k discards the audio for a block that decoded as all zeros