Como algunos sabrán, el 27 de diciembre del 2015 fue un día importante. Desde la versión 2.80 test 5 de Altirra, podemos grabar archivos CAS en el emulador.
La secuencia de uso es la siguiente:
Cassette->New Tape
Cassette->Tape control (presionar botón "rec")
Grabar en el emulador.
Cassette->Save
Tuve éxito inmediato (aceleración c:Patch -cassette SIO desactivada) con SITRE, pero con otros copiadores tuve problemas. Estuve probando el NHP 2.56 y extrañamente, siempre tenía problemas en el último bloque:
Spoiler: MOSTRAR
Phaeron escribió:This is due to a bug in the Atari OS. It has a rare bug where it incorrectly measures the baud rate of the incoming tape signal when the start or end of the sync byte is immediately before the vertical blank interrupt. This is happening consistently on the last record of the tape due to the emulator having fully deterministic timing. It doesn't happen with OS-B or if you manually trigger the tape load to jitter the timing. Some of the guys in Poland made a synthetic testing tape with 10,000 records and found that about 0.1% of the blocks fail on a real Atari with the XL ver.2 OS and the entire tape was never able to load without patching the OS.
As for why it only happens when you save and reload the tape, the CAS format only stores gaps to millisecond precision, so the blocks get shifted by a tiny amount on save. About 1/7000th of a second is enough to avoid (or cause) the bug.
Adicionalmente, otro usuario en el foro agregó un comentario respecto a otro problema con un archivo cas:
A lo que el creador de Altirra contestó:Regarding to the "Bruce Lee Problem"
It still happens that the baudrate is incorrectly set and the bootloader is aborted.
Perhaps this can help.
In altirra 2.70 I get a pokey-divider of a steady $5BE, where in 2.80 it's a steady $5D4.
Todo era muy extraño. Y lo principal, ¿Por qué con SITRE no tuve estos problemas?Phaeron escribió:Internal precision of the CAS translation is improved slightly in 2.80, so the timing changed. $5BE is the closest divider for 2.70 and $5D4 is the closest for 2.80.
As I've already noted, the current failure on this tape and configuration are due to the timing triggering the bug in the OS ROM on that one particular block. I will not make a change to "fix" this that simply moves the problem to another tape or configuration.
Decidí probar el NHP 2.56 con los otros sistemas operativos:
Spoiler: MOSTRAR
Conversando con mr-atari, respetado usuario holandés, me facilitó un poco comentado OS: REV2-UHSIO (2015), que genera correctos archivos cas:
Hay mucho aún por probar: otros nhps, stacs (pelusasoft), los CAIN, etc.REV2-UHSIO is a build from my own MyBIOS (for Atarimax MyIDE-hardware).
Supports SIO,HSIO,bit7_SIO and UHSIO (with external clock).
Also I changed the tape-loading to support digital-playback devices upto 6000 baud.
PD: Adjunto el RE2-UHSIO.ROM (agregarlo a "Firmwares Images"-> "XL/XE Kernel ROMS")