On Sat, Jan 21, 2006 at 07:07:21PM +0000, Philip Kendall wrote: > On Sat, Jan 21, 2006 at 03:50:24PM +0000, Stuart Brady wrote: > > I think this should have been IM1, not IM0. Sound right? > > Yes. IM0 is essentially unused on the Spectrum. > > > I think this should be set equal to iff1? > > Almost certainly. The only time they differ is after an NMI is > acknowledged with interrupts enabled.
Thanks! I've just noticed a rather worrying level of similarity between the +D snapshot code and my new Messenger code. In a Messenger snapshot, the 22-byte header is located at offset 0x3cea, and the RAM dump starts at 0x4000. The Messenger header appears to be identical to the +D header. However, libspectrum's +D code appears to have some bugs, as I reported in http://sourceforge.net/mailarchive/message.php?msg_id=10654899 The first three listed in that message appear to be genuine bugs, as the Messenger's code differs from FUSE's. The last one is probably a bug in Ramsoft's documentation. I have 15 Speclone snapshots (from 5 programs). (Speclone was another Spectrum emulator on the SAM.) Does anyone know whether Speclone used +D snapshots? I had assumed so, only they're 49461 bytes long, and I had expected +D snapshots to be 49152+22 bytes. Cheers, -- Stuart Brady
