Thanks for the report, but for it to be really useful, could you send
a patch that fixes the two problems you mention? "diff -u" is the
preferred format. Thanks!
--david
>>>>> On Sat, 27 Feb 1999 21:08:34 +0100, Svend Daugaard Pedersen <sdp@vip.cybercity.dk> said:
Svend> During my work on porting the Epson backend to the Amiga I
Svend> found the following minor bugs:
Svend> - In the call (line 634)
Svend> 632 buf = (u_char *) head; 633 if (s->hw->is_scsi) 634
Svend> receive (s, buf, 4, status); 635 else 636 receive (s, buf, 1,
Svend> status);
Svend> it is expected that the last two bytes read from the
Svend> scanner form a legal 16 bit integer.
Svend> This is not the case on a M680x0 processor.
Svend> - In line 849 (procedure 'attach')
Svend> dummy_dev.sane.model = (char *) memcpy (str, buf + 16 +
Svend> 8, 8);
Svend> 'buf' is not initialized if the scanner is a parallel port
Svend> scanner. (I have not tested it on a parallel scanner but my
Svend> C compiler told me that there is a potential problem).
Svend> The ported backend has been reported to work with the
Svend> following SCSI Epson scanners:
Svend> GT9000, GT8500, GT7000, GT6500 and Perfection636
Svend> A little addition around line 750 had to be made to make the
Svend> backend recognize the Perfection scanner.
Svend> Svend Daugaard Pedersen
Svend> -- Source code, list archive, and docs:
Svend> http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe
Svend> sane-devel | mail majordomo@mostang.com
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com