Henning, thanks for the review. > Ok, looking at sceptre.c: > > * better use #include "../include/sane/*" for the sane headers. It's > not really necessary but makes clear that the local headers are used > and not the ones that may be already installed.
Done. > * assert can make trouble on AIX. The easiest way is to avoid it. > Otherwise, please use the internal "lassert.h" header (see e.g. > sanei_scsi.c). Done. Used lassert.h > * there are some warnings about signed/unsigned. i don't know if they > are a problem but it should be possible to fix them with casts. All warning fixed. > * the coding style isn't quite GNU (see backend-writing.txt). I'm not > that strict about this and there might be other opinions but I think > same/similar coding style is a good thing (tm). indent should be able to > "fix" it. Done (reluctantly). > Concerning sceptre.conf: > * Don't you want to add a "scsi vendor product" line? This allows > auto-detection at least on Linux, even if there is no /dev/scanner link. > sanei_config_attach_matching_devices does all the magic. Line added. > Sceptre.desc: > * version number is without brackets when the backend is included into > CVS Done. > * version number should represent the version that's returned to > sane_init (something like 1.0.5). Done. > * status is :NEW after the inclusion until after the first release > that includes your backend Done. > * what does "hardware KINPO" mean? This is in the SCSI inquiry? Is > this an alternative name? What about the "Vividscan"? If it's part > of the official name, maybe you can add it also to sceptre.desc so > people searching for it can find the pagge? KINPO is the vendor returned by the SCSI inquiry. However they do not sell scanners directly. That's why I called the backed Sceptre, not Kinpo. > doc/Makefile.in: > * add sceptre (the directory) to BACKDIRS Done. New diff is: http://www.geocities.com/harikus2000/sane/sanecvs-sceptre-20020220-2.diff.gz Frank.
