On Sun, 27 Aug 2006, Julien BLACHE wrote: > Hi, > > I've got a patch (part of which I erroneously committed to CVS some > time ago...) that makes sane-find-scanner tell the user which backends > can be used with the USB scanners that it just detected. > > It's based on a table generated by sane-desc from the description > files (really, 2 tables, one for backends included in sane-backends, > and the other for external backends when we have description files for > them). The tables eat up some space, so the sane-find-scanner binary > gets bigger with this patch.
my only concern there would be if the user sees the name of an external backend, and assumes it is already installed. also- what about the case where a backend is installed, but disabled in dll.conf? > > I'm looking for feedback on that idea; I know it's basically a good > idea and it needed to be done, but the implementation is pretty > static, which may not please everyone. Probably having a directory > where external backends can drop a file would be better, so parsing > the desc files from sane-find-scanner looks like a better idea (would > just need to generate the tables at runtime from the desc files, which > would only duplicate some code from sane-desc). > are the desc files currently installed anywhere? does it make more sense to roll them up into structs in a .so that we could load? that reduces the runtime processing, while allowing external backends to provide additional, smaller files... though honestly, i am not all that sure of the benefit. i mean, if you run sane-find-scanner and see it, but scanimage -L does not, then the backend support is missing anyway? allan > Thoughts ? > > JB. > > -- "so don't tell us it can't be done, putting down what you don't know. money isn't our god, integrity will free our souls" - Max Cavalera
