Henning Meier-Geinitz wrote: > > Hi, > > On Tue, Mar 19, 2002 at 10:58:31PM +0100, Oliver Rauch wrote: > > The testbackend prints a warning: > > warning : invalid size for option halftone-pattern > > > > for this option: > > /* halftone pattern */ > > scanner->opt[OPT_HALFTONE_PATTERN].name = SANE_NAME_HALFTONE_PATTERN; > > scanner->opt[OPT_HALFTONE_PATTERN].title = SANE_TITLE_HALFTONE_PATTERN; > > scanner->opt[OPT_HALFTONE_PATTERN].desc = SANE_DESC_HALFTONE_PATTERN; > > scanner->opt[OPT_HALFTONE_PATTERN].type = SANE_TYPE_INT; > > scanner->opt[OPT_HALFTONE_PATTERN].size = 0; > > scanner->opt[OPT_HALFTONE_PATTERN].constraint_type = > > SANE_CONSTRAINT_RANGE; > > scanner->opt[OPT_HALFTONE_PATTERN].constraint.range = &u8_range; > > scanner->val[OPT_HALFTONE_PATTERN].wa = scanner->halftone_pattern; > > > > I think a constraint_range can have a size of 0 if there is nothing to set. > > In this case the option also is set inactive. > > Same code is in mustek so I thought about it a bit yesterday: > <nitpick> The standard says: "The size must be a positive integer value > of SANE_Word". I don't think 0 is a positive integer... > And "inactive" is only for the setting/reading of the option value, > not the option descriptor. > </nitpick> > Probably in practice it doesn't matter but it won't hurt to e.g. use the > smallest possible size.
Hello Henning, if 0 is a positive integer number is a question that is often discussed. I think it does not make sense to set this value to 1 because in this case the backend has to allocate memory for the value what makes no sense. I think we should explictly allow to set this value to 0 Bye Oliver -- Homepage: http://www.rauch-domain.de sane-umax: http://www.rauch-domain.de/sane-umax xsane: http://www.xsane.org E-Mail: mailto:[email protected]
