From: Bertrik S. <be...@zo...> - 2004-01-04 23:11:37
|
Hi all, Henning Meier-Geinitz sent an e-mail to the sane-devel list with some issues around the niash backend before inclusion into sane. Here's the status on those issues: 'DONE' (fixed but not committed yet, can't access sourceforge cvs) * license: I now have agreement of all developers who contributed code to the backend to change the license to GPL + sane exception. I'll change the license and add the proper headers to the source files. * niash.desc should be moved into doc/descriptions, not backend. * niash.conf contains reference to /dev/usbscanner. Currently we don't need the .conf file anymore. I'll delete it. * compilation didn't work because of C++ style comments. Fixed. * niash.c include order of "sane/sane.h" and "sane/config.h". Fixed. TODO: * check syntax of .desc file. It's outdated. make will complain about the issues. * update manpage, use other manpages as example. * run 'indent -gnu' over source files. * check backend with frontend/tstbackend.c. I did this once when tstbackend was just released, but we should test it again and properly. * update AUTHORS and doc/sane.man * 'illegal' symbols when doing './configure --enable-static'; 'make'; 'make libcheck'. Henning's descriptions of the issues are quite clear and they should not be too hard to fix. Henning, can you clarify the last problem, about the 'make libcheck' warnings? As far as I understand, the problem is that there are 'global' symbols seen by the linker other than the sane API function names, right? There is indeed some trickery in the sources where .c files are directly included (and the corresponding .h files also). If I make all offending functions static, I think I have a problem keeping the sources compatible with my testtool program (which I like to keep working for low-level testing) Is there a better/other way to do this? All the best, Bertrik Sikken |