Hi, On Wed, Nov 21, 2001 at 12:32:37AM +0100, Franz Bakan wrote: > 1. mustek.c and microtek2.c > > As you probably allready know, the code for backends using fork() > needs small changes to work with OS/2. > I can provide diffs for mustek.c, mustek.h and microtek2.c, based on > the work of Yuri Dario, if you are willing to include them.
Just send the mustek stuff to me. If this doesn't break anything, I will include it. Without the diffs, does compilation fail or do the backends just don't work? > Additionally sanei/sanei_thread.c would have to be added to > sanei/makefile.in I have just committed this to CVS. > 2. sanei/sanei_scsi.c > > I would like to make some changes here (OS/2-things only): > - remove semaphore() code because it is no longer necessary. > (This breaks backwards compatibility with aspirout.sys 1.01 > aspirout.sys beta2 or later from Paul Ratcliffe is needed then.) > > - add: > #if USE == OS2_INTERFACE > # define MAX_DATA (64*1024) > #endif > But I'm not shure here. I did not have troubles with this, and it speeds up > the things slightly, but perhaps there is a reason I don't know why the > former (implicit) value was 32*1024. > > - add some debug-code > > sanei_scsi.c modified this way works at least for me (EPSON and Mustek) > and a Microtek-scanner of another OS/2-Sane-User. As long as you don't touch non-OS/2-parts of the code, just do it. Or send patches, if you don't have CVS-access yet. > 3. Allthough I get SANE built on OS/2 it's a little bit tricky. > OS/2 executables need the .exe extension to work. > So in theory things should improve if > > AC_OBJEXT > AC_EXEEXT > > were added to configure.in > I don't know much about these things, but as far as I know > this would also need autoconf 2.13 at least. > At the moment configure.in tells: > AC_PREREQ(2.10)dnl dnl Minimum Autoconf version required. This is no problem. I have used 2.13 for month now. What additional changes are necessary in the Makefiles? Bye, Henning