On Tue, 2013-01-01 at 17:07 +0100, Louis Lagendijk wrote: > On Tue, 2013-01-01 at 16:47 +0100, Louis Lagendijk wrote: > > On Sat, 2012-12-15 at 14:08 +0100, Louis Lagendijk wrote: > > > Hello, > > > I have made a simple patch to saned that gives full systemd socket > > > activation. The big advantage is that it allows debug output to go to > > > systemd instead of /dev/null. I do however need some changes to > > > configure.in. I am not sure how to get these in git as I know that I > > > cannot simply run an autoreconf as that kills our ltmain.sh changes. > > > The autotools on my Fedora 17 seem to be newer than what was used to > > > generate the current configure etc. > > > > > > Could somebody give me some hints on how to update the autofoo stuff or > > > even better do it do? > > > I would also would appreciate a review of my changes to configure.in, as > > > we seem no to use pkg-config and AC_CHECK_MODULES a style I am not used > > > to. > > > I have attached configure.in and frontend/Makefile.am as well as > > > saned.c. > > > Help is very much appreciated > > > > > > Kind regards, Louis > > Does anybody have any objections to the changing saned.c? The systemd changes do allow for all debugging output (backend debugging for example) to be captured by systemd, so I believe that they the changes could be quite useful. They are completely backward compatible (even on a systemd based system we do not need to use them if we do specify StdInput to be the inetd style socket.
I now found a machine that has an older automake 1.11 that did allow me to regenerate the Makefile.in files. I ran autoconf, automake and autoheader. That did regenerate all Makefile.in files. Should I check them all in into git? Or should I only check-in frontend/Makefile.in, include/sane/config.h.in and saned.c? kind regards, Louis
