Heya Sam.

Please help with this one. I'm a total n00b to Linux programming and might get some of the terminology/logic wrong every now and then (ps where can I read up more about it):

1. I modded courier.spec by adding the following:

touch $RPM_BUILD_ROOT/rootfile root=.. while [ ! -r $RPM_BUILD_ROOT/usr/lib/courier/bin/${root}/rootfile ] ; do root=${root}/.. done

for f in dotlock maildirmake maildrop makedat reformail reformime mimegpg deliverquota do %{__ln_sf} /usr/lib/courier/bin/$f $RPM_BUILD_ROOT/usr/local/bin/$f done

for f in makeuserdb pw2userdb userdb userdbpw vchkpw2userdb do %{__ln_sf} /usr/lib/courier/sbin/$f $RPM_BUILD_ROOT/usr/local/bin/$f done

2. This seemed to count dirs upstream in increments of one, but exited with the error 'no job control' without indicating where in the process there was 'no job control'. After changing a typo and rpmbuilding again it reported the error 'file maildirmake already exists'. I don't know if the typo changed things or if the previous rpm rebuild and 'ln -sf' had changed anything for the next rpm build.

3. Actually ran the ./configure; make; make install sequence without the --with-redhat switch and it installed it!

4. Thought while on a good thing to try rpmbuild again, and bam!.. the rpm was built!

Sequentially it seems as if the courier.spec mod had forced the required symlinks which allowed it to compile without the --with-redhat switch, and since it had installed by running the ./configure; make; make install sequence, allowed rpmbuild to build an rpm.

Which leads me to question if it is a broken RPMBuild or something in the Courier compile process. The rpm-build-4.3.1-0.3.1 patch didn't work so I rolled back to 4.3.1-0.3

Has someone reported the RPM breaking? If not/so, how/where would I be able to report it/look it up/follow it up?

If this builds the required RPM, can I accept this built RPM as acceptable and impliment it?

Thanks again for your help.

Stu@

On Thu, 2004-11-18 at 01:18, Sam Varshavchik wrote: