I had choosen sqwebmail 3.5 because I used a Qmail Step-by-Step procedure which shows this version. Now I downloaded SQWM 3.6 and the installation works *almost* fine.
Here is the error I got at the end of the "make install" or "make install-strip" process:
...
make install-data-hook
mksoftlinks /usr/local/share/sqwebmail/html en-us
/bin/sh ./../mkinstalldirs /usr/local/share/sqwebmail
/usr/bin/install -c -m 644 ldapaddressbook.dist /usr/local/share/sqwebmail/ldapaddressbook.dist
/bin/sh ./../mkinstalldirs /var/run/sqwebmail
test -w /etc || exit 0; chown bin /var/run/sqwebmail
chown: bin: invalid user name
make[4]: *** [cache-reminder] Error 1
make[3]: *** [install-am] Error 2
make[2]: *** [install-recursive] Error 1
make[1]: *** [install-recursive] Error 1
make: *** [install-strip] Error 2
"bin" in chown is not a user, even in Belgium :-)
I searched for all "chown" strings and I found the stupid line in sqwebmail3.6/authlib/makefile.in:
install-data-local: install-authldaprc install-authdaemonrc install-authmysqlrc install-authpgsqlrc
test "$(man_LINKS)" = "" && exit 0; \
$(mkinstalldirs) $(DESTDIR)$(man7dir) ; \
for f in $(man_MANS) "" ; do \
test "$$f" = "" && continue ; \
$(INSTALL_DATA) $$f $(DESTDIR)$(man7dir) ; done ; \
cd $(DESTDIR)$(man7dir) || exit 1 ; \
for f in $(man_LINKS) "" ; do \
test "$$f" = "" && continue ; \
rm -f $$f; $(LN_S) authlib.7 $$f ; \
test -w /etc && chown bin $$f && chgrp bin $$f ; \
done ; exit 0
Can you see what is wrong?
Thanks, Guy
Oh!!
(Great googly-moogly. No wonder.)
Well, this explains all your other problems.
If you can... I strongly suggest upgrading to 3.6.0 instead!
If you can't (and if not, why not?!), then yes... you'd still need the mime.types setting. -- And yes, the numlib stuff would be a problem too.
I made a patch which Sam applied to the numlib files a while back (after he figured out what the problem was for me; thanks again, Sam! :) -- and I'll try to send it if you can't upgrade. (Tell me if so, and I'll look around for it. But I really do encourage you to upgrade to the latest version. ... Really! ... Quite strongly!)
As for the global vars, they are useful with at least a few other packages, but probably not all. I'm sure that in some cases, they might (potentially) even cause problems instead of fixing problems. -- I haven't seen that to be the case yet myself, but with all the variety of stuff out there, I'm sure it's true for something. ;-)
-jab
