On Tue, Feb 24, 2004 at 04:35:49PM +0100, Marco Tamanti wrote:
> Compiling a qmail-1.03 patched with qmail-ldap-1.03-20040201.patch I 
> had to do some little changes to the Makefile to elude some errors.
> I don't know if these errors happen in every OS or just macosx (I'm on 
> 10.2.8).
> 
> 1)
> To prevent the error:
> /usr/bin/ld: Undefined symbols:
> _quota_calc
> _quota_get
> _quota_rm
> 
> I added the string "maildir++.o seek_cur.o seek_end.o seek_trunc.o" to 
> the string containing "load qmail-pop3d".


> 2)
> To prevent:
>  make: *** No rule to make target `qmail-pbsdbd.run', needed by `ldap'. 
>  Stop.
> 

1 and 2 are definitifly a bad patch. I just checked with the source on
www.nrg4u.com and the necessary patch hunks for those two are in the
patch.

> 3)
> The last problem was a macosx one: name of directories are case 
> insensitive, so the creation of file "install" overwrites "INSTALL".
> I used the following procedure:
> mv INSTALL INSTALL.txt
> make setup
> mv install install.go
> mv INSTALL.txt INSTALL
> ./install.go
> 

Known issue with HFS disks. Either use a UFS partition for that (I use
that) or have a look at newfs_hfs I think. There is a "turn of the stupid
case insestive behaviour" flag. My mosx is currently not running so I
can't tell you where I found it precisly.

> I still haven't tested the service, but the compilation ended without 
> errors. Just some warnings:
> chkshsgr.c:10: warning: passing arg 2 of `getgroups' from incompatible 
> pointer type
> chkshsgr.c:10: warning: passing arg 2 of `setgroups' from incompatible 
> pointer type
> 

Known issue and there is no easy way around it. chkshsgr is just a helper
app and does not get installed so you can ignore those errors.

> qmail-send.c:1448: warning: passing arg 2 of `substdio_fdbuf' from 
> incompatible pointer type
> qmail-send.c:1449: warning: passing arg 2 of `substdio_fdbuf' from 
> incompatible pointer type
> qmail-send.c:1502: warning: passing arg 2 of `substdio_fdbuf' from 
> incompatible pointer type
> 

These should be fixed now. IIRC this was just a different function
declaration.

> Hope this would be useful to someone, now I'm going to test the whole.
> 
Happy testing.

-- 
:wq Claudio

Reply via email to