Building Subversion with Apache is not for the faint
of heart...
I just crossed the hurdle you are stuck on about 3
days ago... I'm working on Solaris, not GNU/Linux, but
I suspect the issue is the same.
First, Apache version 2.0.49 now incorporates the
changes to recognize BerkeleyDB.4.2 so you no longer
need to run the patch listed in the FAQ. You do need
to be certain that you are setting LD_LIBRARY_PATH to
include the berkeley lib and include directories. I
was able to leave out the line
"--with-berkeley-db=/usr/local/BerkeleyDB.4.2" but I
did need the "--with-dbm=db42" line.
However, I STILL got the same error you are getting
until I made one important discovery... there seems to
be an issue with which version of the compiler you use
to compile the BerkeleyDB...
The way I finally solved it was by downloading
BINARIES of Berkeley from sunfreeware.com and not
trying to compile my own (obviously that site won't
help you, but you should be able to find GNU/Linux
binaries somewhere -- I mention it because I want to
help the next poor schlub who tries to google this
problem). I'll leave as an exercise for the reader to
figure out which versions of what are needed to
successfully get Berkeley binaries that will work..
Once I got Apache and Berkeley to play nice I ran into
a problem with the Subversion build itself. For some
reason the compilation gets confused and tries to
point multiple files at the same alias... When I run
"make" the process seems to finish, but when I try
"make install" it fails because some files don't
exist.
I'm going to try again later to see if I can get past
this hurdle and either post what I learn or else the
logs of what happened to see if anyone can help.
All in all this has been one of the most horrific
installs I've ever tried... The Subversion client is
pretty easy, but getting a server with apache to work
is, well, WORK! :-) Seriously, it has taken days and
days of trying to resolve dependencies. On my box it
takes about 15-20 minutes for "./configure" for apache
to run, so tweak and try has dragged on and on. I can
be pretty stubborn, so I've kept trying. It is
becoming less of a priority, however, since I've been
forced by schedules to revert to CVS for this project.
I still want to get Subversion working, because I
prefer its architecture. So far I've seen two major
deficiencies in Subversion 1) difficult to get server
working and 2) Can't store files on NFS. In my case
the box doesn't have huge storage, but I have access
to NFS shares that DO have huge storage and are backed
up frequently. On balance, however, CVS' deficiencies
still out weigh Subversions... As I understand it this
problem is a problem with BerkeleyDB. I wonder if
there is any chance of or plans for de-coupling
Subversion from Berkeley and making it more database
agnostic?
sorry for rambling...
jeffa
__________________________________
Do you Yahoo!?
Yahoo! Small Business $15K Web Design Giveaway
http://promotions.yahoo.com/design_giveaway/
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Thu Apr 1 15:53:25 2004