What I mean is, I intend to work a bit on an open source project which I already have installed through my system packages. What I'm wondering is if there is a suggested workflow for managing this situation, where I want to keep my system-installed packages but also compile and run the version I'm developing.
My question is general as I've wondered this before with other projects, but this particular instance drove me to ask this question.
The project I'm referring to is rtorrent. The project has a other dependencies of course, but in particular it has one by the same author: libtorrent. The point is that both of these things are newer than my system's packages, so I need to pull them both and build rtorrent against that libtorrent version.
So far I have compiled libtorrent, short of actually doing
make install since that is where it would clash with my system packages I imagine. It ended up dumping the library files in
src/.libs, but I think I will also need access to the header file(s) as well right?.
How should I go forward with building this git-cloned rtorrent? I've never done this before so please excuse the ignorance, but does it all revolve around the
./configure script? Should I simply use use
LDFLAGS to add the library path with
CPPFLAGS to add the include path (for header(s)) with
-I? Or should I re-
./configure libtorrent with a
--prefix common to libtorrent and rtorrent, and then
make install libtorrent?
./configure --help shows these "influential environment variables"
libtorrent_CFLAGS C compiler flags for libtorrent, overriding pkg-config libtorrent_LIBS linker flags for libtorrent, overriding pkg-config
What I'm curious about is, this gives me the impression that even if I were to set
CPPFLAGS, ultimately it gets the information for libtorrent using
pkg-config anyways, so I have to use those instead. Using those did result in a successful
./configure execution, along with a
--prefix parameter I think I'm good to go.