On Sunday, March 17, 2002, at 09:12  AM, Christopher C. Lamer wrote:

> How can I install a program that is not recognized by fink?
> I've tried installing a few things such as dillo and other programs.
> I have tried typing "cd / the location of the file" then typing
> "./configure. After doing this, something seems to happen but once it's
> done, I have no clue what happened, if it installed, or how to use it!

While Fink provides some of the best ports out there, it doesn't hurt to 
understand how to do it "by hand" if you should ever need/want to.  For 
instance, I wanted to customize my MySQL installation somewhat and so I 
did it this way, also I wanted Python 2.2 before it was available via 
Fink so I also compiled it "by hand".

It's pretty easy:  you do the ./configure part, which you already know 
about.  One thing you should know about this is that you're not 
executing a command on your system called "configure", rather you're 
executing a shell script that is traditional packaged in with the source 
code.  That's why you do the "./" part (also why some say having . in 
your PATH environment variable can be troublesome).  To find out what 
kinds of customizations (parameters) you can add to this script as 
arguments, do "./configure --help".  That's the traditional way to get a 
list of what you can pass as a parameter to "configure" script.

After that, you need to actually build the binary from the source.  That 
is done with a command on your system, "make".  This just builds the 
binary, there may be optional arguments to this command but if so I 
don't know about them nor have I ever used them.

Finally, most (not all) binaries can/need to be installed into a certain 
part of your system using the "make install" command.  I suppose 
"install" is an argument to "make", though that's just something I 
thought of now and might not be true.  RTFM would tell me, but I'm 
eating breakfast right now.  This command usually needs to be executed 
as root or with the "sudo" command, so that you can write into parts of 
your system that your regular user account doesn't have write access 
to.  The location of where the install puts the binary stuff is 
determined usually by an argument to the ./configure script 
(--prefix=/path/to/installation, but YMMV so check ./configure --help 
for more info).

It's easy, and a lot of programs build right out of the box on Darwin 
without needing to be ported.  Just make sure you have the developer 
tools installed.  This is good to know if, for some reason, you want to 
add some functionality to a program that wasn't compiled in.  For 
instance, I'm not sure which ./configure parameters are set in the PHP 
binary you can get from Marc Liyanage (maybe it says on his page) or in 
the Fink build of the same name, but if you wanted --with-xslt-sablot 
but it's not compiled in then you have to do it yourself.  (Actually, 
the PHP binary is a bit complicated since it's compiled into Apache 
[depending on how you do it], so that's probably not a good example but 
you get the idea.)

When I first realized I was going to have to compile my own software on 
a Linux box I was kind of intimidated but to be honest with you I never 
use RPMs anymore.  Although remember that Fink provides most of what you 
may need on Darwin, especially for the trickier ones that require 
porting before they will build correctly.

HTH,


Erik


_______________________________________________
Fink-beginners mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-beginners

Reply via email to