SIPB ClueDump #1

How To Build Random Software

Credits and Blurb

Run by: mkgray, yandros, ghudson

Written by: bobmacke

In this session we will talk about finding and building `interesting' programs on the net. We may touch on some issues involved in installing and maintaining programs as it's normally done in the sipb and outland lockers.

NOTE: A lot of the means of usage shown here are Athena-specific. If you are on another system, you will need to find out what to do to use these services (assuming your system offers them).

In This Cluedump:

First things First: Finding Stuff

 Before you can build cool things, you have to find them. FTP works well for this.

ncftp (or ftp)

Usage: 
	add sipb
	ncftp <location>

(for example)

	ncftp ftp.x.org:contrib/...

You should be prompted for a username and password. Most places accept anonymous as the user, and prefer your email address as the password.

At this point you should receive the ftp> prompt. Here are the commands that will work in ftp:

Where to Find Cool Stuff

 So, now you know how to use ftp to find cool stuff, but where to ftp to? Where is the best cool stuff located? archie <foo> finds stuff for you! This searches prospero for the file, and returns the sitename, pathname, and sometimes a little more info. Sometimes, there are a zillion entries listed. In that case, check the hostname endings to find the most "local" host. FTP to the sites with the "nice" endings (.mit.edu, .edu, .gov, etc.) and avoid "weird" endings which could be overseas or something.

If the list is way too long, you can try archie -server <foo> to narrow down the search.

Building Stuff

 Okay, you finally found the cool thing you want. Now you have to build it. How? Read on...

README

Before attempting to build/configure anything, check the directory from which you are getting stuff to see if there is a file called README or sometimes INSTALL. (Actually, any name in all caps should be checked.) You should read these files before anything else. They may have important, helpful, or vital information about installing the stuff in the directory. Sometimes they only have pointless info, but always read them anyway.

gunzip

Gunzip is used to decompress files that are compressed. Compressed files have endings of .tar.gz, .tgz, and sometimes .tar.Z

	add sipb
	gunzip <file>.tar.gz (or .tar.Z or .tgz)

	(Returns <file>.tar)

tar (and gtar)

 Sometimes, what you want is packaged in a neat and tidy file that, because it is just one file, is easy to port around and therefore easy to put it where you want it. However, that one file is actually a package of all the different files required for the software you want to build. To separate it out into the separate files (such that you can actually USE it), you want to use tar.

	tar <commands> <file>.tar

	(example):  tar xvf xteddy.tar
Commands: One thing you want to consider is that you probably should create a subdirectory in which to put the files when you use tar. Otherwise, (at best) you will wind up with a mess of files in your home directory or (at worst) some of the files will have the same name as files that already exist, and those files will be written over.

gtar - Should work similar to tar, except it adds the -z option which runs gzip/gunzip (after compressing)/(before decompressing). Under Linux, gtar is the default and is called "tar" 

	add outland
	gtar <file>.tar.gz

./configure

 This will usually be described in README (yet another argument for reading README before attempting to build anything.)

xmkmf

 
	add x11[m]
	xmkmf <foo>
xmkmf stands for xmake makefile, sometimes known as IMakefile.

Some Makefiles require gmake (which adds a bunch of stuff that not all makes have)

make

In theory, make should nicely build software for you without any trouble. Just type make <foo> and you have a nice bit of software built for you. yay. However, in practice make doesn't always work that way. (boo)

Native Compilers

 Native Compilers are usually a pain in the butt.

	<foo>.c Line 427: Identifier TIOCLN!... undeclared
    (solution:) grep <foo>.c TIOC /usr/include/*.h
bobmacke@MIT.edu