5 Using Red Hat Commercial Linux

It's called the user group scheme, and it's described in the manual. What it basically does is allow you to use groups to allow multiple people to use a common set of files and edit them, transparently. It goes like this:

• You'd like to have a group of people work on a set of files is say, the "/usr/lib/emacs/site-lisp" directory. You trust a few people to mess around in there, but certainly not everyone.
• So you do `chown -R root.emacs"/usr/lib/emacs/site-lisp'. And you add the proper users to the group.
• To allow the users to actually create file in the directory you do `chmod 775 /usr/lib/emacs/site-lisp'.
• This is fine, but when a user creates a new file it gets the group of the users default group (usually users).
• To prevent this you do `chmod 2775 /usr/lib/emacs/site-lisp' which causes everything in the directory to get created with the "emacs" group.
• This is fine, but the new file needs to be mode 664 for another user in the emacs group. To do this you make the default umask 002.
• Well, this all works good, except that if your default group is "users", every file you create in your home directory will be writable by everybody in "users" (usually everyone).
• To fix this, you make each user have a "private group" as their default group.

At this point, by making the default umask 002 and giving everyone a private default group, you can easily set up groups which users can take advantage of without doing any magic. Just create the group, add the users, and do the above chown and chmod on the group's directories.

We only ship xmh because it is part of the standard X11 distribution. It requires the mh mail package, which we do not ship. We haven't seen much need for it so far, so we haven't spent the time on it. If you have a need for it and would like to see it in future Red Hat releases, please let us know. I won't guarantee anything, but if enough people request it...

Some of you may be accustomed to other distributions and the fact that color ls is installed by default. RHCL does not do that, but it is possible. The best way to find info on it is to do:

man color-ls

man dircolors In short, you can probably put something like ``. `dircolors` '' in your .bash_profile file. You may also need alias ls="color-ls --color=yes" in there.

You need to do make zImage or make zlilo instead of just make. make zlilo is the best choice for most beginners as it will not only build the kernel, but also install it as the default for you if the make goes well. make zImage will build a compressed kernel and place it in /usr/src/linux/arch/i386/boot. You must copy it to your root dir and run lilo yourself to use this one.

If you used make, you got a file called vmlinux. This is an uncompressed kernel image, and will NOT work! LILO will only boot a compressed kernel.

Simple. Read the RPM-HOWTO, available in the docs directory on any official RedHat Mirror.

Those directories are simply a template to use to organize your source code. It happens to be the same as the template we use in /usr/src. If you plan to keep your source in /usr/src as well, you can simply delete /usr/local/src. Red Hat does not use it for anything, it is only meant as a convenience for the user.

Lets say you do the following as root:

gcc -o hello hello.c
hello

Why?

By default, `.' is not in your path. You must run:

./hello

This is for security reasons...`.' should be in normal user's path, but not in root's path.

You need to say 'y' to the line that asks:

CONFIG_MODVERSIONS [n]

When you untar the distribution of Netscape, you will get an nls directory. Depending on where you untar the distribution, you will need to set some environment variables. You should add lines similar to the following to your .bash_profile file:

export XKEYSYMDB=${XKEYSYMDB-/home/foobar/lib/X11/XKeysymDB}
export XNLSPATH=${XNLSPATH-/home/foobar/lib/X11/nls}

You most likely didn't install the kernel sources. Find the sources for the kernel on your system and install them. Then go to /etc and make sure that linux-source is a symlink to /usr/src/kernel/linux-your.version.here. If not, make the symlink (something like ln -sf /usr/src/kernel/linux-1.2.13 /etc/linux-source).

Then you need to cd /usr/src/linux and do: make config make and then you can hit control-C after a few seconds (it makes some symlinks that you might need right at the beginning).

It seems that the standards on what addresses to use for printer ports are a little fuzzy. Different manufacturers are using different addresses. To compound the problem, DOS probes for them in a certain order and assigns the first one it finds to LPT1. This leads people to believe that their printer port under linux must be /dev/lp0. It may not be. You may only have one parallel port and it could be LPT1 under DOS and be /dev/lp0, /dev/lp1, or /dev/lp2 under linux. To find out which it is, turn your printer on and connect it to your parallel port. Then do cat /etc/printcap > /dev/lpX where 'X' is 0, 1, and 2 until you get some output. When you do get output, you have found your printer device under linux.

Now you just need to run the print-tool from the control-panel. To get it, do a 'startx' as root and double-click on the print-tool icon in the control-panel. Use the device you found to setup your printer with the print-tool.

There is currently no Openwindows package for RedHat 2.0. There are people who have asked for it, but we just haven't gotten to it yet. We should have a package sometime, so be patient.

Before asking where some particular piece of software is (like diff for example), please see any official RedHat mirror and look in redhat-2.0/RedHat/rpm-contents.gz. This file has a query of all available RPMs and their contents. grep through it for what you want and it will probably lead you to an RPM that you don't have installed. Use glint or rpm -i to install the proper RPM off your CD-ROM.