Re: Finding installed package files
From: Moe Trin (ibuprofin_at_painkiller.example.tld)
Date: Mon, 04 Apr 2005 21:09:05 -0500
In article <7s2dnTBP-evfDc3fRVnfirstname.lastname@example.org>, Rod Engelsman wrote:
>Well I really dislike that the Fedora guys renamed the package. That
>makes it so you can't do a rpm -U with a package directly from
>OpenOffice.org. Their own repository is at least one, sometimes two,
>releases out of date.
Often, the rename is done for a very good reason - changes that make it
incompatible with the original. But again, is there something that you
need in the newer releases?
>It appears that he got his information from the standards you reference.
Linux Standard Base http://www.linuxbase.org/spec/
>there's the rub, IMHO. What's the use of standards if people don't
Standards are wonderful thing;
everyone should have one of his very own
A few years ago, Friday, October 14 was World Standards Day -- in
*some* countries. In America, it was observed on October 11th. In
Finland, it was marked on October 13th. Italy planned a separate
conference on standards for October 18th. - after Shakib Otaqui
>I find this sort of ironic given the emphasis in the open-source culture
>on open standards and adherence thereto. I understand that it's not on the
>same level as Internet standards or file format standards (e.g.,
>OpenDocument), but it would still be helpful.
Part of the blame is market forces. Unlike cars, or TV sets, or breakfast
cereal, much of the product is identical, and you have to do something to
set your product apart. That's why the kernel you get with your distro
is different from just about all others, AND different from the one you
get at kernel.org.
>It's one of the reasons that, for instance, my ISP supports Windows and
>Mac but not Linux. It's not that they're really hostile to Linux, but
>they find it impossible to support the hundreds of distributions and all
>the possible permutations arising from that.
OK, let's look at that for a moment. Both windoze and the Mac are pretty
much standardized. All Linux distributions are actually using the same
ANU ppp package (for the average user, there is little difference between
ppp-2.4.0 and the current 2.4.3), but every idiot and his mule has come up
with a "helper" application, to help people configure pppd. Invariably,
they fsck it up one way or another (you'd think after nine freakin' years
they'd learn not to look for a login prompt any more). Each distribution
has their own, and there is one with the KDE and Gnome desktops, and there
is WvDial (none of which are worth the disk space they occupy, but that's
another rant). So, which one are you going to support? Also remember that
the staff at the ISP help desk often can't even _spell_ ppp, never mind
have a clue how it works. 90+ percent of the customers are using windoze,
and that's what the very underpaid (and undertrained) staff is going to
support. So, when was the last time you rebooted? That's ALWAYS the
>Well... it *is* optional. It's not like a system-level commmand like
>grep or locate; it's an application program and there are options to
>install it or not. So I have to disagree with you there.
Read through the File Hierarchy Standard. It came with the distribution,
as opposed to you paying extra for it, or getting it from a third party
>Well... I know how to do that now. But I didn't at the time. Sometimes
>the problem can be a matter of knowing what questions to ask. You can't
>very well ask *how* to do something if you don't know that the
>"something" can even be done in the first place.
Dinosaur time again - when I started in *nix, there weren't that many
commands to worry about. So a not uncommon trick was
for FILE in `ls $1` ; do
echo $WHAT | grep -q 'nothing appropriate'
if [ $? = "0" ] ; then
echo $WHAT >> /tmp/XXX
Call that script with the name of an executable directory (example 'FOO /bin'
where the script is called 'FOO'), and this script would look at the files
in that directory, and print the header of the man page if one existed, OR
print the filename with the output of the 'file' command. That allowed you to
get a brief one line description of what commands were there, or a clue of
how to look at the file (then, most commonly the ones without a man page were
scripts of some kind). When there were only 3 to 5 hundred commands on the
system, this was doable. But even today, you rarely find someone who uses
more than a tenth of the commands on a system.
>My /usr/bin has 2404 files in it.
>/usr/lib has 2017 files, and 117 folders
>I find other folders like /usr/sbin with 377 files.
>So I'm not sure what your point is here.
You probably will never use more than one command in 15. Hold onto your hat,
this may make your eyes water:
[compton ~]$ history | sed 's/^......//' | tr '|' '\n' | sed 's/^ *//' | cut
-d' ' -f1 | sort | uniq -c | sort -n | tail -20 | column
21 tr 25 whois 35 cd 52 sort 68 vi
22 find 26 diff 39 column 62 zgrep 71 zless
22 wc 26 mv 39 uniq 64 cut 91 ls
24 pack 30 rm 44 sed 65 grep 127 less
'history' has the last 1000 commands I've used in an individual shell (I have
27 terminals open on this desktop, and when counting for history, the above
is considered one command, even though it has ten commands piped together
to make it do what I wanted). But you can see that I'm using a lot of the
same commands. OK, I found one shell where I've actually used 105 different
command - out of the 1300 on this system. That's what I'm referring to.