Linux-Development-Sys Digest #889, Volume #7 Mon, 22 May 00 02:13:08 EDT
Contents:
Re: linux tools advice (Kaz Kylheku)
Re: RH6.1/gcc2.91: #include system search (Daniel Kiracofe)
Re: Checking for I'm swapped (Daniel Kiracofe)
Re: linux tools advice ([EMAIL PROTECTED])
Re: RH6.1/gcc2.91: #include system search (Shaun Arral)
What !@#$ moron colorised g++? (Thaddeus L. Olczyk)
HELP: 8 Bit Linux ? (Dan Mathias)
shutdown -r now ("Sake")
source code of shutdown ("Sake")
2.2.14 and Michael Hasselstein's NAT - kernel OOPS (Steve Frampton)
Re: HELP: 8 Bit Linux ? (Johan Kullstam)
Re: Need ideas for university funded project for linux (David Steuber)
Re: linux tools advice (David Steuber)
Re: Need ideas for university funded project for linux (Christopher Browne)
Re: Need ideas for university funded project for linux (Christopher Browne)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Kaz Kylheku)
Crossposted-To: comp.os.linux.development.apps
Subject: Re: linux tools advice
Reply-To: [EMAIL PROTECTED]
Date: Mon, 22 May 2000 02:24:30 GMT
On Mon, 22 May 2000 08:54:15 +0800, sllai <[EMAIL PROTECTED]> wrote:
>what do I need to develop an GUI application in linux system. Must I develop
>it in linux environment or can I do it in the PC environment and then
>transfer it into the linux based system.
Linux runs on PC's, therefore it is a ``PC environment''.
--
#exclude <windows.h>
------------------------------
From: Daniel Kiracofe <[EMAIL PROTECTED]>
Subject: Re: RH6.1/gcc2.91: #include system search
Date: Sun, 21 May 2000 22:58:08 -0400
Shaun Arral wrote:
>
> I wanted to look at <sys/errno.h> and noticed quite a few of them.
> Now I'm wondering what should be my #include search path be setup as
> (including X) in my IDE??
> Or where do i find this i man'd: gcc,make, looked through HOWTO's.
AFAIK <sys/errno.h> usually refers /usr/include/sys/errno.h
--
/* Daniel */
E-mail: [EMAIL PROTECTED]
Webpage: http://www.cis.ohio-state.edu/~kiracofe
"Fear is only afraid of the absence of itself" - Mediocrates
------------------------------
From: Daniel Kiracofe <[EMAIL PROTECTED]>
Subject: Re: Checking for I'm swapped
Date: Sun, 21 May 2000 23:02:02 -0400
"Dmitry A. Antipov" wrote:
>
> Hello all,
> I have a PID of running process. How can I check for it's swapped or
> not ?
Linux, AFAIK, never swaps whole processes out. It only pages. That
is, it will write 4k (on intel) out a time, but never the whole thing at
once. Therefore, given any pid, the process is not swapped out...
--
/* Daniel */
E-mail: [EMAIL PROTECTED]
Webpage: http://www.cis.ohio-state.edu/~kiracofe
"Fear is only afraid of the absence of itself" - Mediocrates
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.development.apps
Subject: Re: linux tools advice
Date: 21 May 2000 19:22:27 -0700
In article <[EMAIL PROTECTED]>, "sllai" says...
>
>what do I need to develop an GUI application in linux system. Must I develop
>it in linux environment or can I do it in the PC environment and then
>transfer it into the linux based system.
Qt, GTK+, Tcl/Tk, Java/swing to name few.
what do you mean by "pc environment" ?? Linux runs on the "pc" too.
Nasser
------------------------------
From: Shaun Arral <[EMAIL PROTECTED]>
Subject: Re: RH6.1/gcc2.91: #include system search
Date: Sun, 21 May 2000 21:10:22 -0700
Daniel Kiracofe wrote:
> Shaun Arral wrote:
> >
> > I wanted to look at <sys/errno.h> and noticed quite a few of them.
> > Now I'm wondering what should be my #include search path be setup as
> > (including X) in my IDE??
> > Or where do i find this i man'd: gcc,make, looked through HOWTO's.
>
> AFAIK <sys/errno.h> usually refers /usr/include/sys/errno.h
>
> --
> /* Daniel */
> E-mail: [EMAIL PROTECTED]
> Webpage: http://www.cis.ohio-state.edu/~kiracofe
>
> "Fear is only afraid of the absence of itself" - Mediocrates
That file has 1 line in it:
#include <errno.h>.....
never mind i followed the trail of the "errno.h" files that include
"errno.h" in this order (called:)/usr/inclued/sys -> /usr/include/errno.h
-> bits/errno.h -> linux/errno.h -> (/usr/include/<like the rest of
course>)asm/errno.h which is where i finally got what i was looking for.
and the ones in the /usr/lib/bcc/include/* directory are from "dev86.rpm"
??? no dependencies so it's gone...
Thanks
------------------------------
From: [EMAIL PROTECTED] (Thaddeus L. Olczyk)
Subject: What !@#$ moron colorised g++?
Date: Mon, 22 May 2000 04:06:36 GMT
Reply-To: [EMAIL PROTECTED]
Sorry, but this is one is just pissing me off.
I hate Microsoft. I hate them because I think they produce crappy,
buggy software and systems that force other programmers to produce
crappy, buggy software. One of the major reasons for this is that
Microsoft produces software that is "easy to use". The way that
Microsoft produces "easy to use" software also causes two things:
1) those that want to do sophisticated things, find them hard to do,
2) the programmers that march in lockstep to Microshit do well,
but those that don't have a hard time. Examples include things
like not create a new "file/project/window" in an MDI app each
time you open it, automate your building and testing in such a way
that you can check out one script and run it on an arbitrary
machine and other such things, basic "point and click" system \
adminstration, but when you really want to secure something you
have to spend hours. These force programmers on Windows
to mostly work in "hack" mode. Write the code to a feature, test
if the feature works ( but never test if it breaks anything else
). If a person reports a bug, fix the bug, but never test to see
if you've created new bugs.
The thing is that I don't object to making computers "easy to use".
What I object to is making computers "easy to use" for less
sophisticated users, while making it "hard to use" for more
sophisticated users. And I don't think you have to make it hard
to use for the experts to make it easy to use for the nonexperts.
Unfortunately Linux seems to be going this same way. I suspect
this is due to MS hating programmers who have come over to linux
and dragged their MS mentality with them. For those who want to
change linux to Windows, I say this: fuck off. Linux doesn't need you
and is better off without you. There are those who will improve Linux
without making it hard to use for experts.
This I think is a praticular case in point. Programmers are supposed
to be expert users. What do they need to have error messages colorised
for? Frankly I don't want to see people who need crutches like having
the error message, file name, and file line number in different colors
programming for Linux. And the !@#$%^& colorisation is screwing
me up.
Error messages piped to a file have all sorts of crappy escape
codes embedded, making it impossible to read. Error messages generated
when compiling from emacs are screwed up for the same reason.
I can't use next-error. Can anyone help me to fix this?
------------------------------
From: Dan Mathias <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: HELP: 8 Bit Linux ?
Date: Mon, 22 May 2000 00:12:08 -0700
Hello,
Is there any 8 Bit Linux or Unix for the 68HC11 or 8088 cpu ?
Thanks
--
Dan
------------------------------
From: "Sake" <[EMAIL PROTECTED]>
Subject: shutdown -r now
Date: Mon, 22 May 2000 00:14:22 -0400
Hi,
I compiled a single floppy ramdisk only system for my K6 machine out of the
RH6.0. Everything works fine except the following:
When I issued the "shutdown -r now" command, I got "You don't exist. Go
away"
message. and of couse the command doesn't reboot my machine. The machine is
set
to run level 2 and no getty nor login are involved. The ini calls /bin/ash
directly.
Did I miss something ? PLEASE HELP !!
Any e-mail suggestions to [EMAIL PROTECTED] is greatly appreciated.
Thanks in advance.
------------------------------
From: "Sake" <[EMAIL PROTECTED]>
Subject: source code of shutdown
Date: Mon, 22 May 2000 00:31:16 -0400
Hi,
Where can I find the source code of the shutdown utility and/or its
relatives
(halt, reboot, poweroff)
Reply via e-mail is appreciated
[EMAIL PROTECTED]
Thank you very much
------------------------------
From: Steve Frampton <[EMAIL PROTECTED]>
Subject: 2.2.14 and Michael Hasselstein's NAT - kernel OOPS
Date: Mon, 22 May 2000 04:34:27 GMT
Hello:
I've been using kernel 2.0.38 with Michael Hasselstein's NAT patch to
provide 1:1 static NAT in a production environment. As the server seems
to hang under network load (over 100 clients mostly masq'd with about a
dozen using the NAT functionality, the server connected to a 100 Mb/s
switch crashes randomly, but when connected to 10 Mb/s switch it seems
to hold up okay), I would like to replace it with some new hardware
running a modern distro with kernel 2.2.14.
I've got the new box ready to go, and have patched 2.2.14 with Michael's
patch as well as the Solar Designer patch from www.openwall.com. I've
noticed that sometimes (very randomly, but I was able to duplicate it --
perhaps coincidentally -- with a small app that kept malloc'ing 1K of
memory until failure) I've noticed some kernel OOPS (usually but not
always when trying to see the output of /sbin/route). I typed "ksymoops
-m /boot/System.map" and I have attached the output here. It seems to
indicate the failure is occuring in the NAT patches. I looked at the
NAT code but I haven't gotten *THAT* low level into Linux yet. ;-)
I'm wondering if this is a known issue. Also, is the kernel OOPS output
a new feature of 2.2.x (perhaps this bug is the same as is causing the
hard crashes in 2.0.38 but I don't get to see any output?).
Unable to handle kernel paging request at virtual address 49545241
current->tss.cr3 = 02319000, %cr3 = 02319000
*pde = 00000000
Oops: 0000
CPU: 0
EIP: 0010:[<c0151110>]
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010206
eax: ced041e0 ebx: cec18720 ecx: ced041e0 edx: ced041e0
esi: 49545241 edi: 00000049 ebp: cf3ed700 esp: c232dd9c
ds: 0018 es: 0018 ss: 0018
Process route (pid: 12912, process nr: 25, stackpage=c232d000)
Stack: 00000049 cf3ed700 00000049 00ffd700 ffffffff c015a92c cf7206c0
00000002
c015a957 00000002 ced041e0 00000000 00000001 cf7206c0 c232ded4
c02156a0
00000000 ced041e0 00000002 c0166f10 cf7206c0 c0166abc c232de28
00000035
Call Trace: [<c015a92c>] [<c015a957>] [<c0166f10>] [<c0166abc>]
[<c016b78c>] [<c016b80f>] [<c0148724>]
[<c016b78c>] [<c014921f>] [<c014902b>] [<c014903d>] [<c0130be8>]
[<c0130d7d>] [<c01484d0>] [<c0149262>]
[<c01499d5>] [<c0108fa9>] [<c0108ea4>]
Code: 8a 06 89 c2 83 e2 0f 8d 3c 96 24 f0 3c 40 0f 85 ce 03 00 00
>>EIP; c0151110 <ip_nat_mod_fnct+44/430> <=====
Trace; c015a92c <ip_build_xmit+e4/318>
Trace; c015a957 <ip_build_xmit+10f/318>
Trace; c0166f10 <udp_sendmsg+34c/38c>
Trace; c0166abc <udp_getfrag+0/c4>
Trace; c016b78c <inet_sendmsg+0/90>
Trace; c016b80f <inet_sendmsg+83/90>
Trace; c0148724 <sock_sendmsg+88/ac>
Trace; c016b78c <inet_sendmsg+0/90>
Trace; c014921f <sys_sendto+c7/ec>
Trace; c014902b <sys_connect+5b/80>
Trace; c014903d <sys_connect+6d/80>
Trace; c0130be8 <d_alloc+18/150>
Trace; c0130d7d <d_alloc_root+31/3c>
Trace; c01484d0 <get_fd+38/98>
Trace; c0149262 <sys_send+1e/24>
Trace; c01499d5 <sys_socketcall+10d/1e0>
Trace; c0108fa9 <error_code+2d/34>
Trace; c0108ea4 <system_call+34/38>
Code; c0151110 <ip_nat_mod_fnct+44/430>
00000000 <_EIP>:
Code; c0151110 <ip_nat_mod_fnct+44/430> <=====
0: 8a 06 mov (%esi),%al <=====
Code; c0151112 <ip_nat_mod_fnct+46/430>
2: 89 c2 mov %eax,%edx
Code; c0151114 <ip_nat_mod_fnct+48/430>
4: 83 e2 0f and $0xf,%edx
Code; c0151117 <ip_nat_mod_fnct+4b/430>
7: 8d 3c 96 lea (%esi,%edx,4),%edi
Code; c015111a <ip_nat_mod_fnct+4e/430>
a: 24 f0 and $0xf0,%al
Code; c015111c <ip_nat_mod_fnct+50/430>
c: 3c 40 cmp $0x40,%al
Code; c015111e <ip_nat_mod_fnct+52/430>
e: 0f 85 ce 03 00 00 jne 3e2 <_EIP+0x3e2> c01514f2
<ip_nat_mod_fnct+426/430>
Thank you in advance for any insight into this.
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
Subject: Re: HELP: 8 Bit Linux ?
From: Johan Kullstam <[EMAIL PROTECTED]>
Date: Mon, 22 May 2000 04:46:53 GMT
Dan Mathias <[EMAIL PROTECTED]> writes:
> Hello,
>
> Is there any 8 Bit Linux or Unix for the 68HC11 or 8088 cpu ?
no. neither have memory management. why do you need to these ancient
cpus? wouldn't a mips r3k be easier and better?
--
J o h a n K u l l s t a m
[[EMAIL PROTECTED]]
Don't Fear the Penguin!
------------------------------
Crossposted-To:
comp.os.linux,comp.os.linux.development,comp.os.linux.development.apps,comp.os.linux.misc,comp.os.linux.setup,comp.os.linux.advocacy
Subject: Re: Need ideas for university funded project for linux
From: David Steuber <[EMAIL PROTECTED]>
Date: Mon, 22 May 2000 05:00:02 GMT
[EMAIL PROTECTED] (David T. Blake) writes:
' David Steuber <[EMAIL PROTECTED]> wrote:
'
' > What TrollTech is currently doing with Qt 2.x and higher is a
' > good thing. People who produce GPL software can use Qt without
' > worrying about the QPL.
'
' That is not even close to true. Trolltech has rights to a copy
' of everything that even links with QT. They could EASILY take
' your QT linked code, and fold it into proprietary software.
http://www.trolltech.com/products/download/freelicense/
Show me where in this license TrollTech can take GPL code that links
to Qt and make it proprietary.
' From the QT Free license.
'
' If your program links with QT or is a modification of QT, you
' must supply a copy of your program (including source) to
' Trolltech.
This wording does not exist _anywhere_ in the QPL.
' Think about that for a while. They are granted a copy, with
' full rights to the copy. They are not bound in this copy by
' any license you use. Fair use would allow them to use large
' chunks of it in proprietary closed software.
Fair use is not well defined. What sort of code can they take from
you that wouldn't normaly be considered fair use?
' A license is not free if your modifications of the copyright are
' not as free as the original.
Modifications of the copyright? I don't understand. How, for
example, would GPL'd code be made less free? GPL is a very
restrictive license. Once code is covered by GPL, it can not be
uncovered.
Are you telling me that TrollTech owns copyright of your software
simply because you #include some qt files? I think you are wrong.
--
David Steuber | Hi! My name is David Steuber, and I am
NRA Member | a hoploholic.
All bits are significant. Some bits are more significant than others.
-- Charles Babbage Orwell
------------------------------
Crossposted-To: comp.os.linux.development.apps
Subject: Re: linux tools advice
From: David Steuber <[EMAIL PROTECTED]>
Date: Mon, 22 May 2000 05:00:03 GMT
[EMAIL PROTECTED] (Kaz Kylheku) writes:
' Linux runs on PC's, therefore it is a ``PC environment''.
Ah, but linux is most decidedly un-PC. PC is Microsoft. PC is
bending to the will of authority. Linux users are most un-PC in this
respect. We love freedom. We love choice. We do what we please and
damn the social morays imposed by Microsoft and the idiots who worship
the Bill.
Why else are we ostracized by the invention of the winmodem? Why else
would Microsoft try to shut us out by decomoditizing standards through
organizations such as the ITU?
Liberty above all! Beer above liberty!
Yes, I am drunk. Want to make something of it?
--
David Steuber | Hi! My name is David Steuber, and I am
NRA Member | a hoploholic.
All bits are significant. Some bits are more significant than others.
-- Charles Babbage Orwell
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To:
comp.os.linux.development,comp.os.linux.development.apps,comp.os.linux.misc
Subject: Re: Need ideas for university funded project for linux
Reply-To: [EMAIL PROTECTED]
Date: Mon, 22 May 2000 05:44:31 GMT
Centuries ago, Nostradamus foresaw a time when [EMAIL PROTECTED]
would say:
>Mongoose <[EMAIL PROTECTED]> writes:
>> On Wed, 17 May 2000 15:01:04 GMT, [EMAIL PROTECTED] wrote:
>> >Mongoose <[EMAIL PROTECTED]> writes:
>
>> >The way I see it, Linux needs the following, at minimum, before it can
>> >be a legitimate competitor to Windows:
>
>> >1. A streamlined, easy install process;
>
>> Theres distros that have that now. Caldera I think? You can play
>> tetris while linux is installing on your machine.
>
>Dunno. I've just heard very bad things about some of the installers,
>namely that they either work perfectly or not at all. And, of course,
>we should be able to turn off the easy-to-use installer and get our
>hands dirty.
Actually, all you need to do to get at _that_ is to try installing
Linux on a non-IA-32 platform. It gets _real_ easy at that point to
need to get away from the "pretty, barnified installers."
[No, I'm not bitter. I'm just irritated that it was _so easy_ to bash
up my Debian/Alpha system so that it couldn't run userspace stuff so
easily. If you upgrade a PC164 machine to Potato, and _aren't_
already running a kernel more modern than the 2.0.36 that comes with
the Stable release, the clock gets _very_ confused, to the point of
not being able to get init to anything better than run level 1...]
>> >2. An office suite roughly as functional as Office, and at least as
>> > easy to use;
>
>> Staroffice which is basically a clone of MSoffice, and Corel Office
>> Suite. Both very good office suites for linux.
>
>I've used StarOffice (not Corel Office), and it's not roughly as
>functional as Office. Also, it's not GPLed. I have hopes that
>GNOME's office suite will come through (and it's very nice, though
>crash-intensive, so far).
GNOME seems to have The Spreadsheet. Gnumeric is _quite_ impressive.
Unfortunately, AbiWord is _not_ a terribly impressive word processor;
while it may compare quite favorably to MS WordPad, and be quite
suitable for letter writing, it is _not_ much of an option for more
sophisticated document preparation.
In contrast, Kspread doesn't seem terribly impressive, but the
combination of KLyx and KWord look Rather Featureful.
It would not seem to me to be a Total Disaster if they decided to let
the "less impressive components" fail, as subprojects, and depend on
the Other Folks' Good Stuff.
>> >3. A GUI package installation mechanism that's as easy to use as
>> > InstallShield (trivial if we get a file manager for GNOME or KDE); and
>
>> Maybe, theres a few out there but no one uses them except commercial
>> companies. Most programs use the standard configure; make; make
>> install line
>
>Yeah. And that's a serious problem. Do you realize how fucking
>annoying it is to have to install 150MB of source, dedicate 1.5 hours
>to configuring and building, and then find out that there's some God
>forsaken shared library I need to install before it will work? Not
>that I have gone through this several times with XEmacs on RedHat
>boxes, or anything.
>
>I want to click on a damn button and have the program install. I want
>the option to do it by hand if I have to, but installing anything on
>Linux is a nightmare if you have to build it from the source. Note
>also that "make install" will occasionally break, depending on your
>distribution. And they all seem to be going in tangential directions
>on this one.
>
>There's just no excuse for not having an adequate installer. We have
>two excellent package-management tools, dpkg (and apt) and rpm. All
>we have to do is put a shiny new GUI front-end on them.
>
>Not that I am bitter.
I'm far more concerned with there being a solid basis underneath than
there being a pretty veneer on top.
Unfortunately, RPM seems a bit weak in terms of supporting
construction of well-managed _sets_ of packages, in comparison with
the set of dpkg tools.
Frankly, part of what I'd like to see happen is for some of this stuff
to get scripted in automated fashion.
I've got an hourly process that runs:
apt-get update
apt-get -q -y -d upgrade
which checks to see if there is anything out there needing to be
upgraded.
That's not going to automagically _do_ the upgrades; it merely
downloads the updated packages. I get to run dselect [which needs a
prettier face in these modern times...] to actually _install_ them.
In similar manner, I've set up a fair bit of my systems' configuration
to be self-updating via cfengine.
The point I'd make here is that I'm not entirely enthralled about
"shiny GUI front ends."
I think it might be _friendlier_ to set up some automated processes
which are "open" enough to provide hooks for manual maintenance, when
needed, but which Do Appropriate Things without any need for possibly
naive users to _do anything at all._
In effect, the friendliest user interface for system administration is
the user interface that you don't _need_ to use because the computer
did the Right Thing on your behalf.
>[...]
>
>> It seems that there are alot of linux programs out there that do
>> these things people need, its just that its hard to find them all.
>
>Yeah. Because there are 573,283 Linux packages, 572,911 of which do
>exactly the same thing, and 290 of which are cutesy man pages.
>
>We need some Machiavellian masochist to sift through all the packages
>for Linux, pick the best ones, and throw out the rest. I think
>anything that hasn't been changed for 5 years should go; we might have
>to make an exception for e2fsck, but as a rule, it would probably
>eliminate half the packages.
You overspeak, _slightly._
<http://rpmfind.net/linux/RPM/> reports:
"This archive hosts 64126 RPMs representing 99304 MBytes of data"
Much of these represent duplicates, whether via being different
versions, or versions for different RPM-based istributions.
In contrast, Debian has somewhere around 5000 packages these days.
With the _significant_ upside that they are each, at least initially,
required to have an identifiable maintainer.
The "maintainer list" occasionally gets a bit stale; there are some
packages that are unmaintained. If this is important to someone for
some particular package, that's an opportunity to identify a new
maintainer.
>> As for ease of use, most linux users are intellegent computer users
>> and don't need guis to configure and install stuff.
>
>This is a lie.
>
>I'm an intelligent computer user. I have manually, painstakingly
>configured my Debian system by hand, because there are either no
>usable GUIs to do it or they don't get put on my X menu (so I install
>them and forget them). I hate doing it, and I have to refer back to
>the man pages - which, by the way, are indecipherable even if you know
>what you're doing - roughly every two seconds. And then we have the
>fact that UNIX folks just love to abbreviate, and apparently consider
>it a matter of personal style and creativity. I use "fn" for
>"function," God help me, but I'm not so big a moron that I'd do it in
>a configuration file. Most of the otherwise-intelligent people who
>write the programs we use every day are guilty of that and worse
>crimes, though.
>
>I much, much, much prefer being able to right-click on something and
>hit "Properties." I also like being able to press F1 when the mouse
>is over a confusing field and get an explanation of it. (The
>explanation often isn't a help, and I expect that would carry over to
>Linux, but at least there's no flipping around between screens.)
>
>Please excuse the rant. But Linux has been a pain in the ass to
>configure since I started using it in the early 90's, and it's
>improved not at all since then.
One problem, it seems to me, is that there has only been limited
effort put into building common tools to help configure stuff.
Yes, there's Linuxconf, and COAS, and such. It is interesting that
COAS is getting rearchitected, because it proved prohibitively
difficult for people to add modules to it.
"As it turns out, most people found developing for COAS too
complicated. I will admit freely that it is very complicated; in our
next version, we're trying to cut down on some of the more esoteric
things." <http://www.coas.org/whatnext.html>
I would speculate that the same is true for Linuxconf; the site
<http://www.solucorp.qc.ca/linuxconf/modules.hc> only lists a limited
number of third party modules.
Contrast with WebMin, which combines a large number of standard
<http://www.webmin.com/webmin/standard.html> and third party
<http://www.webmin.com/webmin/third.html> modules. Admittedly, the
wish list <http://www.coastnet.com/~ken/webmin/wish.html> is rather
large...
I _seriously wish_ that these tools took the approach of not merely
modifying files, but rather generating some form of "closures."
Common Lisp defines closures thus:
"lexical closure n. a function that, when invoked on arguments,
executes the body of a lambda expression in the lexical environment
that was captured at the time of the creation of the lexical closure,
augmented by bindings of the function's parameters to the
corresponding arguments."
The idea is that rather than modifying the file, the tools should
generate _little programs_ (e.g. - closures) to modify the files.
That way, the tool may store the _program_, and provide the ability to
rerun as needed. (Theoretical CS Geek Aside: When you compile a C
program using GCC, what GCC does is to transform C code into a whole
bunch of machine language closures.)
<wild-eyed-dreaming>
For instance, if WebMin generated Cfengine
<http://www.iu.hioslo.no/cfengine/> scripts, this would mean you'd run
WebMin, and it would _generate scripts._
It should log these in a spool directory, so that:
- If you wanted, you could grab some of them and add them to the
Cfengine "cleanup" that runs regularly.
More mundane, and more critical, uses include:
- You get to look back at the spool to see what was done in the past.
Whether this is called "tracing back," "tracking," or "auditing,"
it's useful from a _control_ perspective.
- You may not understand how to configure a particular facility. So
you use the tool to help walk you through. Then you wind up with a
script that you can, if you wish, look at, to see what it did and
how. Good for learning.
The latter things were provided on AIX by the SMIT utility, some
years back; it would log whatever you did with the tool in a format
that would actually let you run it through SMIT to redo it...
</wild-eyed-dreaming>
>> This is the problem though, they don't care enough to create
>> programs to help newbies install and use linux and so linux is being
>> held back.
>
>I care enough. I'm just no good at GUI programming.
>
>> I don't see linux taking off any time soon either but the more help
>> it gets, the more popular it will be.
>
>I don't think we should squander this opportunity. The reason I get
>so locquatious when it comes to Linux is that I really like some parts
>of it, and really hate others. Same thing for Windows, but the really
>funny part is that the two are, for me, almost perfectly complementary.
>I see an opportunity for us to improve Linux so that it can be like it
>is now, or like Windows, or like anything at all, and change between
>the two with only about five minutes' effort.
>
>It's just that there's such a huge opportunity here, and it seems like
>so few people are willing to take advantage of it. Myself included,
>but I, unfortunately, don't have a separate computer available to do
>Linux development. (Is there a Linux for SGI boxes yet?)
I find it irritating that these [UNIX and MacOS] are the only systems
that people appear capable of being aware exist; there are doubtless
useful things that can be learned from:
a) VMS
b) MVS
c) Perhaps MacOS?
d) Lisp Machines
e) There are still things from Multics not emulated on more recent
systems.
--
Why do we drive on parkways and park on driveways?
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/oses.html>
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To:
comp.os.linux,comp.os.linux.development,comp.os.linux.development.apps,comp.os.linux.misc
Subject: Re: Need ideas for university funded project for linux
Reply-To: [EMAIL PROTECTED]
Date: Mon, 22 May 2000 05:44:32 GMT
Centuries ago, Nostradamus foresaw a time when Mongoose would say:
>On Wed, 17 May 2000 21:10:43 GMT, [EMAIL PROTECTED] wrote:
>>We need some Machiavellian masochist to sift through all the packages
>>for Linux, pick the best ones, and throw out the rest. I think
>>anything that hasn't been changed for 5 years should go; we might have
>>to make an exception for e2fsck, but as a rule, it would probably
>>eliminate half the packages.
>
> Ya this is true, freshmeat needs some better organization, or a
>voting system. Since everyone can make applications for free, that
>leads to so much crap being made.
When the database is not actually linked to installable, working code,
that's not an incredibly useful "organization."
The best vote is a functioning software package. Whether that be:
a) A source RPM, or
b) A source .deb, or
c) A Ports package.
--
"Windows NT was designed to be administered by an idiot and usually
is..." -- Chris Adams" <[EMAIL PROTECTED]>
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/linuxsysconfig.html>
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.development.system) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Development-System Digest
******************************
