Linux-Misc Digest #839, Volume #20 Tue, 29 Jun 99 00:13:08 EDT
Contents:
Re: exiting Gnome caused core dump (Michel Catudal)
Re: Linux loses in NT tests (Alex Lam)
Re: Automated benchmarking? (was: Re: Opportunity for FreeBSD) (Seth Van Oort)
Re: nohup not working (Adrian Hands)
Re: Documentation issues. (Leslie Mikesell)
Re: changing default os at boot (Adrian Hands)
Re: Wordperfect and color printing (John Hong)
Routing two Internet Networks ("T.J. Weber")
Re: Linux loses in NT tests (Alex Lam)
Re: Mac-emulation on Linux? ("Anthony D. Saxton")
Machine dies mysteriously--Help please! (Zhengdong Zhang)
Re: Documentation issues. (vepxistqaosani)
----------------------------------------------------------------------------
From: Michel Catudal <[EMAIL PROTECTED]>
Subject: Re: exiting Gnome caused core dump
Date: 28 Jun 1999 22:00:47 -0500
bruce wrote:
>
> RH6.0
>
> Every time I exit Gnome using Ctl-Alt-Backspace, my system core dumps.
> No big deal, but is there a better way to quit X?
Yes do as you would do in winblows
--
use OS/2 for a crash proof work environment
use Linux for safe and quick internet access
use Winblows to test the latest viruses
http://www.netonecom.net/~bbcat/
We have software, food, music, news, search,
history, electronics and genealogy pages.
------------------------------
From: Alex Lam <[EMAIL PROTECTED]>
Subject: Re: Linux loses in NT tests
Date: Mon, 28 Jun 1999 19:32:31 -0700
Silviu Minut wrote:
>
> Alex Lam wrote:
>
> > Silviu Minut wrote:
> > >
> > > > Redhat SUX big time anyway....
> > > >
> > > >
> > >
> > > Not trying to defend RedHat (although I'm running it), just trying to
> > > understand, but why does it suck that bad?
> >
> > I have nothing but problems with RH since 4.0, first, it killed one of
> > my monitor during auto probing.
>
> I still fail to understand what RH has to do with that. Ain't X -probeonly
> an Xserver thing? Was your monitor supported?
>
Yes, it's a generic svga with 75hz sync.
> > 2nd. It never allows me to install
> > completely in all the boxes that I've tried. (5, from old P-133 to brand
> > new dual cpu full U2WSCSI.)
> >
>
> Strange. I never had any problems installing RH5.0, RH5.2 and RH6.0. It's
> true, I did skip the autoprobe of the monitor cause I knew it could lock up
> the computer, and I installed and configured the Xserver afterwards.
>
> >
> > But I was able to install SuSE, Slackware, FreeBSD on all the same boxes
> > that chocked RH.
> >
> > The RH installer never let me pass the probing stage, both in text and
> > GUI modes.
> >
>
> So if you never got to install RH, how can you say it's all junk?
> Maybe it is, but from what you're saying I don't see enough justification.
>
RH never did allow me to finish the installation. don't tell me I
didn't give RH a chance. I tried Slackware first - my first ever
Linux encounter, a few years ago, pre-ELF days - . Yes it took me 30
plus hours, but it did installed properly.
Then I tried Redhat later, hoping that it'll be easier to install...
No, RH never even allowed me to finish a single installation in 5
different boxes.
It keep crashing and freezing, with both GUI and text modes... At one
time, I did managed to get to the point to reboot from Lilo... During
boot up, got a kernel panic
and the whole thing crashed and hosed itself out. I thought it's M$
BLUES all over again.
Then, I tried other distro. All were able to install smoothly with a
SINGLE pass. With the same hardware.
Maybe RH just don't like me.
So, I do have plenty of justification to call RH junk. I know RH is
very popular. If I want to run something popular, I'd stay with M$.
Alex Lam.
> >
> > Alex Lam.
> > --
--
*remove all the Xs (upper case X) if reply by e mail.
** no more M$ Windoze.
------------------------------
From: Seth Van Oort <[EMAIL PROTECTED]>
Crossposted-To: comp.unix.bsd.freebsd.misc,comp.os.linux.advocacy
Subject: Re: Automated benchmarking? (was: Re: Opportunity for FreeBSD)
Date: Mon, 28 Jun 1999 22:19:59 -0500
[EMAIL PROTECTED] wrote:
>
> Donn Miller <[EMAIL PROTECTED]> wrote:
> >snip<
> : But yes, I suggest we do some sort of benchmark of FreeBSD against other
> : OSes to see where we stack up. Ignorance is a dangerous thing.
>
> Agreed.
>
> Along these lines I've been thinking that an ongoing, automated
> benchmark test might be interesting. Here's how I think it might be
> able to work:
>
> Take test box foo. Install the latest of everything on it (*BSD
> (stable and current), Linux (stable and beta), Solaris x86, NT,
> Win2k). Create a custom bootloader that cycles through OSes on each
> reboot. Create an automatic update system invoked on boot (or
> shutdown?) to upgrade the system and its software (Samba, Apache,
> benchmark software, et al). -*BSD (and Linux?) would simply be cvsup
> && make world && build kernel.
>
> Create benchmark load test boxes bar, baz, and boo. Run NT on one,
> 98 on another, FreeBSD on another, etc.
>
> Run automated benchmarks say, once a week and post the results to a
> web page with a graph showing how the systems have progressed (or
> regressed) over time and with respect to each other. Only benchmark
> systems which allow freely distributed source would be allowed, for
> obvious reasons.
>
> Durring the week the box would run in "inspection mode", such that
> the entire box would be available to inspect via anon FTP (read only
> of course:-). -Do to MS/Solaris licenses, some parts might only be
> inspectable by MD5 checksum I guess, but maybe full access could be
> granted to those that could somehow prove they have a license for NT
> and/or Solaris, such as though from MS/Sun?
>
> The general idea would be to try and create a single, stable,
> undisputable benchmarking server. If anyone has a problem with how
> XYZ isn't tuned correctly, it can be fixed. If anyone has a problem
> with how benchmark software XYZ is biased to YZQ OS or doesn't
> measure "real world" loads well, they can submit another (open
> source) benchmark to be run in the normal que.
>
> I have no idea where such a server would be best hosted. Certainly
> not PC Labs and just as likely not UC Berkeley either for the same
> reasons. Maybe some small tropical island like Jamaica were
> I...err...I mean whoever is the sysadmin can maintain it properly
> and in an unbiased fashion. :-)
>
> In short, no one gets to cheat, no one gets to lie, and the "winner"
> (for the week at least:-) gets honest bragging rights for the Indy
> 500 of benchmarks. :-)
>
> Thoughts?
It would be one of the most interesting things I've seen in a while. Who
wants to do it? :-)
Seth
>
> --
> -Zenin ([EMAIL PROTECTED]) Caffeine...for the mind.
> Pizza......for the body.
> Sushi......for the soul.
> -- User Friendly
------------------------------
From: Adrian Hands <[EMAIL PROTECTED]>
Subject: Re: nohup not working
Date: Mon, 28 Jun 1999 20:45:10 -0400
Reply-To: [EMAIL PROTECTED]
I just did EXACTLY as you described and, when I logged out, silly
continues to live.
Please note that nohup does NOT prevent silly from being a child process
of your shell.
When you logout, silly is sent signal 1 (HUP).
Since you have not coded a handler for the signal, it takes the default
action (program exits).
nohup prevents silly from dying by trapping the HUP signal so that silly
never gets it,
but it does not change the "parent process id" of the process.
You can simulate logging out by sending the HUP signal to the process:
$ kill -HUP 23154
or, on Linux:
$ killall -HUP silly
If you really want silly to NOT be a child of the shell, try this:
$ (./silly &) &
[1] 23153
$
[1]+ Done ( ./silly & )
$ ps -f
UID PID PPID C STIME TTY TIME CMD
adrian 13874 9237 0 Jun26 ttyp5 00:00:01 /bin/bash -i
adrian 23154 1 51 20:37 ttyp5 00:00:04 ./silly
adrian 23156 13874 0 20:37 ttyp5 00:00:00 ps -f
$
This works because it starts a subshell which then runs "silly &" and
exits.
The "Done" message is the result of this intermediate subshell exiting.
silly lives on, with "init" (pid 1) as it's adoptive parent.
Note that silly is still associated with my terminal (ttyp5).
If you're writing a daemon take a look at the "Unix FAQ".
It has several excellent suggestions for things all daemons should do,
including how to get rid of the association to the terminal.
If you're writting
Siemel Naran wrote:
>
> I find that nohup is not working.
>
> I made the following program
> // silly.cc
> int main() { while (true) ; }
> And compiled it with
> g++ --pedantic silly.cc -o silly
>
> Then I run the program with this
> nohup silly
> /usr/bin/nohup silly &
> In tcsh, nohup is a builtin shell command. So the first version uses
> the builtin nohup whereas the second version uses the nohup in
> /usr/bin/. In bash, nohup is not a builtin. So both versions use the
> nohup in /usr/bin/.
>
> Now here is the relevant result of "ps f".
> 21917 qd S 0:00 -sh
> 21933 qd R N 0:03 \_ silly
> We see that process/program "silly" is running under sh, just as if we
> had typed "silly &". In other words, process "silly" is a child process
> of "sh". But this is precisely what I don't want because when I log out
> of the window with "exit", process 21917 and its child process 21933 are
> killed. So "silly" is killed.
>
> My questions.
> 1. How can I keep a program running after I log out (I thought nohup
> does the job)?
> 2. Is there something wrong with GNU nohup (both the versions in tcsh
> and /usr/bin/tcsh)?
>
> Thanks.
>
> --
> ----------------------------------
> Siemel B. Naran ([EMAIL PROTECTED])
> ----------------------------------
------------------------------
From: [EMAIL PROTECTED] (Leslie Mikesell)
Crossposted-To: gnu.misc.discuss,comp.unix.bsd.misc
Subject: Re: Documentation issues.
Date: 27 Jun 1999 18:04:40 -0500
In article <7l68h4$[EMAIL PROTECTED]>,
Peter da Silva <[EMAIL PROTECTED]> wrote:
>>Is there some reason SDF (Simple Document Format) hasn't caught
>>on? http://www.mincom.com/mtr/sdfsite-2.001/
>>It claims to be author-friendly and has translators to several
>>output forms.
>
>It's a step backwards. Now documentation formats are aligning on SGML DTDs,
>why go back to "dot commands".
So a person can type it without needing a DTD-aware editor?
>Also, it seems to be rather procedural.
Some people think that way. Let the output formatters take care
of the differences. Does anyone ever think of 'start-of-object'
when you type a paragraph, or do you just note the breaks between
them?
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: Adrian Hands <[EMAIL PROTECTED]>
Subject: Re: changing default os at boot
Date: Mon, 28 Jun 1999 21:03:29 -0400
Reply-To: [EMAIL PROTECTED]
Well, /sbin/lilo can definately be used to change the default to nt
using either the "-R" or "-C" switches (see "man lilo").
You can make linux your default and use "lilo -R nt & shutdown -r now"
to boot nt in the morning, but I suspect you want your users to be able
to boot durring the day and get nt by default, so...
You can use LOADLIN.EXE to load linux from windows. I haven't tried it
with nt, but it works great with win95. As much as I hate to suggest
making m$ your default, it sounds like you should make nt your default
and use LOADLIN to load linux at night.
-Adrian
"J. Grant Kessler" wrote:
>
> I'm looking for a way to have a computer default to booting windows nt
> during the day and linux at night. During the day, the computers must
> boot windows nt, but at night, I would like to set an at command to reboot
> the computer out of windows nt and automatically boot into linux. By
> morning, I'd like a cronjob to reboot linux and boot into windows nt. I'm
> trying to automate this process so that the computers can be utilized for
> parallel processing in linux(not nt) at night, and go back to their normal
> duties as nt workstations during the day.
>
> I was wondering if there are any boot loaders which will change the
> default operating system at boot based upon time. I have not found any.
>
> Is there a way in windows nt to modify the mbr using an at command?
> Dskprobe.exe appears to do this, but also appears to need human
> intervention and cannot be run by an at command automatically.
>
> Any suggestions or help will be greatly appreciated.
>
> Thanks
> Grant
------------------------------
From: [EMAIL PROTECTED] (John Hong)
Subject: Re: Wordperfect and color printing
Date: 29 Jun 1999 03:34:52 GMT
Vinh Le ([EMAIL PROTECTED]) wrote:
: Just a quick question: Does Wordperfect 8 support color
: printing in the retail version? The free downloadable
: version has the option ghosted. Obtaining the key
: doesn't add the feature either.
Shouldn't this be dependent on your printer or the selected
printer driver?
------------------------------
From: "T.J. Weber" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,linux.net
Subject: Routing two Internet Networks
Date: Mon, 28 Jun 1999 22:44:57 -0600
Reply-To: [EMAIL PROTECTED]
Hi, I have a question, and it might be fairly complex.
We currently have two Internet connections coming into our building. My
company hosts several web sites and does some Internet work. I'm trying
to get re-familiar with the route daemon, and the basics on configuring
routes.
We have two Internet connections, both T1, and for simplicity I'll use
local IP addresses.
ISP #1 -- 192.168.1.0/25 (default route)
ISP #2 -- 10.201.30.0/24 (new route, not configured)
So we have the two Internet connections, and the link from ISP #1 has
been working for over a year, just fine without problems. Now we have
our new, ISP, ISP #2 .... which I'm trying to routes set for all IP
addresses on 10.201.30.0/24 to go to the router, set to 10.201.30.1.
In a nutshell, I want all traffic with IP addresses on the
10.201.30.0/24 block to route through the router on 10.201.30.1, and all
the IP addresses on 192.168.1.0/25 to go through on 192.168.1.1. Here's
the diagram:
current default route: 192.168.1.1
10.201.30.0/24 --> 10.201.30.1 --> Internet
192.168.1.0/25 --> 192.168.1.1 --> Internet
What my current problem is that the IP addresses set on the
10.201.30.0/24 block get routed through 192.168.1.1, which is the
current default route. I've tried adding the network to Linux and
configuring the default route for that network, but it still doesn't
work. When I set a website to an IP on 10.201.30.0/24, all the traffic
goes out on 192.168.1.0/25.
Please advise if anyone has any thoughts on which commands I would issue
to route the traffic to the place I want it to go to.
Thanks in advance,
--t.j. weber
--
T.J. Weber | Providing your business with COMPLETE
Interplanetary Media | computer & Internet solutions!
phone: 847.205.5200 | ----- SARRZY INTERNET SOLUTIONS -----
fax: 847.205.5201 | web: http://www.ipmedia.net
e-mail: [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
====================================================================
He's not dead, he's / You have the right to remain
electroencephalographically / silent. Anything you say will
challenged. / be misquoted and used against you.
====================================================================
------------------------------
From: Alex Lam <[EMAIL PROTECTED]>
Subject: Re: Linux loses in NT tests
Date: Mon, 28 Jun 1999 19:38:44 -0700
Johan Kullstam wrote:
>
> Silviu Minut <[EMAIL PROTECTED]> writes:
>
> > So if you never got to install RH, how can you say it's all junk?
> > Maybe it is, but from what you're saying I don't see enough
> > justification.
>
> on the other hand, the *primary purpose* of a distribution is to get
> (at least a critical mass of) a working linux on your system.
> everything else is just icing on the cake. hence, if the install
> doesn't work, then the distribution fails.
>
> just so you know i have no axe to grind against redhat, i have been
> using redhat since 3.0.3 and am running 6.0 now. i have personally
> had good luck with it and my two machines. ymmv.
>
Please read my other reply here. Yes, I do have plenty of axe to grind
against RH.
I tried it on 5 machines, RH chocked on all 5 , but the other distros
were able to installed properly and smoothly on the same 5 machines.
I tried RH 4.0, 4.2, 5.2.
Non were able to even finish the initial installation.
Would you call that junk if that happens to you? I don't care if RH
is the most popular distro, but it certainly is making the most noises
lately, especially since its IPO.
"Empty vessel makes the most noise" surely applies here.
Alex Lam.
> --
> J o h a n K u l l s t a m
> [[EMAIL PROTECTED]]
> Don't Fear the Penguin!
--
*remove all the Xs (upper case X) if reply by e mail.
** no more M$ Windoze.
------------------------------
From: "Anthony D. Saxton" <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.advocacy,comp.os.linux.development.system,comp.os.linux.powerpc
Subject: Re: Mac-emulation on Linux?
Date: Tue, 29 Jun 1999 03:44:48 GMT
Matt Denton wrote:
> In article <7gjd3s$cdo$[EMAIL PROTECTED]>, "FM"
> <[EMAIL PROTECTED]> wrote:
>
> > 2. Buy a Macintosh and dual boot with Linux/MacOS
>
> This is probably the best option - LinuxPPC is pretty solid, MKlinux
> (easier to install?) is a bit sluggish as it runs on top of a Mach
> microkernel instead of natively but is a decent port of Linux nonetheless.
>
> Me, I'm looking forward to Sheepshaver (http://www.sheepshaver.com) MacOS
> VM for Linux to see how that works...
>
> --
> Matt Denton
> San Francisco, USA
SheepShaver's fine for those of us privilidged few using the PowerPC version
of linux. It won't help anyone that doesn't have a PREP, BeBox or PowerMac
though.
Tony
------------------------------
From: Zhengdong Zhang <[EMAIL PROTECTED]>
Subject: Machine dies mysteriously--Help please!
Date: Mon, 28 Jun 1999 21:53:25 -0500
I am using Red Hat Linux 5.0 on an old 486 machine. It used to work pretty
well. Today when I tried to log in (I have not used it for a while), I
found the machine was not responsive. So I reset the machine. It worked
for several minutes, then went dead. I reset it again, the same followed.
I don't have a clue what could go wrong. Please help. Any suggestion will
be appreciated.
Zhengdong Zhang
[EMAIL PROTECTED]
------------------------------
From: vepxistqaosani <[EMAIL PROTECTED]>
Crossposted-To: gnu.misc.discuss,comp.unix.bsd.misc,comp.text.sgml,comp.text.xml
Subject: Re: Documentation issues.
Date: Mon, 28 Jun 1999 22:08:45 -0400
Christopher Browne wrote:
<snip!>
> The problem that I see with SGML (that is equally true for XML) is that
> it introduces two additional languages (e.g. - the DTD and the
> application-specific markup language) without diminishing the need to
> understand at least two others, namely:
>
> a) Whatever is being used to interpret the parsed document, and render
> it into...
> b) The output format.
>
> With Jade, this involves:
> a) The DTD,
> b) The language described by the DTD,
> c) DSSSL,
> d) {HTML|TeX|nroff|Postscript|...}
<snip!>
Perhaps my lack of sgml sophistication is showing, but what is the difference between
your latter a) and b)?
As for the rest of your point: I am currently working on a project that eliminates
two of your steps. The idea is to write a TeX (style or class) file that can process
native sgml. TeX works on *.sgml files somewhat less efficiently than on *.tex files
-- but it can be done. Since I am quite fluent in TeX, this is the ideal solution for
me. Note that I do not interpret the DTD; I don't even strictly require it (though,
in practice, it can be quite helpful). All I need is the assurance that there is a
DTD and that the *.sgml files parse correctly.
The problem with this approach is that each DTD and set of sgml files conformant to
that DTD are sui generis and new TeX files would have to be developed for each new
project -- but we already do that, though in a somewhat simpler way.
While I appreciate the problems with sgml/xml (bloated syntax, poor and/or expensive
tools, enormous consultancy costs), the goal is a worthwhile one -- and I really
don't see any other way to get to it.
Already I am routinely faced with the problem of going back in time to deal with
ancient (1995 and earlier) TeX and LaTeX files. A format so general that it is
unlikely to require that kind of effort is a goal well worth pursuing.
Fred
------------------------------
** 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.misc) 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-Misc Digest
******************************
