Linux-Development-Sys Digest #371, Volume #6 Wed, 3 Feb 99 13:14:21 EST
Contents:
Re: Rewriting IDLE Process - Need Strategic Advice - part 1 (Kalle Olavi Niemitalo)
Re: New free widget library: Notif-0.1 (Adam P. Jenkins)
v2.2.1 crond filling up disk (M Sweger)
Re: Linux Phase 2: A Consumer Operating System (steve mcadams)
Re: Internal PCI modem (Donato Marrazzo)
Re: New free widget library: Notif-0.1 (John Rowe)
Re: VESA VGA graphics Console at 1280X1024 (Donal O'Sullivan)
VESA VGA graphics Console at 1280X1024 (Donal O'Sullivan)
Re: svgatextmode + riva128 = ? (Walter van der Schee)
kernel: N_TXTOFF < BLOCK_SIZE. Please convert binary. (Michael Spalinski)
IrDA for Linux ? (Claudio Leonel Salvadori)
Re: New free widget library: Notif-0.1 (jedi)
Re: Kernel is too big (Horst von Brand)
Re: Parallel C for Linux (Frankie East)
Re: Linux on ARM7 ?? (Mitch Davis)
Re: Easy(?) kernel question. (cano_jonathan)
Re: glibc-help (Andreas Jaeger)
Re: GNUPro for Linux - Recommendations (Frankie East)
2.2.1 ("Me")
Re: Linux Phase 2: A Consumer Operating System (John Hasler)
Unix/Advanced Computing People (Lakshmi Natarajan)
Re: 2.2.1 (Jonathan Stott)
Re: glibc-crypt-2.0.112 somewhere ? (Ronald Cole)
----------------------------------------------------------------------------
From: Kalle Olavi Niemitalo <[EMAIL PROTECTED]>
Subject: Re: Rewriting IDLE Process - Need Strategic Advice - part 1
Date: 03 Feb 1999 10:23:40 +0200
I believe the scheduler must always have one task running -- the idle
task if nothing else. If your defragmenter is going to read the hard
disk, it will have to wait for the read operation to finish before it
can use the data. This would leave the scheduler with no runnable
tasks. So you should add a new low-priority task instead of using the
existing idle task. That way the HALT loop could still work too.
Disclaimer: I haven't read the Linux scheduler.
------------------------------
Crossposted-To:
comp.windows.x,comp.os.linux.advocacy,comp.os.linux.development.apps,comp.os.linux.x
Subject: Re: New free widget library: Notif-0.1
From: [EMAIL PROTECTED] (Adam P. Jenkins)
Date: 02 Feb 1999 19:19:10 -0500
[EMAIL PROTECTED] (Joseph H Allen) writes:
> In article <[EMAIL PROTECTED]>,
> jik- <[EMAIL PROTECTED]> wrote:
> >What drag and drop protocol did you use, Xdnd?
>
> My own- it will only work within the client (I.E., from one window to
> another within the same client). This certainly needs to be fixed; I was
> not aware that there was a real drag and drop protocol for X. I'll take a
> look at Xdnd. Are there other/better/more widely used protocols than this?
Xdnd is the wave of the future. Both GNOME and KDE will use it in
their desktops.
--
Adam P. Jenkins
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (M Sweger)
Subject: v2.2.1 crond filling up disk
Date: 3 Feb 1999 13:13:53 GMT
Reply-To: [EMAIL PROTECTED]
Hi,
I've got Doslinux77 which is v2.2.1 and when I login and do top I
see that crond is eating up 60% of the cpu and syslogd is using 33%.
I have a 33mhz intel pc. The crontab is commented out and even though
I do a crontab -d to kill any cron running, it doesn't make a difference.
The /var/log/cron,/var/log/debug/var/log/messages are filling up with
messages saying it has removed user, and the user isn't specified. THis
continues for 40000 times or xxxxx times over and over filling up the
disk. I know I can kill the crond daemon, but this still is a problem.
--
Mike,
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (steve mcadams)
Crossposted-To: alt.os.linux,comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: Linux Phase 2: A Consumer Operating System
Date: Tue, 02 Feb 1999 21:15:04 GMT
[Snipped for brevity, quoted material marked with ">"]
On Tue, 2 Feb 1999 13:40:42 +1100, [EMAIL PROTECTED] (Screwtape)
wrote:
>There's (at least) two possible design routes. If we just want to get Linux on
>the desktop, we'll take all the fine openness and flexibility that makes Linux
>great, and sweep it under the carpet to deliver a very Windows-like experience:
>assumptions being made about what you're trying to do, arbitrary limitations
>imposed, treating the user like they are brain-dead, and so on.
Ugh. Windows, from a usability standpoint, has only one point of
merit: it is usually possible as a total newbie to go into a program
and make it do something useful. Nevermind that beyond that, the
program may be crippled, force you to jump back and forth between the
mouse and keyboard, yada yada. Aside from this, and the fact that
it's the only GUI interface most of its users have ever seen, it has
little or no usability value.
>On the other hand, my personal attitude is this: "Linux everywhere" be damned.
>Let Linux succeed on its technical merits, not because we're pushing it so hard.
>Let's not rush Linux to the desktop - let's make a new GUI, or even a new *type*
>of GUI. Something that has the same kind of power and flexibility as the shell,
>if that's possible, something where nothing is dumbed down. I really have no
>idea what that would be, but I hope the folks at GNOME and KDE realise they
>don't just have to copy Windows, or even the Mac. Let them make something
>beautiful and new. We're not gonna press them to do it quickly, as long as they
>do it *RIGHT*.
YES! I agree, as you might guess. (I've been working on a new (type
of) GUI the last couple months or so, though I'll admit that lately
I've gotten sucked into the newsgroups and haven't devoted much coding
time to it. I also got kinda nailed by some of the linuxen who seemed
inclined to say "hey, we already got too many gui's, leave off that,
nobody cares" and some others who said "you'll never make a nickel
writing GPL code". But attempting to stay on topic, I agree there are
good ways and bad ways to go about the effort. -steve
========================================================
so what? - http://www.codetools.com/showcase
------------------------------
From: Donato Marrazzo <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Re: Internal PCI modem
Date: Fri, 29 Jan 1999 13:07:14 +0100
What about "HCF" modem as the mine?
Can i hope?
Robert Krawitz ha scritto:
> Yup, that statement makes it a Winmodem. "HSP" means "Host Signal
> Processing".
------------------------------
Crossposted-To:
comp.windows.x,comp.os.linux.advocacy,comp.os.linux.development.apps,comp.os.linux.x
From: John Rowe <[EMAIL PROTECTED]>
Subject: Re: New free widget library: Notif-0.1
Date: Wed, 3 Feb 1999 10:17:34 GMT
Emile van Bergen <[EMAIL PROTECTED]> writes:
> Joseph H Allen wrote:
>
> 1. Where to record state data? If you have two callback functions which
> need the same data to decide their actions upon, you end up with a lot
> of global variables.
I have to disagree with this. Just about every datum refers to some
object and belongs inside the structure corresponding to that
object. Provided the toolkit has a simple way of passing things to
the callback the callback only ever modifies structure members. Of
course, with some of them that can be pretty hard :-(
Assuming your application places no restrictions on the number of data
sets it can work upon and the number of windows per data set then
you're going to have to do this anyway. You don't store data about an
object inside a function.
IMHO static data is very rare and usually the result of some
unpleasant compromise.
John
------------------------------
From: Donal O'Sullivan <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: VESA VGA graphics Console at 1280X1024
Date: Wed, 03 Feb 1999 14:08:35 +0000
Apologies for the previous message, I attached it to this thread by
accident.
--
Regards,
Donal
------------------------------
From: Donal O'Sullivan <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: VESA VGA graphics Console at 1280X1024
Date: Wed, 03 Feb 1999 14:00:59 +0000
I have just set up the VESA VGA graphics Console. It works fine
except at a resolution of 1280x1024. At this resolution, my monitor
shuts down every few minutes for about two seconds. It then recovers
and everything is fine for a while. This also happen under X.
Has anyone else seen this problem?
I am currently running Kernel 2.1.1 although the problem also occured
when using kernel 2.1.122.
My motherboard is an Asus TX97-E with a Cyrix 6x86Mx 233MHz processor
My graphics card is a Raven 3D (Voodoo Banshee).
--
Thanks,
Donal
------------------------------
From: Walter van der Schee <[EMAIL PROTECTED]>
Subject: Re: svgatextmode + riva128 = ?
Date: Wed, 03 Feb 1999 15:30:55 +0100
Hello Andrew,
The way I see it, you can do this using VESA-framebuffer support built
into the 2.2.x series kernels.
It asks for VESA-2.0 or higher, and I have used it successfully on
2.1.130-tru 2.2.1 on a Diamond Viper V330 (Riva 128), which has support
for VESA 3.0.
Read the /usr/src/linux/Documentation/fb/vesafb.txt for details.
Walter
------------------------------
From: Michael Spalinski <[EMAIL PROTECTED]>
Subject: kernel: N_TXTOFF < BLOCK_SIZE. Please convert binary.
Date: 28 Jan 1999 15:29:07 -0500
Does anyone know what this means? From /var/log/messages:
Jan 28 15:12:46 riemann kernel: N_TXTOFF < BLOCK_SIZE. Please convert binary.
This is the 2.2.0 kernel, with aout support compiled as a module. The rest
of the setup is RH5.2 ... This message appears whenever I try to run an old
a.out binary.
BTW, insmod binftm_aout succeeds, lsmod shows that it is there.
------------------------------
From: Claudio Leonel Salvadori <[EMAIL PROTECTED]>
Subject: IrDA for Linux ?
Date: Wed, 03 Feb 1999 23:41:16 +0000
Do anyone know about any implementation of the
IrDA protocol stack for Linux ?
Thanks.
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (jedi)
Crossposted-To:
comp.windows.x,comp.os.linux.advocacy,comp.os.linux.development.apps,comp.os.linux.x
Subject: Re: New free widget library: Notif-0.1
Date: Tue, 2 Feb 1999 17:12:18 -0800
On 2 Feb 1999 17:55:29 -0600, Peter Samuelson <[EMAIL PROTECTED]> wrote:
>[Joseph H Allen <[EMAIL PROTECTED]>]
>> I was not aware that there was a real drag and drop protocol for X.
>> I'll take a look at Xdnd. Are there other/better/more widely used
>> protocols than this?
>
>Look at XEmacs for a sample implementation of multiple DND protocols.
>Between CDE, KDE, GNOME, OffiX and probably others, there are quite a
>few to choose from. I think KDE and GNOME might be using the same
>protocol but I don't know for sure.
What parlour tricks can you do with Xemacs and Offix dnd?
My prefered WM and file mangler use OffiX.
--
Herding Humans ~ Herding Cats
Neither will do a thing unless they really want to, or |||
is coerced to the point where it will scratch your eyes out / | \
as soon as your grip slips.
In search of sane PPP docs? Try http://penguin.lvcm.com
------------------------------
From: [EMAIL PROTECTED] (Horst von Brand)
Subject: Re: Kernel is too big
Date: 3 Feb 1999 16:06:12 GMT
In article <793ccv$dhi$[EMAIL PROTECTED]>, Dan Wilder wrote:
>Chiyu Wang <[EMAIL PROTECTED]> writes:
>[ ... ]
>
>>However, after running /sbin/lilo, a message shows on the screen
>
>>Kernel /boot/vmlinuz-2.0.36-0.7.new is too big
>>^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
In RedHat, you can do just:
make install
make modules
make modules_install
This takes care of everything.
--
Horst von Brand [EMAIL PROTECTED]
Casilla 9G, Vi�a del Mar, Chile +56 32 672616
------------------------------
From: Frankie East <[EMAIL PROTECTED]>
Subject: Re: Parallel C for Linux
Date: 03 Feb 1999 08:16:57 PST
Tom Goodale wrote:
> Vitor Pedro Bonucci Pias wrote:
> >
> > I have my system running Dual pentium II 450.
> >
> > The system runing fine with kernel-2.2.0-pre7,
> >
> > and i would like to know if there is a Parallel C
> >
> > compiler for Linux to explore the Parallelisme.
> >
> > Thanks
> >
> > Pedro Pias
>
> I don't know of any parallel c compilers, but if you can put up with
> distributed memory parallelism, you could use pvm or mpi.
>
> Tom
The best available are Protland Group at www.pgroup.com and Kai.
Portland offers C/C++/F90/and HPF along
with a good parallel debugger and an excellent profiler. Both PGroup
and Kai are OpenMP compliant meaning they follow a standardized method
of allowing you to specify which processor you want which process to run
on (otherwose you'd simply be multithreading, which of course you can do
without these compilers to get some speedup).
The advantage to PGroups is that they supply the main compilers for the
big guns - Crays, Tandems, Origins, etc.
So if you need to move code between your computer and a supercomputer,
or you have code from a supercomputers application that you want to move
to Linux that would simplify things greatly. The Terra Earth simulation
was just recelty ported from supercomputers to a Linux Beowolf cluster
this way.
...Frankie
------------------------------
From: Mitch Davis <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.arm
Subject: Re: Linux on ARM7 ??
Date: Wed, 03 Feb 1999 12:19:42 +1100
Roger Gammans wrote:
>
> BTW, This means you will need a 32bit wide memory system since the
> kernel won't be using the thumb instructions. (At least I think so, I
> haven't done any hardware myself with the 7tdmi).
ARM7 instructions are 32 bits, and Thumb instructions are 16 bits.
This is totally independent of the external data bus size, which
is left up to the manufacturer to choose.
In other words, you can run ARM7 code on a processor which only
has a 16-bit external data bus, and you can run Thumb code on a
processor which has a 32-bit data bus.
So whether you /need/ a 32-bit wide memory system is independent
of which instruction set you use.
Regards,
Mitch.
--
| mailto:[EMAIL PROTECTED] | Not the official view of: |
| mailto:[EMAIL PROTECTED] | Australian Calculator Opn |
| Certified Linux Evangelist! | Hewlett Packard Australia |
------------------------------
Crossposted-To: comp.os.linux.misc
Subject: Re: Easy(?) kernel question.
From: cano_jonathan <[EMAIL PROTECTED]>
Date: Wed, 03 Feb 1999 02:32:59 GMT
Frank Hale <[EMAIL PROTECTED]> writes:
> cano_jonathan wrote:
> >
> > I'm looking at the 2.0.35 linux kernel and I have a question
> > (primarily for the intel platform but info about others is welcome).
> >
> > When a process context is changed, doesn't the memory manager TLB need
> > to be flushed? Is there a specific x86 instruction or register that
> > does this? Where is the code that causes this to happen? I've looked
> > at
> >
> > #define switch_to(prev,next) in
> >
> > linux/include/asm-i386/system.h
> >
> > but I don't know the x86 instruction set well and can't seem to find
> > the code I'm looking for.
> >
>
> Were you joking with this question or is it really an easy kernel
> question? I have been using linux for sometime now and wouldn't consider
> this an easy question.
I can see how it might be considered a joke but no, I wasn't joking.
I imagine that the question is quite easy for the right sort of
person: someone who knows the x86 instruction set and architecture
well (e.g. someone who has had to write assembler instructions for
kernel scheduler)
While there are many bright minds who don't work on the kernel and
therefore don't know the answer to my question that doesn't mean that
the answer isn't easy --- it is just obscure.
Sorry for the confusion, no doubt caused by my idiosyncratic world
view.
--jfc
--
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Jonathan Cano, IGS 6k*
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
------------------------------
From: Andreas Jaeger <[EMAIL PROTECTED]>
Subject: Re: glibc-help
Date: 03 Feb 1999 17:45:52 +0100
>>>>> Horst Piening writes:
> need some help with glibc:
> After installing the glibc with egcs-1.1.1 ( with the help of the
> Glibc2-HOWTO) my test program is still linked with the old libc5.
Have a look at the linking stage with gcc -Wl,-verbose and check which
libraries your program is linked against.
Andreas
--
Andreas Jaeger [EMAIL PROTECTED] [EMAIL PROTECTED]
for pgp-key finger [EMAIL PROTECTED]
------------------------------
From: Frankie East <[EMAIL PROTECTED]>
Subject: Re: GNUPro for Linux - Recommendations
Date: 03 Feb 1999 08:35:17 PST
> I haven't used DDD so I cannot compare them. But it does offer mixed source
> and assembly in debugging. And teh ability to change a registers contents at
> runtime. So with this I can not only get to the line that is blowing me up
> but step through every instruction that line compiled to and when I think I've
> located the error I can change registers to test my hypothesis. It is nice
> not having to change, compile, change, compile, ... to find the tricky bugs.
The popups and memory watch are nice. It will display the entire arrays
contents for example. And the gcc and egcs are newer. The C++ compiler is
improved over the freely available version of egcs with better template support
(I think they claim 37 levels of nesting though you couldn't pay me enough to
try to find out :). All in all I am very satisfied with it, but I keep both
Cygnus and default gcc installed. Then I set my search path for root to always
find the default gcc first, and my users to find Cygnus. Cygnus optimizes
application code nicely but optimizations aren't something you want to play with
when compiling kernels. The original gcc is very much designed to compile the
linux kernel (and vice versa).
I went ahead and got the bundle with Source Code Navigator. I am very pleased
with it over all but I wish they'd do like slickedit and let me use 'vi' mode in
the editor. Then I'd have the power of include graphs and symbol hopping AND be
able to not continually be annoyed at having to *mouse* around. But, since they
use ClearCase here, and I use SCCS and RCS, and Source Navigator integrates with
them all. I am happy. It is alot easier to get around 100000 lines of code
with it than with vi in an xterm for damn sure!
Hope I've helped...
...Frankie
------------------------------
From: "Me" <[EMAIL PROTECTED]>
Subject: 2.2.1
Date: Wed, 3 Feb 1999 09:47:03 -0900
after a few attempts of compling a new kernel it still will not boot
it goes to loading linux then locks up tight..
any ideas?
------------------------------
From: John Hasler <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: Linux Phase 2: A Consumer Operating System
Date: Wed, 3 Feb 1999 16:03:56 GMT
Christopher Browne wrote:
> RPM doesn't do what dselect does; it corresponds reasonably closely to
> dpkg. And while there have been some efforts to build 'package list
> managers' for use with RPM, none are as usable as of yet as dselect.
If you like dselect, wait til you see gnome-apt.
--
John Hasler
[EMAIL PROTECTED] (John Hasler)
Dancing Horse Hill
Elmwood, WI
------------------------------
From: Lakshmi Natarajan <[EMAIL PROTECTED]>
Crossposted-To:
comp.unix.programmer,comp.unix.advocacy,comp.unix.misc,comp.os.linux.misc,comp.os.linux.development.apps
Subject: Unix/Advanced Computing People
Date: Tue, 02 Feb 1999 21:22:49 -0500
Hi!
I am looking to network with people involved in Unix and advanced
computing
for both personal and professional reasons. I have been seriously
interested in the above for a long time. I have just joined USENIX and
hope it will be good for this purpose.
I am programmer, but my work environment does not have the concentration
in Unix that I am looking for.
I live in lower Westchester county, New York. I have got Linux on my PC
and
have been studying Richard Stevens' APUE as well as Bach's The Unix
Operating System. I am interested in Unix system programming,
administration, OS kernels, network programming, parallel processing
(don't
know much about it), Perl, Tcl/Tk, C, C++, Java, ... the whole bit!
I would appreciate some leads.
Thanks,
(Mr.) Lakshmi Natarajan.
--
####################################################
# Not by bread alone, or by music or by science, #
# but through all of them. #
####################################################
------------------------------
From: [EMAIL PROTECTED] (Jonathan Stott)
Subject: Re: 2.2.1
Date: 3 Feb 1999 17:54:23 GMT
Reply-To: [EMAIL PROTECTED]
In article <799ta0$[EMAIL PROTECTED]>, Me <[EMAIL PROTECTED]> wrote:
>after a few attempts of compling a new kernel it still will not boot
>it goes to loading linux then locks up tight..
>any ideas?
Did you remember to run LILO after building the new kernel? If you're sure
you did, run it again anyway and then tell us what the last thing the kernel
prints before things lock up.
-JS
--
Jonathan Stott xxxxxxxxxxxxxxxxxxxx
icbm://41.30.14N/81.36.36W/ [EMAIL PROTECTED] [EMAIL PROTECTED]
Physicist for hire - http://poly.phys.cwru.edu/~jstott/resume.html
------------------------------
From: Ronald Cole <[EMAIL PROTECTED]>
Subject: Re: glibc-crypt-2.0.112 somewhere ?
Date: 02 Feb 1999 18:02:54 -0800
Andreas Jaeger <[EMAIL PROTECTED]> writes:
> You should really read Ulrich's announcements about the test releases:
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> The crypt add-on can be found at
>
> http://www.ozemail.com.au/~geoffk/pgp/
>
> The 2.0.111 version is fine.
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
I would love to read Ulrich's announcements.
But, having royally pissed off Ulrich, I was reluctant to ask him why
there are no searchable maillist archives of the libc-alpha maillist
made available via http. I can find archives for egcs and many
different linux lists and, as a consequence, avoid asking many
embarrassingly stupid questions by making judicious use of them.
It's really too bad that I can't say the same about libc-alpha. I
haven't had to subscribe to a maillist in almost ten years and have
been very pleased with the signal to noise ratio in my mailbox. I'm
not about to change that now...
--
Forte International, P.O. Box 1412, Ridgecrest, CA 93556-1412
Ronald Cole <[EMAIL PROTECTED]> Phone: (760) 499-9142
President, CEO Fax: (760) 499-9152
My PGP fingerprint: 15 6E C7 91 5F AF 17 C4 24 93 CB 6B EB 38 B5 E5
------------------------------
** 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
******************************
