Linux-Development-Sys Digest #629, Volume #6 Fri, 16 Apr 99 08:14:23 EDT
Contents:
ncurses question ("Dan Miller")
Write access problem of UFS filesystem (Solaris x86) (Chun Li)
aic7xxx-5.1.14-2.0.36: unable to apply patch (Dr. Peer Griebel)
RE: Download file like NetVampire or Getright ??? ([EMAIL PROTECTED])
where to download free CORBA for Linux? ("Enosh Chang")
Re: Timer Function in Linux? (Gerard van der Sel)
Re: threads, how????? (Peter Samuelson)
Problems with portmap after changing parameters in the kernel (Ulrich Eckhardt)
Re: Idea: Make a seperate "i686" tree for Redhat Linux 6.0 (Christopher Browne)
Kernel memory and fbcon issues (Timo Ketola)
The Fastest way to prepare for a successful IT career!
([EMAIL PROTECTED])
Re: M-systems binary only drivers & GPL (Peter Samuelson)
Re: Q: Is it possible to compile binaries for a different processor? (Craig Graham)
Re: pcibios_* functions (Peter Samuelson)
Re: CodeWarror for Linux (was: Re: Programming tools for ...) (Des Herriott)
Re: Timer Function in Linux? (Robert Kaiser)
Re: PCI DMA to user space possible? (Robert Kaiser)
RE: Download file like NetVampire or Getright ??? (Vincent)
Sound cards: how many samples have been processed? (Nick Bailey)
----------------------------------------------------------------------------
From: "Dan Miller" <[EMAIL PROTECTED]>
Subject: ncurses question
Date: Thu, 15 Apr 1999 16:30:22 -0700
Is there some way to make control characters appear as the standard 1-byte
characters, rather than as ^-chars?? For example, addch(2) displays
^B rather than the filled-in happy face... I'd like to modify this
functionality
if I possibly can...
------------------------------
From: Chun Li <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.misc
Subject: Write access problem of UFS filesystem (Solaris x86)
Date: Thu, 15 Apr 1999 22:57:30 -0700
Hi, there,
I have Linux (SuSE 6.0) and Solaris x86 (2.6) two operating systems
installed in my PC. Recently I updated the Linux kernel to version
2.2.5, and configed the kernel with read and write access to the
UFS filesystem (Solaris x86 filesystem). However, after the upgrade,
I can only have read access from Linux to Solaris (via mount Solaris
x86 filesystem to Linux, of course). I can NOT write any file to the
Solaris x86 filesystem. From documentation of the kernel configuration,
it is said that the newer kernel (2.2.x) has the ability to write
file in UFS filesystem (like Solaris x86 one). Did anyone experience
this problem and know how to fix it? Thanks in advance for your helps.
Regards,
-Chun
P.S. When you reply this post, please cc me a copy of email, as
I'm afraid of missing your valued post. Thanks again.
------------------------------
From: [EMAIL PROTECTED] (Dr. Peer Griebel)
Subject: aic7xxx-5.1.14-2.0.36: unable to apply patch
Date: Fri, 16 Apr 1999 09:00:41 GMT
Hi,
I'm trying to apply the newest aic7xxx patch to the current 2.0.36
kernel. I downloaded a stock kernel and tried to apply the patch using
patch -p0 < aic7xxx-5.1.14-2.0.36
The result is most of the hunks fail! Only two hunks succede :-(
Some examples examples:
Hmm... Looks like a unified diff to me...
The text leading up to this was:
==========================
|diff -U 3 -rN linux-2.0.36/Documentation/Configure.help
|linux/Documentation/Configure.help
|--- linux-2.0.36/Documentation/Configure.help Fri Jan 8 23:07:01 1999
|+++ linux/Documentation/Configure.help Tue Feb 2 13:58:57 1999
==========================
Patching file linux/Documentation/Configure.help using Plan A...
Hunk #1 failed at 1622.
1 out of 1 hunks failed--saving rejects to
linux/Documentation/Configure.help.rej
...
Hmm... The next patch looks like a unified diff to me...
The text leading up to this was:
==========================
|diff -U 3 -rN linux-2.0.36/include/scsi/scsicam.h linux/include/scsi/scsicam.h
|--- linux-2.0.36/include/scsi/scsicam.h Fri Jan 8 21:20:29 1999
|+++ linux/include/scsi/scsicam.h Tue Mar 23 01:12:22 1999
==========================
Patching file linux/include/scsi/scsicam.h using Plan A...
Hunk #1 succeeded at 14 with fuzz 2.
done
What am I doing wrong? I need this patch because I'm having much
trouble with the current aic7xxx driver.
Thank you.
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: alt.linux,alt.os.linux,comp.os.linux.x
Subject: RE: Download file like NetVampire or Getright ???
Date: Fri, 16 Apr 1999 09:12:27 GMT
In article <[EMAIL PROTECTED]>,
Vincent <[EMAIL PROTECTED]> wrote:
> I'm looking for a software which download file in multiple time like the
> NetVampire or Getright under Windows ?
> If this software exist, where can I found it ?
>
> Thanks
>
> [EMAIL PROTECTED]
>
>
try wget (man wget)
it is powerfull but not easy to use
Cedric
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: "Enosh Chang" <[EMAIL PROTECTED]>
Subject: where to download free CORBA for Linux?
Date: Fri, 16 Apr 1999 17:28:50 +0800
Hi, all:
I am a beginner of Linux. And I want to learn CORBA.
So, could you tell me where to download free CORBA for Linux?
Thanks a lot
------------------------------
From: Gerard van der Sel <[EMAIL PROTECTED]>
Subject: Re: Timer Function in Linux?
Date: Fri, 16 Apr 1999 09:45:26 +0200
Reply-To: [EMAIL PROTECTED]
Peter Samuelson wrote:
>
> [Mridula Shrikant Pradhan <[EMAIL PROTECTED]>]
> > I would like to know whether in Linux the timer function is available
> > in microseconds.
>
> Oh, you can get it in microseconds -- see nanosleep(). But it won't be
> that accurate. The Linux scheduler (by default on x86) only wakes up a
> process every 10000 microseconds, and there's no guarantee your process
> would be the next one to wake up at that time, unless it were un-niced
> appropriately. (Not even then, really.)
>
> The realtime clock, which you can read all about in
> /usr/src/linux/Documentation/rtc.txt, gives you resolution up to 8kHz,
> which of course is 120us. However, to use it you have to compile RTC
> support into your kernel (the config option isn't selected by default,
> at least on x86).
>
> --
> Peter Samuelson
> <sampo.creighton.edu!psamuels>
As a newbie (3 weeks on a Linux box) I found RTC functionality. (RedHat
5.2, kernel 2.0.36 everything installed). When you have the correct
hardware the resolution is even better: 32kHz (which is 30 us). I use it
at home for an driver for my modelrailroadtrack.
I have a question about RTC on linux. The only thing I could find was
that it is necessary every 11 minutes, when you have installed a 'DFC'
(atomic radio clock) board on your PC. Is there an other use for this
device?
--
Met vriendelijke groet,
Gerard van der Sel
Mailto:[EMAIL PROTECTED]
"De dinosaurussen hadden hun komeet, wij hebben de Windows computer" -
me
"The box said: 'install on Windows 95, NT 4.0 or better'.
So I installed it on Linux."
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Crossposted-To: comp.os.linux.development.apps
Subject: Re: threads, how?????
Date: 15 Apr 1999 20:34:56 -0500
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Sander Zijlstra <[EMAIL PROTECTED]>]
> I want to make three threads running simultaniously, and processing
> on the same data. One thread has to read in the data from a socket
> (later on it will be a I/O-device), while the next thread performs a
> FFT on that data, and finally the last thread demodulates it (because
> the data represents a modulated signal) and writes it to a file...
A simplistic approach would be to use a pipe, exactly as though they
were separate processes (and indeed they could be separate processes).
Call pipe() to get a pair of file descriptors. The first thread
write()s data in one FD and the next thread read()s it from the other
FD. This limits parallelism due to the relatively small buffer Linux
will allocate for the pipe.
As someone else already suggested, what you probably want are circular
buffers. Create a buffer of, say, 32k. Have two global ints for the
start and end of the data. Thread #1 writes into the buffer and only
modifies `end', blocking if `(end-start)%bufsize' would be less than 0
(buffer full). Thread #2 only modifies `start' as it reads data,
blocking (see sched_yield()) on `(end-start)%bufsize > 0' (buffer
empty). Same between Thread #2 and #3. No mutexes required, since
nobody is modifying the same variables.
It also sounds like Thread #1 might not be doing anything useful and
could be eliminated. Depends on the complexity of "read in the data
from an I/O device", I suppose -- if you need quasi-realtime and
buffering it might be necessary.
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: Ulrich Eckhardt <[EMAIL PROTECTED]>
Subject: Problems with portmap after changing parameters in the kernel
Date: 16 Apr 1999 09:45:52 GMT
Hi,
i want to use the informix database with kernel 2.2.5 For doing
this i have to change the kernel parameters NR_OPENFILES
NR_OPEN and __FD_SETSIZE to 5120 in the files fs.h,
posix_types.h and limits.h
The database works fine now with this kernel. But
portmapper failes with the error message :
informix:~ # portmap -dv
svc_run: - select failed: Invalid argument
portmap[4481]: run_svc returned unexpectedly
Aborted
So i got the portmap_5.beta.tar.gz package and compiled this,
but i got the same error message. Then i had a look in the code
and found a call to svc_run() in the main procedure of portmap.c
But i haven't found the function declaration for svc_run. Where
can i find this function or has anybody an idea whats happens here.
Thanks for any help
Uli
--
Ulrich Eckhardt Tr@nscom GbR
http://people.frankfurt.netsurf.de/uli http://www.transcom.de
Lagerstra�e 11-15 A8
64807 Dieburg Germany
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: comp.os.linux.misc
Subject: Re: Idea: Make a seperate "i686" tree for Redhat Linux 6.0
Reply-To: [EMAIL PROTECTED]
Date: Fri, 16 Apr 1999 05:47:34 GMT
On 8 Apr 1999 17:07:45 -0400, Alexander Viro <[EMAIL PROTECTED]> wrote:
>In article <[EMAIL PROTECTED]>,
>Johan Kullstam <[EMAIL PROTECTED]> wrote:
>>not if you're from a commonwealth country - which includes new zealand
>>(where is the old zealand btw?). in *english* (as opposed to american
> ^^^^^^^ - Zeeland.
>
> Gaak... Across the North Sea (looking from England, that is).
>Netherlands. Heck, they *really* don't teach history and geography
>in schools, or what?
Not even in commonwealth countries... :-(
I was not aware of that; the surprising factor is that it was named
after a "foreign" nation, or am I showing *complete* ignorance in being
unaware of a transition of New Zealand from Dutch control to English
control?
--
"UNIX was not designed to stop you from doing stupid things, because
that would also stop you from doing clever things." -- Doug Gwyn
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/lsf.html>
------------------------------
From: Timo Ketola <[EMAIL PROTECTED]>
Subject: Kernel memory and fbcon issues
Date: Fri, 16 Apr 1999 10:34:45 GMT
Hi folks,
I'm sorry if these are too trivial - but...
I have put Linux 2.2.5 into a nonstandard hardware - that is no BIOS,
LCD display...
Its working now but 3 questions arose:
1) Display hardware needs a memory block from the main ram aligned to a
64K boundary. How do I best allocate this. Now I reduced the memory size
reported to the kernel and ioremap'ped the top memory block into the
virtual memory. But as this is not in control of the driver I think its
not the correct way.
2) The display is 16 level gray scale, 4 bpp, 2 pixels in a byte. In
what visual type does it map nicely. It seems to me that the fbcon
treats it as a pseudocolor device. At least the appearing logo is
indexed color picture - I think. I don't have very clear idea how this
affects for example the applications.
3) The kernel message says that 412k of the memory is reserved. I'm
afraid that the 640k-1M area is reserved and lost. If thats true, how
can I get them back.
Thanks in advance!
I don't mind if pointed to FAQs, docs or correct newsgroups.
--
Timo
--
Epec Oy
Timo Ketola
Box 194
FIN-60101 Sein=E4joki
direct: +358 6 2170 222
switch: +358 6 2170 111
fax: +358 6 2170 110
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Subject: The Fastest way to prepare for a successful IT career!
Date: Thursday, 15 Apr 1999 20:43:09 -0600
For More Information:
http://www.cbtnow.com/index.html?ASCID=134
The Fastest way to prepare for a successful IT career!
CBT Systems offers the fastest and easiest way to prepare you for a successful career
in the ever-expanding Information Technology (IT) industry. Our Self-Study Courses are
100% computer-based training (CBT) available through electronic download or CD-ROM to
give you flexibility unmatched by traditional training methods.
For More Information:
http://www.cbtnow.com/index.html?ASCID=134
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: M-systems binary only drivers & GPL
Date: 15 Apr 1999 21:38:15 -0500
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
Oh, fun, yet another GPL debate. Why *do* I bother to get involved.
[Tristan Wibberley <[EMAIL PROTECTED]>]
> The GPL makes no statement that static linking is simply always
> illegal. If the kernel has been specially compiled to keep the code
> text for the driver well separated and clearly identifiable, then it
> is legal according to the GPL.
...but practically speaking, how are you going to manage to do that?
Yes, you can figure out which part of the kernel is binary-only by
using System.map, but I can't, for example, just recompile the *rest*
of my kernel. To do that, the distributors must distribute .o files of
their own code, making sure no GPL code is in those .o files (which
could be tricky if they want to use inline functions from GPL header
files).
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: Craig Graham <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.alpha,comp.os.linux.questions,comp.os.linux
Subject: Re: Q: Is it possible to compile binaries for a different processor?
Date: Fri, 16 Apr 1999 07:50:20 +0000
Chris Gahan wrote:
>
> I was wondering if it is indeed possible to compile x86 binaries, for
> example, on an Alpha. I thought about this, and it definitely seems
> possible. I figure that if you get the x86 version of GCC, and
> compile it on an Alpha, it would spit out x86 binaries. Am I way off
> with this, or is this something that I can do right now? :)
Yeah - I used to compile Playstation (MIPSr3000) binaries with GCC
on DOS and Linux - it's pretty easy to build a cross compiler, just
get the full GCC & binutils source, and configure with the target and
host set to whatever you need.
Craig.
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: pcibios_* functions
Date: 15 Apr 1999 21:58:05 -0500
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Tim Linehan <[EMAIL PROTECTED]>]
> > I am trying to compile a program that uses the
> > pcibios_read_config_byte () and similar functions
[Peter Pointner <[EMAIL PROTECTED]>]
> These functions are kernel functions, and I think you can't use them
> from user space.
Correct. That's why <linux/pci.h> marks those functions as off-limits
to a user program via `#ifdef __KERNEL__'.
> You can use the proc interface, e. g. "cat /proc/pci" for a 2.0.x
> kernel. On 2.2.x the path and content might be different.
If you have the 2.0-/proc/-pci-compatibility option on, /proc/pci still
works. Otherwise you have to read /proc/bus/pci/devices and decode the
numbers, which you can do using information in /usr/share/misc/pci.ids
from the pciutils package.
For reading and writing the pci config info, look for files
(representing devices) in the subdirectories (representing PCI busses)
under /proc/bus/pci/.
> Of course you can read these "files" from your program. For 2.2.x
> kernels, there is a utility called lspci (which gets the information
> probably via proc).
Yes, the pciutils package, i.e. lspci and setpci, is a perfect example
of how to use the new interface correctly. (The "sample
implementation", if you will.)
> Peter
Peter
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: [EMAIL PROTECTED] (Des Herriott)
Crossposted-To:
comp.os.linux.advocacy,comp.os.linux.development.apps,comp.os.linux.help,comp.unix.programmer
Subject: Re: CodeWarror for Linux (was: Re: Programming tools for ...)
Date: 16 Apr 1999 11:21:52 GMT
On Sat, 10 Apr 1999 15:27:33 -0500, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>
> That pretty much sums up all the responses here: that the lack of
> progress is because of the emotional attachment most afficionados here
> have to obsolete paradigms along with failure to even attempt to see
> outside of their little, little boxes. Good luck dinosaurs.
Don,
You've spouted a lot of rhetoric and provided very little in the way of
solid argument as to just why a GUI should be better than a CLI for
software development.
You've responded to the reasoned arguments of others with insults.
You've made some fairly gross generalisations, all of them false. The
paradigms we Unix people are *not* obsolete, despite what some large
software corporation would like you to believe.
Graphical IDE's are *not* objectively better than command-line
development environments. If you prefer them, fine, good luck to you,
but many of us are extremely productive with a Unix command-line. If
you can't handle that, no problem, use your IDE with its brightly-
coloured buttons and ever-so-helpful dialog boxes.
But don't accuse us of lack of progress or obsolesence, because it
simply isn't the case.
--
Des Herriott, Oracle Corporation UK Ltd.
[EMAIL PROTECTED]
- speaking for myself, not my employer.
------------------------------
From: [EMAIL PROTECTED] (Robert Kaiser)
Subject: Re: Timer Function in Linux?
Date: 16 Apr 1999 10:20:02 GMT
Hi,
In article <7f10re$5ae$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Mridula Shrikant Pradhan) writes:
> I would like to know whether in Linux the timer function is available in
> microseconds.
Depends on what you want a timer function to do: If you want to
suspend/resume processes under timer control (e.g. sleep()/usleep()
nanosleep()), your timing resolution is limited to the system ticker
period (see also Peter Samuelson�s response).
However, if you just want to know the current time, I think
gettimeofday() will give you resolution down to microseconds.
Hope this helps
Rob
============================================================
Robert Kaiser rkaiser AT sysgo DOT de
Sysgo RTS GmbH
Mainz, Germany
------------------------------
From: [EMAIL PROTECTED] (Robert Kaiser)
Subject: Re: PCI DMA to user space possible?
Date: 16 Apr 1999 08:21:39 GMT
In article <[EMAIL PROTECTED]>,
Bryan Hackney <[EMAIL PROTECTED]> writes:
> Robert Kaiser wrote:
>>
>> With the mmap() approach it would look something like this:
>>
>> ioctl(framegrabber_device, TELL_BUFFER_SIZE_TO_DRIVER, buffersize);
>> /* .. driver allocates buffer memory.. */
>>
>> ioctl(framegrabber_device, GRAB_VIDEO_IMAGE, 0);
>
> No, the frame device will tell the driver how big it is, and anyway, the
> memory is on the frame device. See below.
Not in the case of the device I had to deal with: it just DMA�s
a stream of video data into the host CPU�s RAM. There is no
RAM on the PCI board itself (at least none that would be visible to
the programmer). And the size of the buffer (i.e. the picture
I�m acquiring) is defined by the application (and it may change
from picture to picture).
> The scatter/gather of pages is handled transparently by the MMU. It really
> does look like one giant contiguous space in user space, provided that the
> pages get mapped in correctly.
>
> [snip]
>
> Now, in the case of frame buffers where big contiguous areas are required,
> this memory is on the device itself. You are _not_borrowing_ pixel memory
> from the system, you are mapping the device's memory into the system. Big,
> big, important distinction.
Yes of course, but this is a whole different story. (AFAIK, this is
basically how XFree86 acesses the graphics card�s video memory), but
this is not what I was talking about. The framegrabber I have to
deal with does not have local video memory. It uses the motherboard�s
RAM for that.
> If you use malloc() to get virtual memory to copy into and out of some
> frame device, you are going to copy into physically scattered but virtually
> contiguous pages.
>
> I though the main purpose here is to minimize copying.
No, the only copying that happens is the PCI busmaster writing
_directly_ into the user space buffer that I obtained from malloc().
The CPU does not get involved in the copying at all.
> Other unixes use readv() and writev() to avoid copying. I think all read()/
> write() calls are copying calls in all unixes I've dealt with at this level.
This is definitely not true for the unixes I have dealt with!
The point is that, as long as you don't write device drivers, you
don't even _see_ wether a particlar driver uses DMA or copying to
implement read()/write(). The only way a user could tell is by
looking at the speed: if a device can transfer more than -say-
20 Megs per second without being a CPU hog, then the device most
likely uses user-space DMA. And this is how it should be: it
is none of the application�s business to know about such hardware
details!
readv() and writev(), AFAIK, are a completely different story: They
are yet another complexion that was apparently introduced to deal
with the reading/writing of _virtually_ scattered data in a single
system call. They are in fact rarely used and I do not think they
have anything to do with DMA (otherwise, the size of the chunks you
can legally pass to readv()/writev() would have to by limited to
the system�s page size.
> You can implement a much cleaner device interface with mmap() and a couple
> of ioctls() than with readv()/writev().
But what could be cleaner than a simple read()/write() ?
Maybe that framegrabber thing was a bad example after all as it
apparently bears the implication that a frameGRABBER always
includes a frameBUFFER. The board I had in mind (and those _do_
exist) is more like a fast data acquisition board, a fast A/D
converter, if you will. It might even be a harddisk I want to read
from real fast.
You know, it's really strange: nobody seems to have a problem with
the fact that read()/write() transfer user data from/to a device. But
as soon as the underlying hardware is capable of doing this with DMA,
everybody seems to just assume that it can�t be done with the same
plain and simple read()/write() API. -- Why ??
Cheers
Rob
============================================================
Robert Kaiser rkaiser AT sysgo DOT de
Sysgo RTS GmbH
Mainz, Germany
------------------------------
From: Vincent <[EMAIL PROTECTED]>
Crossposted-To:
alt.linux,alt.os.linux,comp.os.linux.x,comp.os.linux.development.apps,comp.os.linux.development,comp.os.linux.hardware
Subject: RE: Download file like NetVampire or Getright ???
Date: Fri, 16 Apr 1999 09:38:37 +0200
I'm looking for a software which download file in multiple time like the
NetVampire or Getright under Windows ?
If this software exist, where can I found it ?
Thanks
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Nick Bailey)
Subject: Sound cards: how many samples have been processed?
Date: Fri, 16 Apr 1999 12:28:39 +0100 (BST)
This is a quesion about Studio, our small sound editor
(http://www.ee.leeds.ac.uk/homes/NJB/Software/Studio/screens.html) but
the implications are wider than that, so I am asking around for
opinions. I think "c.o.l.dev.system" is the place to ask this
application oriented question, for reasons which I hope will be clear.
We are going to be releasing a pre-version1 shortly. Among the things
we need to do is to find the number of samples played by the sound
system in order to move a line along the audio graph. Development is
happening under Solaris (for reasons of space and access: our PC/Linux
lab gets locked after 5pm -- Unix System 9-5? 8-), which has a facility
to read the number of samples the sound card has processed via an
ioctl. As far as I know (please let this be wrong!!), there is no such
facility in OSS.
The C program responsible for spitting out the audio files will have to
be ported, and it is I who will probably have to do it. After poking
around, there seems to be a choice of action:
1. Use ALSA. This has a call to do what I want.
struct snd_pcm_playback_status {
...
int scount; /* number of bytes processed from
playback start (last underrun) */
...
};
I've not personally got a problem with this, but Red Hat (again, as far
as I know) doesn't have ALSA support, and it would also be nice to have
the program working without requiring the installation of a whole new
sound system. I'm all for ALSA, but it might be an overkill for just
this one problem.
2. Change the structure of the Solaris version like this:
Solaris:
fork(): parent plays sound file
child does while(1){ usleep(a bit); printf("%d", ioctl(get bytes
processed)); }
Tcl/tk app sits reading this output through a pipe and updating the
display.
Linux/OSS:
ioctl(make soundframents short); /* to get reasonable resolution when
child does ioctl()? */
shmget(a piece of memory to store byteswritten in);
/* leaving out all the SYSV IPC nonsense... */
fork(): parent plays sound f. keeping byteswritten up to date
child does while(1){ usleep(a bit); printf("%d", byteswritten -
ioctl(empty buffer space)); }
Now, this looks like a dirty hack to me, and I'd prefer to use method 1,
but at least (2) won't break when people don't install ALSA properly...
The trouble I just don't know the status of the sound stuff in the
kernel. Is 2.2 basically ALSA? Are Red Hat/Debian/etc going to start
using it? I see there are .deb files. What are the issues with ESD?
Will that stop (1) above working at all?
I'd appreciate any opinions from people with sound experience (no pun
intended!)
Nick/
PS: This is a genuine request for comment, I don't want to start a RH vs
Debian vs Your-Fave-Dist flame war! We use RH /and/ Debian /and/
Slackware, and like them all. 8-)
--
Dr. N.J.Bailey-----------------------------------------------
Lecturer in Electronic and Electrical Engineering
University of Leeds, Woodhouse Lane, Leeds,
LS2 9JT. UK.-------------------------------------------------
http://www.ee.leeds.ac.uk/homes/NJB/
------------------------------
** 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
******************************
