Linux-Development-Sys Digest #23, Volume #8 Sat, 15 Jul 00 20:13:10 EDT
Contents:
Re: Play mpeg without Xwindows (Scott Smith)
snafu 0.11 available ("Uncle")
Setting Environment Variables?? (Sonic Hedgehog)
Re: virtual time alarm? (Kaz Kylheku)
Sound programming. Is there a library for it? ([EMAIL PROTECTED])
Re: Raw disk access and the 'raw' program ([EMAIL PROTECTED])
Re: pure virtual function. (Mario Klebsch)
cannot compile tk ([EMAIL PROTECTED])
Re: pure virtual function. (Kaz Kylheku)
a file security proposal (Michael Gu)
RS-232 port logger? (Micah Lax)
Re: are PCI drivers char drivers? (Nix)
Re: GetModuleFileName() in linux (Nix)
Re: RS-232 port logger? ([EMAIL PROTECTED])
Re: a file security proposal (Kaz Kylheku)
Re: a file security proposal (Michael Gu)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Scott Smith)
Crossposted-To: comp.os.linux.hardware,comp.os.linux.misc
Subject: Re: Play mpeg without Xwindows
Reply-To: [EMAIL PROTECTED]
Date: Sat, 15 Jul 2000 05:43:04 GMT
>>> Is it possible to play a movie (mpeg or quicktime) without using Xwindows,
>>> only in Shell screen.
>>>
>>> Maybe using frame buffer ?
>>>
>>> Thanks a lot for your help.
>>>
>>> thierry
To view mpegs at the console in Linux (and hear them), go to freshmeat.net
look up smpeg. On smpeg's page they will have the source (maybe binaries)
and also a link to SDL which you will need.
You will need SVGA and have to tell SDL to support it (by default, it won't).
--
Scott Lacy Smith <[EMAIL PROTECTED]>
Student of Computer Science
"Nullus Anxietas" Denton, Texas, US
The University of North Texas
------------------------------
From: "Uncle" <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.development.apps,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.portable,comp.os.linux.security,comp.os.linux.setup
Subject: snafu 0.11 available
Date: Sat, 15 Jul 2000 07:04:18 GMT
Hello folks. Got a new version for everyone to try out.
point yer browser at:
http://www.geekcave.net/snafu.shtml
to get it.
New things in 0.11:
- configuration script
- rpmgt - rpm package manager/viewer for cmdline junkies
- some more menu selections
- the beginings of centralization and configuration stuff.
I'm planning to redesign the way snafu works next. What
I'm hopeing to work out is a way to customize the menus
and their functions without much fuss from within snafu.
The idea is that as problems aries or your installing something
new, snafu can be tailored to help speed up the debug/config
process by providing access to the files your editing and views
of the log entries you need to see.
While I have ideas on how im going to go about this, I'd love
to get ideas on how to go about modifying the menus with in the
program. Also, I will be porting this over to SCO sometime next week.
I would like to get this ported to as many platforms as I can.
Unfortunately, I only have access to linux and SCO unix. If anyone
would care to set up s.n.a.f.u. on any of the unix flavors and their way
of logging/configing, i would love to hear from you. [EMAIL PROTECTED]
I'm also planning on adding some monitoring deamons that can be
setup and control inside snafu or via the traditional command line.
I've recived some requests to have snafu watch for things like
reversce ICMP checksum errors and the like and have it
email them when it happens. I'm looking for anything y'all
would want to be notified of occuring at the time it happens.
Thats it this time. Thanks to everyone's interest and emails.
look forward to hearing from you.
Uncle
[EMAIL PROTECTED]
http://www.geekcave.net
------------------------------
From: [EMAIL PROTECTED] (Sonic Hedgehog)
Subject: Setting Environment Variables??
Date: Sat, 15 Jul 2000 07:59:56 GMT
I have installed the latest version of Redhat just recently now at
RedHat 6.2 as I am interested in learning a little about X Windows
programming. Qt is installed with Redhat choosing the development
option and this set the QTDIR environment setting to /usr/lib/qt.2.1.0
- this version didn't have the makefiles for the tutorials so I
downloaded another version. I renamed the directory to a more typing
friendly name of 'qt' - so now I need to change the QTDIR setting to
/usr/local/qt. But I have no idea how to do this. The Install file
told me to create a '.profile' file in the home directory (for bash
default for Redhat) but this didn't work. I read somewhere that
there was a setenv command but this didn't work either. Would someone
please tell me what I am doing wrong?!?
The install program also mentioned that problems could occur finding a
library and to run the sbin/ldconfig program, so I ran it to see what
happened, and this program is trying to run libraries that no longer
exist on the computer. Would someone please tell me how I modify this
file so that no longer happens?
A big thankyou in advance to all those who read the above message.
And an even biigger thankyou if you can help.
Thankyou and regards,
Vaughan Egan --- [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Kaz Kylheku)
Crossposted-To: alt.os.linux.mandrake
Subject: Re: virtual time alarm?
Reply-To: [EMAIL PROTECTED]
Date: Sat, 15 Jul 2000 09:03:55 GMT
On Fri, 14 Jul 2000 20:45:57 -0400, Elvis Chen <[EMAIL PROTECTED]> wrote:
>I ran some in-house program that was written in C/C++. The same program
>ran fine in another SMP-PC (with Mandrake 7.0). However, when I ran it
>under Mandrake 7.1, sometimes it finishes, but sometimes it is terminated.
>When it is terminated, I got an error message:
>
>"virtual time alarm".
It means that your program caught an unhandled SIGVTALARM. This would
be a strange thing to happen if you are not using setitimer()
with ITIMER_VIRTUAL.
The only time I have seen this error happen was many years ago on in an obscure
and little known user space threading library which used the virtual timer
alarm as the basis for a context switch. Its signal handler was set up using
the unreliable signal function rather than sigaction so once in a while the
program would get the signal with its pants down.
--
#exclude <windows.h>
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.development.apps
Subject: Sound programming. Is there a library for it?
Date: Sat, 15 Jul 2000 15:21:43 GMT
Graphics on Linux arn't a problem. You can use X windows or vgalib amongst
other things. Sound though I haven't figure out how to do short of getting
some miserable beep using \07 or the XBeep() function in Xlib.
Is there some sort of standard sound library in linux? The man pages and
HOWTOs have proved fruitless.
Thanks for any help
NJR
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Raw disk access and the 'raw' program
Date: Sat, 15 Jul 2000 17:25:03 GMT
On 15 Jul 2000 01:18:46 GMT Myke Morgan <[EMAIL PROTECTED]> wrote:
| 2.2.12 says:
|
| % raw -qa
| Cannot open master raw device '/dev/raw' (No such file or directory)
|
| 2.2.16 says:
|
| % raw -qa
| Cannot open master raw device '/dev/rawctl' (No such device)
It's probably a matter of the node in /dev not having been created.
Find out the major/minor device code for these, read the man page
for "mknod", login as root, and fix it.
--
| Phil Howard - KA9WGN | My current websites: linuxhomepage.com, ham.org
| phil (at) ipal.net +----------------------------------------------------
| Dallas - Texas - USA | [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Mario Klebsch)
Subject: Re: pure virtual function.
Date: Sat, 15 Jul 2000 18:06:11 +0200
[EMAIL PROTECTED] (Kaz Kylheku) writes:
>I understand. I don't really see the point of signatures. The mere presence of
>functions that have certain names and type signatures does not an interface
>make! If I understand it correctly, if I have void Lottery::Draw() and void
>Widget::Draw(), then the signatures feature lets me use the Draw function of
>both classes as common interface, even though neither class declared any
>conformance to such an interface. This could result in severe semantic errors.
I never used signatures in gcc, but this description reminds me of
interfaces in Java. What is the difference? In Java, any class that
implements an interface must be declared to implement that interface.
73, Mario
--
Mario Klebsch [EMAIL PROTECTED]
PGP-Key available at http://www.klebsch.de/public.key
Fingerprint DSS: EE7C DBCC D9C8 5DC1 D4DB 1483 30CE 9FB2 A047 9CE0
Diffie-Hellman: D447 4ED6 8A10 2C65 C5E5 8B98 9464 53FF 9382 F518
------------------------------
From: [EMAIL PROTECTED]
Subject: cannot compile tk
Date: Sat, 15 Jul 2000 20:14:02 +0100
Hi,
I am trying to compile Tk on LinuxPPC, but I keep on geting this error, any
suggestions...
./tkUnixDraw.c: In function 'TkScrollWindow':
./tkUnixDraw.c:95: void value not ignored as it ought to be
make: *** [tkUnixDraw.o] Error 1
If you know the location of a binary with the Xauthority security turned
OFF, then thats even better.
Thanks
Moray
------------------------------
From: [EMAIL PROTECTED] (Kaz Kylheku)
Subject: Re: pure virtual function.
Reply-To: [EMAIL PROTECTED]
Date: Sat, 15 Jul 2000 19:22:10 GMT
On Sat, 15 Jul 2000 18:06:11 +0200, Mario Klebsch <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] (Kaz Kylheku) writes:
>
>>I understand. I don't really see the point of signatures. The mere presence of
>>functions that have certain names and type signatures does not an interface
>>make! If I understand it correctly, if I have void Lottery::Draw() and void
>>Widget::Draw(), then the signatures feature lets me use the Draw function of
>>both classes as common interface, even though neither class declared any
>>conformance to such an interface. This could result in severe semantic errors.
>
>I never used signatures in gcc, but this description reminds me of
>interfaces in Java. What is the difference? In Java, any class that
>implements an interface must be declared to implement that interface.
Sorry about getting off topic with discussions about C++ which almost everyone
in here hates.
Signatures are type loose. If some class has functions that have certain names
and types, then you can independently declare a signature type which also has
the same functions in it having the same type signatures. You can then convert
a pointer to the class object to a pointer to the signature (without a cast, I
believe), and use the pointer as an interface to that object. Calls through the
signature functions get routed to the corresponding functions in the class. So
it's sort of like introducing an interface or a pure abstract base class type
of thing without requiring the class declare it. Call it ``interface stealing''
if you like.
There is an adequate description of signatures in the GNU info manual for gcc,
under the node C++ extensions.
--
#exclude <windows.h>
------------------------------
From: Michael Gu <[EMAIL PROTECTED]>
Crossposted-To: sci.crypt,comp.os.linux.security
Subject: a file security proposal
Date: Sat, 15 Jul 2000 19:38:49 GMT
If this has already been done, please ignore this message.
I suggest to add file encryption capability to the system ( kernal, or
whatever ). The general idea I have is following:
1. system distinguish whether a file is encrypted or not
2. when access an encrypted file, system will get the key by some
means. e.g. prompt for a password, read user config file ...
3. after having aquired the key, system will handle
encryption/decryption transparent to the calling party.
4. when create an file, the function call has an option to make it
encrypted.
In conclusion, this will enable system to handle user-level,
individual-file-level file security, and can be transparent to the
application. ( system can handle the key, so the application does not
even know the file is encrypted. )
The advantage of this is that:
1. Maximum compability. Any application can become secure file
access application even without change.
2. Flexibility. Key can be different for user, individual file ...
3. Performance. If well designed, performance should be better than
application handling security.
Extra cost:
each file may need some extra bytes for key information.
Implementation:
It should be pretty straight forward to implement. On the cypher side,
we probably need a hash function and a block cypher.
------------------------------
From: [EMAIL PROTECTED] (Micah Lax)
Crossposted-To:
linux.redhat.development,alt.os.linux,comp.os.linux.misc,comp.os.linux.development.apps
Subject: RS-232 port logger?
Date: Sat, 15 Jul 2000 19:43:23 GMT
Hi,
I need to log all incoming data from an RS-232 port on a linux machine
to a disk file.
Can anyone point me in the right direction?
Cheers,
[EMAIL PROTECTED]
------------------------------
From: Nix <$}xinix{$@esperi.demon.co.uk>
Subject: Re: are PCI drivers char drivers?
Date: 14 Jul 2000 22:23:43 +0100
Joe Pfeiffer <[EMAIL PROTECTED]> writes:
> the basic distinction
> between a char device and a block device is that a block device
> contains a filesystem -- in other words, it's a disk. This is really
> independent of whether it's a PCI device, though I can't remember
> having come across a PCI device that wasn't in fact a char device.
ITYM `a block device uses the page|buffer cache'. (Although this has got
a bit more complicated now with the addition of raw block devices.)
--
`I can guarantee it's no problem in my network, and if I don't get some
sleep soon, I'll guarantee it will become a problem in your network.'
--- Chris `Saundo' Saunderson deals with a late-night phone call
------------------------------
From: Nix <$}xinix{$@esperi.demon.co.uk>
Subject: Re: GetModuleFileName() in linux
Date: 14 Jul 2000 23:23:20 +0100
[EMAIL PROTECTED] writes:
> Actually, on linux systems (with /proc fs), reading the /proc/self/exe
> symbolic link gives the information.
No, not always.
> For example the ls command gives
>
>
> lrwx------ 1 basile parents 0 jui 14 18:08 /proc/self/exe -> /bin/ls
Not if it's been unlinked. And binaries should still work even after
they've been unlinked; this lets them, e.g., be upgraded totally
transparently to the people using them.
--
`I can guarantee it's no problem in my network, and if I don't get some
sleep soon, I'll guarantee it will become a problem in your network.'
--- Chris `Saundo' Saunderson deals with a late-night phone call
------------------------------
Crossposted-To:
linux.redhat.development,alt.os.linux,comp.os.linux.misc,comp.os.linux.development.apps
Subject: Re: RS-232 port logger?
From: [EMAIL PROTECTED]
Date: Sat, 15 Jul 2000 20:10:19 GMT
[EMAIL PROTECTED] (Micah Lax) writes:
> Hi,
>
> I need to log all incoming data from an RS-232 port on a linux machine
> to a disk file.
If the other end of that jack just goes to a standard serial port, you
can write a simple program to pipe the ouput of that port to a file.
tail, or possibly cat, might be good choices.
You can use setserial to set up the port and assign it a serial tty
(ttySx).
--
Eric P. McCoy ([EMAIL PROTECTED])
"Dude... my hands are huge. They can touch anything but themselves...
oh, wait."
------------------------------
From: [EMAIL PROTECTED] (Kaz Kylheku)
Crossposted-To: sci.crypt,comp.os.linux.security
Subject: Re: a file security proposal
Reply-To: [EMAIL PROTECTED]
Date: Sat, 15 Jul 2000 20:39:45 GMT
On Sat, 15 Jul 2000 19:38:49 GMT, Michael Gu <[EMAIL PROTECTED]> wrote:
>If this has already been done, please ignore this message.
>
>I suggest to add file encryption capability to the system ( kernal, or
>whatever ). The general idea I have is following:
It's spelled ``kernel'' and there are already solutions for this, like
CFS from AT&T Research or block device encryption.
> 1. system distinguish whether a file is encrypted or not
> 2. when access an encrypted file, system will get the key by some
>means. e.g. prompt for a password, read user config file ...
Hmm, read a plaintext user config file. And what will protect that?
Prompt for a password at access time? What if it's a GUI application?
At one time, some goofballs decided to make low level file access
calls of an operating system generate prompts at the user. The result was the
infamous ``abort, retry, fail, ignore'' question of MS-DOS.
CFS has a solution in the form of a command that creates plain-text mapping
which appears to be an NFS mounted filesystem. The encryption key is held
within the CFS daemon, which acts as the encryption/decryption agent on
behalf of the user; the user's application manipulate the plaintext tree,
and the CFS daemon performs encrypted I/O on the encrypted tree which is
stored somewhere else, possibly on another machine.
A user can have many trees, with distinct keys. When the user wishes to access
his or her encrypted directory, he or she invokes the CFS command for creating
the plaintext mapping. At that time, the password is entered. The program
passes the authentication to the CFS daemon, which creates the plain view in a
directory tree that is NFS mounted from the deamon by the user's kernel.
> 3. after having aquired the key, system will handle
>encryption/decryption transparent to the calling party.
> 4. when create an file, the function call has an option to make it
>encrypted.
So each application that creates files has to be modified to support
this new option in the file creation system call! Good plan!
--
#exclude <windows.h>
------------------------------
From: Michael Gu <[EMAIL PROTECTED]>
Crossposted-To: sci.crypt,comp.os.linux.security
Subject: Re: a file security proposal
Date: Sat, 15 Jul 2000 23:23:36 GMT
Kaz Kylheku wrote:
> On Sat, 15 Jul 2000 19:38:49 GMT, Michael Gu <[EMAIL PROTECTED]> wrote:
> >If this has already been done, please ignore this message.
> >
> >I suggest to add file encryption capability to the system ( kernal, or
> >whatever ). The general idea I have is following:
>
> It's spelled ``kernel'' and there are already solutions for this, like
> CFS from AT&T Research or block device encryption.
>
Sorry for the spelling, I am not a computer yet.
>
> > 1. system distinguish whether a file is encrypted or not
> > 2. when access an encrypted file, system will get the key by some
> >means. e.g. prompt for a password, read user config file ...
>
> Hmm, read a plaintext user config file. And what will protect that?
>
Did I say it's a plaintext user config file? Even so, why can't it be protected?
>
> Prompt for a password at access time? What if it's a GUI application?
>
I don't see why a GUI application can not pop up a window.
>
> At one time, some goofballs decided to make low level file access
> calls of an operating system generate prompts at the user. The result was the
> infamous ``abort, retry, fail, ignore'' question of MS-DOS.
>
Well, I am OK with these message, at least they tell you something is wrong. It's
better than something dies without a notice.
>
> CFS has a solution in the form of a command that creates plain-text mapping
> which appears to be an NFS mounted filesystem. The encryption key is held
> within the CFS daemon, which acts as the encryption/decryption agent on
> behalf of the user; the user's application manipulate the plaintext tree,
> and the CFS daemon performs encrypted I/O on the encrypted tree which is
> stored somewhere else, possibly on another machine.
>
> A user can have many trees, with distinct keys. When the user wishes to access
> his or her encrypted directory, he or she invokes the CFS command for creating
> the plaintext mapping. At that time, the password is entered. The program
> passes the authentication to the CFS daemon, which creates the plain view in a
> directory tree that is NFS mounted from the deamon by the user's kernel.
>
> > 3. after having aquired the key, system will handle
> >encryption/decryption transparent to the calling party.
> > 4. when create an file, the function call has an option to make it
> >encrypted.
>
> So each application that creates files has to be modified to support
> this new option in the file creation system call! Good plan!
>
Please try to think more freely! If the system handles password,
encryption/decryption, the application won't even know whether the file is
encrypted or not! Why it has to be modified?
>
> --
> #exclude <windows.h>
------------------------------
** 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
******************************
