Linux-Development-Sys Digest #403, Volume #6 Mon, 15 Feb 99 03:14:54 EST
Contents:
Re: Ncurses with g++ ("T.E.Dickey")
Re: glibc 2.1 ;) (Johan Kullstam)
"No Available PTYS" with 2.2.0 (David Moenssen)
DIEGO RAMOS ("Fans club de Diego Ramos, \"Loquitas por Diego\"")
Re: graphics lib?
Shared Directory Access ("FiSTY")
Re: 2.2/Red Hat 5.2 upgrades disaster! (Johan Kullstam)
Re: @@GLIBC (Wayne Schlitt)
Re: Ncurses with g++ (Bryan Hackney)
Managing User Accounts through code ("Mark")
Re: SMP Support ("garry flood")
Re: Help : Time Measurement (Mark Hahn)
Re: Looking for something to do (Glen Turner)
2.2/Red Hat 5.2 upgrades disaster! ("Frank T. Lofaro")
Re: @@GLIBC (Andreas Jaeger)
Re: Shared Directory Access (Felix Rauch)
----------------------------------------------------------------------------
From: "T.E.Dickey" <[EMAIL PROTECTED]>
Subject: Re: Ncurses with g++
Date: 15 Feb 1999 02:32:40 GMT
RD Alexander <[EMAIL PROTECTED]> wrote:
> Is there any reason why ncurses won't work with g++ and libg++?
> Following the instructions on the man page, I added '#include ncurses.h' to
> my C++ program. Afterwards it wouldn't compile - gcc complained of "parse
> error before 'char' on line 15". There isn't a 'char' on line 15. Removing
> the extra include allowed the program to compile again.
what version (g++ and ncurses)?
I occasionally test-compile ncurses with g++, but it did not compile
properly a few years ago before some additional work. Versions starting
with 4.0 should compile with g++ unless you are using some version which
I've not built with.
> --
> Rob Alexander (u8k62 [at or around, you know] ugk.keele.ac.uk)
> Probably
--
Thomas E. Dickey
[EMAIL PROTECTED]
http://www.clark.net/pub/dickey
------------------------------
Subject: Re: glibc 2.1 ;)
From: Johan Kullstam <[EMAIL PROTECTED]>
Date: 14 Feb 1999 10:02:53 -0500
Daren Scot Wilson <[EMAIL PROTECTED]> writes:
> Scary, the trouble some linuxers have had with the new glibc. Just today I
> compiled glibc 2.1, the new kernel 2.2.1, and toyed with egcs but didn't like
> it - programs it made ran slower than those form gcc 2.8.1, and it griped about
> things in the source code. Yet gcc compile glibc and linux just fine, despite
> the warnings in the glibc FAQ. And my Netscape 4.5 is running just fine,
> sending you this news email.
>
> Anyhow,the linux kernel version is important. I had upgraded the kernel to
> 2.2.1 before upgrading glibc. PPP dialing didn't work. Thought my ISP went
> bonkers. Upgraded the glibc to 2.1, everything's fine. Other people who
> upgraded the glibc first, using an older kernel, had other kinds of problems.
>
> Lesson: linux 2.2.1 (or 2.2.0) goes with and only with glibc 2.1, and older
> goes with older. Don't mix 'em.
not true. i am using linux-2.2.1 and glibc-2.0.7pre-whatever from
redhat. i did not have the ppp problem you experienced. i have
ppp-2.3.5 installed on my system. the kernel has a matching ppp-2.3.5
part.
i think a newer user ppp will work with older kernel ppp. ppp is a
bit kludgy since half is in the kernel and half is in user-space.
making these two parts agree sounds like a good idea to me.
as always ymmv.
--
Johan Kullstam [[EMAIL PROTECTED]] Don't Fear the Penguin!
------------------------------
From: David Moenssen <[EMAIL PROTECTED]>
Subject: "No Available PTYS" with 2.2.0
Date: Sun, 14 Feb 1999 22:44:51 -0500
hello all,
I've downloaded kernel version 2.2.0 and have encountered a bothersome
problem: when I start my desktop environment (KDE), I am not able to
bring up any xterms. Once I kill the X-server, there's a message that
says:
xterm: no available ptys
Does anyone know why this is happening? Is there some utility that I
need to download to update my devices?
any suggestions?
------------------------------
From: "Fans club de Diego Ramos, \"Loquitas por Diego\""
<[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.development.apps,comp.os.linux.hardware,comp.os.linux.m68k,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.setup,comp.os.linux.x,comp.os.ms-windows.win95.misc,comp.os.ms-windows.win95.moder
Subject: DIEGO RAMOS
Date: Sun, 14 Feb 1999 22:22:09 -0300
Si te gusta el actor Diego Ramos y queres asociarte a un fans club, queres
cambiar datos o tenes alguna informacion personal de el, escribinos a
[EMAIL PROTECTED]
Gracias. Angie
------------------------------
From: [EMAIL PROTECTED] ()
Subject: Re: graphics lib?
Date: 14 Feb 1999 17:39:29 GMT
ncurses
On Sat, 13 Feb 1999 17:05:40 -0500, Ricky Nelson <[EMAIL PROTECTED]> wrote:
>This is a multi-part message in MIME format.
>--------------D97A6CC3253258DE2C037C45
>Content-Type: text/plain; charset=us-ascii
>Content-Transfer-Encoding: 7bit
>
>I am new to Linux programming and I am trying to find a standard way of
>controlling the display in text mode using C++. Currently I use MS
>v1.52 compiler for DOS and it has a graphics library containing
>functions like:
>
>_settextposition()
>_clearscrean()
>_setvideomode()
>_setbkcolor()
>_settextcolor()
>etc...
>
>My question is what is the standard graphics library that is used in
>Linux. I am using Redhat v5.2.
>
>I looked at svgalib and ncurses but I couldn't seem to find a way to
>move the cursor around the way I am used to. Also is union REGS just a
>DOS way of accessing the video display? I couldn't seem to find any
>documentation on this in Linux either. Please forgive me for my
>ignorance, and any help would be greatly appreciated :)
>
>Ricy Nelson
>[EMAIL PROTECTED]
>
>
>--------------D97A6CC3253258DE2C037C45
>Content-Type: text/x-vcard; charset=us-ascii;
> name="ricky.vcf"
>Content-Transfer-Encoding: 7bit
>Content-Description: Card for Ricky Nelson
>Content-Disposition: attachment;
> filename="ricky.vcf"
>
>begin:vcard
>n:Nelson;Ricky
>x-mozilla-html:TRUE
>adr:;;2180 South Loudoun Street #120;Winchester;VA;22601;USA
>version:2.1
>email;internet:[EMAIL PROTECTED]
>note:ICQ#: 15049905
>x-mozilla-cpt:;-1
>fn:Ricky Nelson
>end:vcard
>
>--------------D97A6CC3253258DE2C037C45--
>
--
Michal Sabala aka Saahbs
Linux'er since 0.97 :)
UIUC Class of 2002; ECE
Linux, hardware, C, Html,
aviation, rc-air models
http://fly.to/saahbs
------------------------------
From: "FiSTY" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup,linux.redhat.misc
Subject: Shared Directory Access
Date: Sun, 14 Feb 1999 23:42:50 -0500
I'm trying to give a group of users full read/write/modify access to a
common directory. However, I can only manage to give read/write access, not
modify.
Does anyone know how to do this?
...Matt
------------------------------
Subject: Re: 2.2/Red Hat 5.2 upgrades disaster!
From: Johan Kullstam <[EMAIL PROTECTED]>
Date: 14 Feb 1999 23:50:00 -0500
[EMAIL PROTECTED] (N1ho) writes:
> Go browse http://www.linuxhq.com/change22.html for their information
> on upgrading to V2.2(.1). It's not clear to me that you can use
> glibc-2.1 for some things. Also, the kernel development was done
> using gcc-2.7.2.3, not egcs,
false. kernel development has used *both* gcc272s and the various
egcs incarnations.
> and I believe there were a number of issues with the kernel code
> that preclude compiling it with egcs at the moment. You might also
> need to update your modutils as well as several other components,
> some of which weren't in the RedHat 5.2 distribution. There is a
> long discussion on that page about the ttyp issues that you had
> problems with.
the 2.2.* kernels work just fine with, e.g., egcs-1.1.1. i am running
one now. i have been running egcs compiled kernels since october
(devels since 2.1.124). i used egcs on 2.0.* with a couple of patches
to ioport.c and string.h. egcs and 2.0.* is *not* recommended. but
egcs and 2.2.* is fairly stable.
modutils needs to be 2.1.121. i think there's an rpm in either
rawhide or contrib.
i don't know about the unix98 pty's. i hear you need glibc-2.1
(which i am not running). perhaps things like xterm and rxvt need
recompiling too.
--
Johan Kullstam [[EMAIL PROTECTED]] Don't Fear the Penguin!
------------------------------
From: Wayne Schlitt <[EMAIL PROTECTED]>
Subject: Re: @@GLIBC
Date: 14 Feb 1999 23:21:48 -0600
In <[EMAIL PROTECTED]> Andreas Jaeger <[EMAIL PROTECTED]>
writes:
>
> Let's assume that the semantics of bzero have changed between glibc
> 2.0 and 2.1 (that's just an assumption - it's not true). In that case
> glibc 2.1 would have two versions of bzero: [EMAIL PROTECTED] and
> bzero@@GLIBC_2.0. Your old binaries that have been linked against
^^^^ [EMAIL PROTECTED] ?
> glibc 2.0 and don't know of symbol versioning would be linked against
> the old interface (bzero@@GLIBC_2.0). But if you would compile a new
> program against glibc 2.1, it would be linked against the changed
> interface ([EMAIL PROTECTED]).
If the semantics of a library change, aren't you supposed to up the
library version number and just have libc.so.6 and libc.so.7? What
about all the other libraries in the system, should they do versioning
also? If not, why? If they do, how hard is it to add versioning to
all the libraries?
I gotta admit, that I'm slightly boggled by the need for this.
-wayne
--
Wayne Schlitt can not assert the truth of all statements in this
article and still be consistent.
------------------------------
From: Bryan Hackney <[EMAIL PROTECTED]>
Subject: Re: Ncurses with g++
Date: Mon, 15 Feb 1999 06:16:10 +0000
Reply-To: [EMAIL PROTECTED]
The curses and ncurses APIs are pathetic ancient
messes. The problems you are having are probably due to one of the
hundreds of macros in curses.h.
You need to isolate to the greatest extent possible the curses
calls, and include curses.h in as few (one) of your sources as
possible.
BH
RD Alexander wrote:
>
> Is there any reason why ncurses won't work with g++ and libg++?
>
> Following the instructions on the man page, I added '#include ncurses.h' to
> my C++ program. Afterwards it wouldn't compile - gcc complained of "parse
> error before 'char' on line 15". There isn't a 'char' on line 15. Removing
> the extra include allowed the program to compile again.
>
> --
> Rob Alexander (u8k62 [at or around, you know] ugk.keele.ac.uk)
> Probably
--
Bryan Hackney / BHC / bhackneyatexpress-news.net
*
------------------------------
From: "Mark" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,linux.dev.admin,linux.dev.newbie
Subject: Managing User Accounts through code
Date: Mon, 15 Feb 1999 08:26:11 +1000
I need to write a daemon that adds/removes users.
My programming experiance revolves around NT and linux is new to me.
I have the deamon up and running etc now i "just" need to maniplate the user
accounts.
How do i do it?
This is what ive come up with:
call adduser/useradd - You need to enter in details manually - how do you
capture
the input stream the adduser process uses to send it the user details?
Manually do it all - how do i create the encrypted field in the passwd file
or shadow file?
Im looking for the most easy/solid/portable solution i can get my hands on
and any help would be much appreciated.
Sorry if these questions seem a bit basic but i just cant get my hands on an
easy to use linux programming manual (ive been spoilt with VC++ help
manuals).
Mark
[EMAIL PROTECTED]
------------------------------
From: "garry flood" <[EMAIL PROTECTED]>
Subject: Re: SMP Support
Date: Sun, 14 Feb 1999 22:35:40 -0800
Now that we know how to tell if the new 2.2 kernel knows about more than 1
CPU, does anybody know of what types of applcations are written w/threads
and can take advantage of more than one CPU. ie compilers, linkers,
Powertools. I heard rumors that Adobe Photoshop in the NT environment is
thread aware, and runs nicely with SMP.
Bucky4me wrote in message <[EMAIL PROTECTED]>...
>I recompiled my kernel for smp but do not know how to check if the OS see's
>both. How can I do this?
>
>[EMAIL PROTECTED]
------------------------------
From: Mark Hahn <[EMAIL PROTECTED]>
Subject: Re: Help : Time Measurement
Date: 15 Feb 1999 07:18:40 GMT
> One of the first steps I have to do is the measurement of the real time
> interval between two consecutive executions of one function in a "for" loop.
> My problem is the multitask execution. Does anyone know if there
> is a way to accurately measure that time?
multitasking is not a problem, since if there is multitasking,
no measurements will be reasonable. that's why you should use
a real (wall-clock) time measurement, such as gettimeofday or:
typedef unsigned long long u64;
inline u64
gettick() {
u64 ticks;
asm volatile ("rdtsc" : "=A"(ticks));
return ticks;
}
if you really think that you want to try dealing with virtual time
while other processes are perturbing your CPU utilization (and thus cache),
you can use times().
------------------------------
Date: Mon, 15 Feb 1999 18:08:11 +1030
From: Glen Turner <[EMAIL PROTECTED]>
Subject: Re: Looking for something to do
hwj <[EMAIL PROTECTED]> writes:
> The assignment is supposed to take 6 months to complete at 3 days
> efficient work pr. week. This includes writing a report. There has
> got to be a fair amount of programming involved, and the focus
> should be on networking. I am considering to replace the routing
> database with a level compressed tree, but i am afraid there is not
> enough work in that job.
>
> Got any ideas?
IEEE 802.1Q VLAN and priority support including QoS routes
from gated.
--
Glen Turner Network Specialist
Tel: (08) 8303 3936 Information Technology Services
Fax: (08) 8303 4400 The University of Adelaide 5005
Email: [EMAIL PROTECTED] South Australia
------------------------------
From: "Frank T. Lofaro" <[EMAIL PROTECTED]>
Subject: 2.2/Red Hat 5.2 upgrades disaster!
Date: 15 Feb 1999 00:52:40 GMT
I am so frustrated I could explode!
I upgraded to the 2.2 kernel and upgraded all the required programs
and libs. And now my system is hosed as a result.
I am using egcs-1.1.1 with glibc-2.1 and kernel 2.2.1-ac5. (same
troubles with 2.2.0 and 2.2.1 though). This is on a Pentium II (thank
goodness, since I am recompiling so much)
I followed all the instructions to enable Unix98 PTY support and the
system did not use it. So I then (to force the issue) deleted all the
/dev/[pt]ty[p-z]? and /dev/[pt]ty[a-e]? devices. This happened:
telnetd: All network ports in use.
# script
Out of pty's
Terminated
#
xterm: no available ptys
So back the old style ptys went. cd /dev; ./MAKEDEV pty
That was fun, now wasn't it?
I am using glibc-2.1 and I had performed all the steps in
/usr/src/linux/Documentation/Changes to get it to work.
Does it work for anyone?
When I had installed util-linux2.9i I had to turn off PAM support. If
I had it turned on, it gave a compile error. With it off, it seems to
work.
Here's the compile error:
gcc -c -pipe -O3 -m486 -fomit-frame-pointer -I. -I../lib -Wall
-I/usr/include/ncurses -DNCH=0 -DSBINDIR=\"/sbin\"
-DUSRSBINDIR=\"/usr/sbin\" -DLOGDIR=\"/var/log\" -DVARPATH=\"/var\"
-DUSE_PAM=1 -DUSE_TTY_GROUP login.c
In file included from /usr/include/security/pam_misc.h:27,
from login.c:133:
/usr/include/security/_pam_macros.h:12: parse error before
`__extension__'
/usr/include/security/_pam_macros.h:12: `__len' undeclared here (not
in a function)
/usr/include/security/_pam_macros.h:12: initializer element is not
constant
/usr/include/security/_pam_macros.h:12: parse error before `if'
/usr/include/security/_pam_macros.h:12: conflicting types for
`__retval'
/usr/include/security/_pam_macros.h:12: previous declaration of
`__retval'
/usr/include/security/_pam_macros.h:12: warning: data definition has
no type or storage class
/usr/include/security/_pam_macros.h:12: parse error before `}'
login.c: In function `main':
login.c:797: warning: unused variable `wtmp'
login.c:295: warning: unused variable `oldSigHandler'
login.c: In function `rootterm':
login.c:1174: warning: `more' might be used uninitialized in this
function
make[1]: *** [login.o] Error 1
make[1]: Leaving directory
`/usr/local/src/util-linux-2.9i/login-utils'
make: *** [all] Error 1
Also, I get this fatal compile error I get when I try to have
util-linux-2.9i use the Unix98 pty support:
gcc -pipe -O3 -m486 -fomit-frame-pointer -I. -I../lib -Wall
-I/usr/include/ncurses -DNCH=0 -DSBINDIR=\"/sbin\"
-DUSRSBINDIR=\"/usr/sbin\" -DLOGDIR=\"/var/log\" -DVARPATH=\"/var\"
-DHAVE_OPENPTY -s script.c -o script
/tmp/ccbV6IYa.o: In function `getmaster':
/tmp/ccbV6IYa.o(.text+0x284): undefined reference to `openpty'
collect2: ld returned 1 exit status
make[1]: *** [script] Error 1
make[1]: Leaving directory `/usr/local/src/util-linux-2.9i/misc-utils'
make: *** [all] Error 1
I give up on Unix98 pty support. As for PAM... I download pam 0.66 and
compiled and installed that. Now su is hosed, and that d*mn
util-linux2.9i still won't compile with PAM turned on! (same stupid
error)
Here is what it did to my su:
su: error in loading shared libraries: /lib/libpam.so.0: undefined symbol: stat
Let me reiterate, all my libs and programs are upgraded and I am
running glibc-2.1!
I don't have anywhere near this kind of trouble with NT!
NT service packs go in a whole lot easier than this.
Heck, NT *reinstalls* go a heck lot easier than this.
Now Red Hat 5.2 was supposed to be 2.2 ready!
There is no Red Hat that really supports the 2,2 kernel.
And even if there was, having to install a whole distribution is
overkill.
I am not an idiot or a whiner. I have dealt with a lot of problems
(bugs and design mistakes) with Red Hat 5.1 and 5.2 (over a dozen
each) without a peep about it, until it got intolerable, and even then
I mostly just grinned and beared it. I have been doing Linux since
1993. I have contributed code to the kernel and have help fix bugs
here and there.
Well, now I have to go and figure out how to unhose my system.
(All this is time I could've spend enjoying life, or contributing to
linux. Now I will do neither. Hmm, I am beginning to see why NT is so
popular...)
------------------------------
From: Andreas Jaeger <[EMAIL PROTECTED]>
Subject: Re: @@GLIBC
Date: 15 Feb 1999 08:14:48 +0100
>>>>> Wayne Schlitt writes:
Wayne> In <[EMAIL PROTECTED]> Andreas Jaeger
Wayne> <[EMAIL PROTECTED]> writes:
>>
>> Let's assume that the semantics of bzero have changed between glibc
>> 2.0 and 2.1 (that's just an assumption - it's not true). In that case
>> glibc 2.1 would have two versions of bzero: [EMAIL PROTECTED] and
>> bzero@@GLIBC_2.0. Your old binaries that have been linked against
Wayne> ^^^^ [EMAIL PROTECTED] ?
You're right - that was a typo of myself.
>> glibc 2.0 and don't know of symbol versioning would be linked against
>> the old interface (bzero@@GLIBC_2.0). But if you would compile a new
>> program against glibc 2.1, it would be linked against the changed
>> interface ([EMAIL PROTECTED]).
Wayne> If the semantics of a library change, aren't you supposed to up the
Wayne> library version number and just have libc.so.6 and libc.so.7? What
Wayne> about all the other libraries in the system, should they do versioning
Wayne> also? If not, why? If they do, how hard is it to add versioning to
Wayne> all the libraries?
Changing the major number is the traditional way when functions change
their semantics. But it's easier to just use symbol versioning since
then your old binaries will still work - and it gives us more freedom
to fix broken functions.
Other libraries could use symbol versioning also but they don't need
to. It's just creating a version script and using it when linking.
Wayne> I gotta admit, that I'm slightly boggled by the need for this.
Going from libc.so.6 to libc.so.7 isn't acceptable at all since it
means changing your whole system to have libc4,
libc5,libc6,libc7,libc...
But going from glibc 2.0 to 2.1 with symbol versioning you don't have
to change your whole system.
Andreas
--
Andreas Jaeger [EMAIL PROTECTED] [EMAIL PROTECTED]
for pgp-key finger [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Felix Rauch)
Subject: Re: Shared Directory Access
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup,linux.redhat.misc
Date: 15 Feb 1999 09:09:34 +0100
[Followup-To set to comp.os.linux.setup]
In comp.os.linux.development.system FiSTY <[EMAIL PROTECTED]> wrote:
> I'm trying to give a group of users full read/write/modify access to a
> common directory. However, I can only manage to give read/write access, not
> modify.
In my understanding, writing a directory means modify it. Probably you
should give some more details about your problem.
- Felix
--
Felix Rauch, research assistant @ ETH Zurich, Institute for Computersystems
Homepage: http://nice.ethz.ch/~felix/ (includes PGP public key)
Email: [EMAIL PROTECTED] -> This article contains my personal views only <-
------------------------------
** 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
******************************
