Re: ucd snmp
On Fri, Sep 10, 1999 at 03:27:19PM -0500, Nathan E Norman wrote: to upgrade our prodution servers to potato (and libc 2.1). In the potato pkg list libucdsnmp3.6 depends on libc6 =2.1; am I screwed? (I admit to not investigating fully whether libucdsnmp3.6 needs libc 2.1 That dependency is auto-generated. If you build with libc6 2.0, you'll get a dependency on plain libc6. # Remove execute bit from MIBS. chmod -x debian/tmp/usr/share/snmp/mibs/* dh_movefiles dh_movefiles: I was asked to move files from debian/tmp to debian/tmp. make: *** [install-stamp] Error 1 I'm not a maintainer and admittedly pretty stupid about how to build debian packages correctly - I used fakeroot dpkg-buildpackage -B -uc Any ideas? I guess dh_movefiles is part of debhelper? You need a 2.0.x version of debhelper. You should be able to install the potato version on a slink system without any problems. David -- David Engel [EMAIL PROTECTED]
Re: SNMP
On Fri, Aug 27, 1999 at 08:38:16PM +0200, Ralf G. R. Bergs wrote: Why is it that I get incorrect results for my ATM interface atm0? The interface thruput that MRTG displays is only about one half (or even less) of the real thruput. Is this due to a problem in the SNMP agent, or is it a bug in the ATM driver or protocol stack? How can I debug the problem? Are the stats in /proc/net/dev correct? If they are, it is most likely a bug in the SNMP agent. If they are not correct, the problem is probably in the driver. Let me know if it's the former. David -- David Engel [EMAIL PROTECTED]
Re: SNMP
On Thu, Aug 26, 1999 at 04:42:33PM -0400, Ken Long wrote: I just ran some package upgrades today and some of the packages that got updated were the snmp packages. (now at 3.6.2-8) I see that they now also mention something about being UCD. The new SNMP packages are built using the UCD SNMP source. The old SNMP packages were built using a different, Linux-specific SNMP source. Well, my question is this...after upgrading the packages, my MRTG machine can no longer talk to my machine to get readings. Watching in xconsole, I'm seeing a message come up saying: Aug 26 16:36:06 lusitania snmpd: Connection from XXX.XXX.XXX.XXX refused The new SNMP agent uses a different configuration file format from the old one. The default configuration only allows access to the MIB2 system group. You will need to edit /etc/snmp/snmpd.conf to allow greater access. David -- David Engel [EMAIL PROTECTED]
Re: SNMP
On Thu, Aug 26, 1999 at 05:27:04PM -0400, Ken Long wrote: Well, since reading your message, I've been trying.pretty unsuccessfully, however. I notice with the default configuration, I can't even connect from the localhost! I don't suppose there's a nice little HOWTO file somewhere, is there? Please see /usr/doc/snmp/FAQ.gz. David -- David Engel [EMAIL PROTECTED]
Need a nice, large, fixed font
Hi, Does anyone know of a large, nice looking, fixed-size font for X? With my poor vision, I need something slightly bigger than the standard 10x20 font for extended use. The 12x24 font is about the right size (maybe a tad too large), but it too ugly. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ld.so.conf question
On Tue, Apr 07, 1998 at 01:39:59PM -0400, Brian White wrote: I have the following ld.so.conf file: /usr/X11R6/lib/Xaw95 /usr/X11R6/lib/Xaw3d /usr/local/lib /usr/lib/libc5-compat /lib/libc5-compat /usr/X11R6/lib /usr/X11R6/lib /usr/X11R6/lib Question: Shouldn't /usr/X11R6/lib be before the libc5-compat directory Not necessarily. As long as you don't have any duplicate libraries that aren't tagged properly, it doesn't matter what order they are in. and shouldn't there only be one of them? There should only be one, but the duplicates shouldn't cause any problems. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ld.so.conf question
On Tue, Apr 07, 1998 at 03:10:11PM -0400, Brian White wrote: /usr/X11R6/lib/Xaw95 /usr/X11R6/lib/Xaw3d /usr/local/lib /usr/lib/libc5-compat /lib/libc5-compat /usr/X11R6/lib /usr/X11R6/lib /usr/X11R6/lib Question: Shouldn't /usr/X11R6/lib be before the libc5-compat directory Not necessarily. As long as you don't have any duplicate libraries that aren't tagged properly, it doesn't matter what order they are in. The problem I'm encountering is that linking -lXmu seems to cause the libc5-compat version of libXtk to get included and then I get warnings about libc5/libc6 conflicts. I was wondering if this could be the cause. Could be. What does ldd report for your libc5-compat version of libXmu? David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: snmpd weirdness
On Mon, Mar 09, 1998 at 12:03:48PM +0100, Peter Gervai wrote: So snmpd. It seems - and some fellow admins reported that they noticed the same effetct - that snmpd's ifInOctets and ifOutOctets are incorrect. More specifically, they are usually the half of the real values, and in my case it seems that ifIn and ifOut is almost the same. Kernel support for ifInOctets and ifOutOctets was not added until kernel version 2.1.15. When snmpd is run on older kernels, the values of ifInOctets and ifOutOctets are estimated based on the number of received and transmitted packets. To get accurate values of ifInOctets and ifOutOctets, you must run kernel 2.1.15 or later. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: ld.so linking conflict
On Mon, Nov 24, 1997 at 03:54:59PM -0500, Brian White wrote: then I get the warning: ld: warning: libdl.so.2, needed by /usr/lib/libtk8.0.so, may conflict with libdl.so.1 I have been unable to find where this libdl.so.1 is coming from since the dl libraries are: You probably have a left over link in /lib. Perhaps you 'forced' something that conflicts with libdl1-dev. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: kde for debian's menu program
On Sat, Oct 18, 1997 at 12:38:33AM -0400, Paul Miller wrote: Has anyone added a kde converter for Debian's menu program (like those of all afterstep, fvwm2/95, etc.)? If so, where can I download it? A new version of KDE for Debian was uploaded today. It should hit ftp.debian.org and mirrors in a day or two. It will support Debian's automatic menu system. FYI, there are some problems with the new version and it will probably take a revision or two to sort things out. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Debian on Dell Latitude Laptop
Has anyone successfully installed Debian 1.3.1 on a Dell Latitude XPi laptop? If so, what it the secret? When I try booting from the install/rescue disk, it gets about halfway through Loading linux... and then reboots. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
[coteau@westriv.com: Network Problem]
-Forwarded message from coteau [EMAIL PROTECTED]- Return-Path: [EMAIL PROTECTED] Received: from elvis.sw.ods.com ([160.86.13.7]) by elo.sw.ods.com with esmtp id m0weKxO-00014ZC (Debian Smail-3.2 1996-Jul-4 #2); Wed, 18 Jun 1997 08:39:30 -0500 (CDT) Received: from master.debian.org (primer.i-Connect.Net [206.190.143.13]) by elvis.sw.ods.com (Netscape Mail Server v2.01) with SMTP id AAA8466 for [EMAIL PROTECTED]; Wed, 18 Jun 1997 08:39:28 -0500 Received: (qmail 10072 invoked by uid 834); 18 Jun 1997 13:39:30 - Delivered-To: [EMAIL PROTECTED] Received: (qmail 10070 invoked from network); 18 Jun 1997 13:39:29 - Received: from ctc.ctctel.com (206.30.29.2) by master.debian.org with SMTP; 18 Jun 1997 13:39:29 - Received: from coteau.ctctel.com (wr-32.dialup.ctctel.com [206.30.31.162]) by ctc.ctctel.com (8.8.3/8.8.3) with SMTP id HAA27971 for [EMAIL PROTECTED]; Wed, 18 Jun 1997 07:46:43 -0600 (MDT) Message-ID: [EMAIL PROTECTED] Date: Wed, 18 Jun 1997 07:44:10 -0600 From: coteau [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] X-Mailer: Mozilla 3.01 (Win95; I) MIME-Version: 1.0 To: [EMAIL PROTECTED] Subject: Network Problem Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi there, We have successfully loaded Linux (Debian 1.3) and have the xdm working properly. We are using fvwm95 as the window manager. The problem we are having is, when we try to telnet or ping to our other unix and linux machines, the error message came nerwork is unreachable. Other machines are running (SGI)Irix 5.3 and Debian 1.1 . We checked the cables and all the necessary files, as mention in the manuals. These appears to be in order. When we type the ps -ef command, no daemons appear to be running. If you have any suggestion for us. Please e-mail us, we'll be thankful. Sincerely, Riaz Khan -End of forwarded message- -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Bo has been Frozen -- Beta Test
On Apr 5, Brian C. White wrote Bo has now been officially frozen! If you'd like to start upgrading Great! However, unstable still points to bo meaning that anything put in unstable will also magically appear in frozen. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081
Re: code-names
On Apr 3, Douglas L Stewart wrote On Thu, 3 Apr 1997, Jason Gunthorpe wrote: Can we use glibc NOW along with the normal c lib? I have a program that will require the use of pthreads and it might be a good idea to use a thread safe C library, I can likely get by without, but.. libc 6 hasn't been officially released yet and it'll be a little while before any distribution includes it as a stable package. I'd suggest just No, libc6 has officially been released, and aside from some minor utmp annoyances, it is quite table. The only reason libc6 is in the experimental section rather than the unstable (i.e. development) tree is because I didn't want it's presence to confuse anyone and slow down the release of Debian 1.3. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081
Re: tk4.2: had to ln -s ld-linux.so.1 ld-linux.so.2
On Mar 27, Rick Macdonald wrote I had to link ld-linux.so.1 to ld-linux.so.2 in order to run tcl7.6/tk4.2. Get tcl76_7.6p2-2 and tk42_4.2p2-2. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081
Re: wish4.2 - the disappearing file
On Mar 25, Jean Pierre LeJacq wrote I'm not sure if this has been reported as a bug yet. Yes, it has, and it's already been fixed. wish4.2 has been linked against /lib/ld-linux.so.2 which hasn't been released yet as a debian package. You can see this by using strings on the wish4.2 executable. ld-linux.so.2 is the dynamic linker for glibc which is available in the experimental section. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081
Re: glibc in 1.3?
On Mar 26, Francesco Tapparo wrote I'm wondering if glibc 2.0 (i.e. libc.so.6) will be in debian 1.3. This release of libc is already out, but I not see it in debian. It's in the experimental section. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081
Re: problems with the new libc6 packages
On Mar 15, Joseph Skinner wrote I've just installed the libc6 packages and find that this seems to break gcc which sort of makes it useless. The error generates is that crtBegin.o is missing. You need to use gcc-2.7.2.2. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081
Re: Unstable vs. Stable
On Feb 20, Rick Macdonald wrote I live off unstable all the time, and seem to have less problems than the average bear! Be forewarned, after the release of Debian 1.3.x, we will be switching to glibc 2.x, aka Linux libc 6. When that happens, the unstable tree will definitely live up to it's name, at least for a while. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Q: ldso_1.8.10-1_i386.deb
On Feb 20, David Puryear wrote When I do dpkg -i ldso_1.8.10-1_i386.deb here is what I get: (Reading database ... 18262 files and directories currently installed.) Preparing to replace ldso 1.8.5-1 (using ldso_1.8.10-1_i386.deb) ... Unpacking replacement ldso ... Setting up ldso (1.8.10-1) ... ldconfig: warning: can't open /usr/lib/libc.so (No such file or directory), skipping ... I didn't have these warnings when I installed ldso_1.8.5-1_i386.deb Is this ldso problem or is it other packages not being upto-date? You probably have mismatched run-time and development packages. For example, your libc package might be at version 5.4.23 while your libc-dev package is still at version 5.4.20-1. The version numbers are supposed to match exactly. Dpkg is supposed to prevent this from happening but doesn't always seem to do it. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: ld-linux.so refuses to link the libraries I need...
On Feb 19, Carlos Carvalho wrote I've put some libraries in /usr/local/mydir, and made symlinks to /usr/local/lib. However, ld-linux.so doesn't link them. Of course /usr/local/lib is in /etc/ld.so.conf. Even if I put /usr/local/mydir in /etc/ld.so.conf the program doesn't run. ldconfig -v shows nothing in /usr/local/lib. Why doesn't it use the symlinks? LD_LIBRARY_PATH works but I'd like to not depend on it... It depends on which version of ldconfig you are using and how you setup the symlinks. Give me some specifics and I'll tell you what you are doing wrong. David -- David EngelODS Networks [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: More diald problems.
On Sat, 11 Jan 1997 [EMAIL PROTECTED] wrote: On Sat, 11 Jan 1997, David Engel wrote: I have a nagging problem with 'defaultroute' that maybe you can help me with. Everytime diald drops the link due to inactivity, it deletes the default route. After that, diald won't bring the link back up for non-loopback addresses because there aren't any routes. I have to manually force the link back up. Any ideas? It's a bug in diald that has been fixed with the newest release of diald. However, the newest release hasn't been debianized yet by the maintainer (I don't think). I am going to do it myself if a new diald.deb doesn't show up in unstable soon. If you want a copy, email me. I didn't know there was a new version. Thanks. Diald now restores the routes and interfaces to the state they were before the link was brought up. However, diald still won't bring the link back up automatically after it has taken it down. The dummy SLIP interface acts likes it is passing packets (the TX counter increments), but diald never seems to notice and reactivate the real SLIP interface. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: More diald problems. (fwd)
On 12 Jan 1997, James LewisMoss wrote: Go to the diald web site and get the patch. It has a bug that once it takes ot down once it won't bring it back up. The patch fixes it. (Is running fine on my machine) The new diald site is unreachable right now. Would you mind mailing the patch. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: More diald problems.
On 10 Jan 1997, Guy Maor wrote: Kevin Traas [EMAIL PROTECTED] writes: 2. From the docs, I thought diald would establish the default route to the gateway automatically. Am I wrong in my assumption? If so, is what I did to fix the problem the right way to go about it? The right way is to add `defaultroute' to /etc/ppp/options. I have a nagging problem with 'defaultroute' that maybe you can help me with. Everytime diald drops the link due to inactivity, it deletes the default route. After that, diald won't bring the link back up for non-loopback addresses because there aren't any routes. I have to manually force the link back up. Any ideas? David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Grrr! Dangling links in /usr/lib/
David Lutz writes: I am upgrading my mostly Debian 1.1 system to Debian 1.2. According to symlinks I have these dangling links: dangling: /usr/lib/libbfd.so.2.7.0.3 - libbfd.so.2.7.0.3.dpkg-tmp dangling: /usr/lib/libopcodes.so.2.7.0.3 - libopcodes.so.2.7.0.3.dpkg-tmp This is a hold over from a problem between dpkg and ldconfig. Just reinstall the binutils package and you should be fine. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Some notes on Debian experiences
Alan Eugene Davis writes: I wanted to compile a new kernel. Remembering that Debian has a screwy header file setup, ^ You forgot a , IMO, here. I had to relink as described in the /usr/src/linux/README. For the umpteenth time. You don't need these links to compile the kernel. I haven't had those links on any of my systems in over a year and it hasn't stopped me from compiling hundreds of kernels. BTW, libc6 will use a different approach. I need to double check, but I think it sucks all of the information it needs from the kernel headers at build time. You won't need any kernel headers in /usr/include, symlinked or otherwise, to compile programs with it. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: gdb and dynamically loaded code
H.J. Lu writes: Douglas Bates writes: I have version 4.16-2 of the gdb package installed. When I start gdb I get the warnings warning: Unable to find dynamic linker breakpoint function. warning: GDB will be unable to debug shared library initializers warning: and track explicitly loaded dynamic code. It would be very helpful to me to be able to set breakpoints within dynamically loaded code. Is there something I can change in the set-up for gdb to allow me to track explicitly loaded dynamic code? Apparently, gdb can not handle a stripped ld-linux.so.1. The symbols gdb wants can be found in the .dynsym section so this could be considered a bug in gdb but I'm not sure. H.J., what do you think? How do you debug a stripped binary? Unless I am wrong, if you want to debug ld-linux.so.1, you have to compile it with -g and don't strip it. It doesn't cost much in disk space. H.J., I think you misunderstood. We aren't debugging ld-linux.so.1. We are debugging an ordinary binary. When gdb starts a binary, it looks up the address of _dl_debug_state so it can set a breakpoint in the dynamic linker to trap the loading of shared libraries. With a stripped ld-linux.so.1, gdb fails to find the address of _dl_debug_state and issues the warning shown above. However, _dl_debug_state is contained in the .dynsym section (which doesn't get stripped) of ld-linux.so.1, so I'm thinking that it should be possible for gdb to get the address from there even when ld-linux.so.1 is stripped. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- This message was delayed because the list mail delivery agent was down.
Re: libc-5.4.13
Karl M. Hegbloom writes: Will Debian 1.2 have libc-5.4.13? It's on sunsite; says it's a bugfix. Unless Bruce vetoes it, yes. I uploaded it last night. It should show up in the next few days. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: bo is full of symlinks to rex?
Rick Macdonald writes: When the change occured, my mirror of unstable lost a bunch of files. Now I have to get both rex and bo. I think I can continually mirror unstable only, by careful use of the following option in mirror, but I haven't actually tried this yet: [description of flags_[no]recursive+L option deleted] Anybody got a better method to mirror only the unstable tree efficiently? This is exactly what I do to mirror just unstable. It works fine. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: GnuStep Re: X is painful
Bill Bumgarner writes: Now that I have an answer on the libc front (ie; 1.2 will ship with = 5.4.7), I'm going to reboot to Linux and produce 'official' 5.4.7 = compliant GCC 2.7.2.1 + GnuStep Threading Patches in short order. =20 BTW: For now, I'm using the built in MIT_POSIX_THREADS. Who is = maintaining LinuxThreads-- it seems like a superior package and I would = like to move to it as soon as possible... BUT: If I move to using LinuxThreads, gcc will DEPEND on the LinuxThreads = package. Does that offend anyone? This makes me nervous. What exactly are you proposing to do to gcc? Why would it have to depend on any thread package? I sent a message the other day, but it bounced because the lists were down. IMO, we should stick with MIT pthreads as included in libc5. It has been available for quite a while and is reported to work, though I don't know of any Debian packages using it. We can switch to LinuxThreads when we go to libc6. FYI, LinuxThreads is already available as a compile-time add-on for libc6, so I would expect it to be well supported. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- This message was distributed manually by [EMAIL PROTECTED] after the list initially failed to distribute it.
Re: g77 failure
Hamish Moffatt writes: with it enabled. Ideally, the GCC package will have gcc with this enabled, and the GNAT package will not supply a gcc at all. Hopefully this will happen for the next GNAT package version, which is awaiting GNAT 3.06 (based on gcc 2.7.2.1). When I was maintaining gcc, I asked multiple times for the maintainer of GNAT to send me the necessary files, but I never heard from him. It only takes 2 or 3 files being in the right place and gcc's configure script will automatically include Ada support. Fortran support is already done in this way. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- This message was distributed manually by [EMAIL PROTECTED] after the list initially failed to distribute it.
Re: ldconfig warning : inconsistent soname??
Martin Konold writes: Similar problems here with: ldconfig: warning: /usr/X11R6/lib/libncurses.so.1.9.9e has inconsistent soname ( libncurses.so.3.0), skipping What in the worlkd is ncurses doing in /usr/X11R6/lib? It's supposed to be in /lib. Please install the Debian version which is named correctly. ldconfig: warning: /usr/lib/libhistory.so.2.0.1 has inconsistent soname (libread line.so.2.0), skipping This is a known bug in the readline package. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Dynamic Linking
GREENE KENNETH ADAM writes: I'm trying to get GCC to do dynamic linking (I'm new to Linux, was using DJGPP on a M$-DOG system). I can get it to do static linking, but not dynamic. when I do gcc -dynamic -o test test.o it generates a static linked ELF binary (according ot ldd). Some of your lib*.so symlinks might have been deleted accidentally. Reinstall the appropriate *-dev packages and see if that fixes it. If this is not the appropriate place for this message, my humblest apologies. Yes, this is the right place. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Anyone tried nm on libc5
Paul Chau writes: I tried running nm on the libc5 and it said that no symbols found. Can anyone explain to me why? You have to use -D to see the dynamic symbols. All of the regular symbols were stripped out. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: JDK
Evan Thomas writes: kazem wrote: Is anybody can help describe me how i can install JDK in my box. Or if there is a package for it. As far as I know there is no Debian package. You can download the Linux port from http://www.blackdown.org/java-linux.html. The site points to installation instructions. It installs in the /usr/local tree so it won't conflict with any Debian stuff. JDK v1.0.1 has been available for some time. Search for jdk*. I mailed the maintiner about upgrading it to v1.0.2, but I never got a reply. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Netscape 3.01 libc5_5.4.7 crashes
Guy Maor writes: LD_PRELOAD=/usr/lib/libgnumalloc.so netscape That should be LD_PRELOAD=libgnumalloc.so.5 ... David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: nm /lib/libc.so
Bernd Eckenfels writes: some softwre PAckages run nm on libc to get the available functions. This wont work with ELF anymore. nm only prints symbols for .a files. Arent the symbols in .so needed for linking? You have to use 'nm -D'. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: does libc5.4.7 compatible with netscape 3.0
Randy Gobbel writes: Yes, I would say it's a bug in the Debian version of libc 5.4.7. Some functions have been moved from the C library to the compiler, and applications compiled with older versions of gcc need help to work with libc 5.4.7. I stumbled across a fix for the problem at Are you sure about this? I tried H.J. Lu's precompiled version and it causes Netscape to bus error also. Someone else suggested that it was a Java-related problem in netscape. I haven't had any problems since I disabled Java support. http://nightflight.com/~pcg/medkit.html This page is empty as far as I can tell. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Is this acceptable?
Tristan McCann writes: I was just wondering if this was acceptable: binaries created with Libc5.4.7 are not compatable with libc5.2.18 I was wondering if it was acceptable if I make debian packages to put in the re= quires box that it requires libc5.4.7 in the unstable section. This is OK as long as the packages you are building also go in the unstable tree. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Problems building Python 1.4
Paul Barrett writes: I'm trying to build Python 1.4 on my Debian box to test the alpha release of the Numerical module. The 'configure' script fails when I include the option '--with-readline' because it cannot find the termcap library. Is this library included in the 'libreadline' library? Readline appears to depend on it and it is nowhere to be found. If this is the case I'm sure Guido would like to know about it, so he can modify the Python configure script. You need to link with ncurses instead of termcap. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: does libc5.4.7 compatible with netscape 3.0
Lawrence Chim writes: Once I upgrade libc to 5.4.7-1, netscape cannot run and it said bus error. Then I downgrade to 5.2.18, netscape runs again. Lawrence reported this to me earlier. I had not seen any problems at that time but am now. Is anyone else having problems with libc 5.4.7 and/or know of the fix? David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: gcc 2.7.2.1 + objc support-- debianized package available
Bill Bumgarner writes: I have made gcc 2.7.2.1 with objc multithreading support (snapshot 960906) = available via: ftp://ftp.thoughtport.net/pub/debian/gcc_2.7.2.1-2_i386.deb ftp://ftp.thoughtport.net/pub/debian/ It is a debianized package; ie-- it is designed to be installed under = Debian linux. The '-2' is simply because the source package I downloaded = off of one of the debian source mirrors was '-1'; this is an incremental = improvement over '-1', so it is '-2'...=20 Unless you are going to take over the gcc package, please you a different numbering scheme. There already is a gcc 2.7.2.1-2 package in the Incoming directory waiting to be installed. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: libc5-dev upgrades
Arrigo Triulzi writes: I know that it is intended that kenel header files come with libc5-dev and are not symlinks to the kernel source tree. Ok, this said let us come to my humble request: I would dearly like to use the symlink to my current kernel source on all boxes, this is because of some driver work people are doing. Whenever I upgrade the symlinks zap the /usr/src/linux/include/linux directory and I have to reinstall the include files by hand... Would it be possible to have an option preventing dpkg from following symlinks (or is it already there) so that I can save my kernel setup from zapping when I upgrade libc5-dev? Please read /usr/doc/libc5/FAQ.gz. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1001 E. Arapaho Road (972) 234-6400 Richardson, TX 75081 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: g77 and gcc-2.7.2.1
Hamish Moffatt writes: I wrote: Similar problem (I think) with GNAT Ada. GNAT 3.04 (and hence the Debian package, and also GNAT 3.05 which is not available as a package) is based on gcc 2.7.2. Until GNAT is updated to be based on gcc 2.7.2.1 and then a pkg of the same released, you have to stick with gcc 2.7.2 ... Gcc 2.7.2.1 is a bug fix only release. It should be safe to bump the version number of GNAT to 2.7.2.1 and re-release it. David writes: Gcc 2.7.2.1 was a bug-fix only release so it should be safe to just change the version number for GNAT to match it. But GNAT diverts the GCC binary, and the GCC binary needs to be GNAT should not do this. You should send me the appropriate files to make the standard gcc driver support Ada. Look at the current Debian source to see how fortran is already handled. recompiled to recognise Ada code, or the GCC binary from GNAT won't work properly because it's 2.7.2 not 2.7.2.1, etc. It might be possible to hack it together but not worth it. As soon as GNAT is updated and then the Debian package, I'll upgrade GCC as well ... David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Linux binary? (Tkined)
Juergen Schoenwaelder writes: [EMAIL PROTECTED] (David Engel) said: DavidAside to Juergen. What do you think of Dr. Ousterhout's proposed Davidimprovement to package loading? I will probably add it, or something Davidsimilar, to the Debian version of Tcl 7.5 before packaging Scotty David2.1.2. Well, you know that I am not 100 % happy with John's solution. Anyway, scotty will use whatever mechanism will appear in Tcl 7.6 to pickup packages automatically. (I still hope that John modifies its scheme a bit.) What don't you like about it? I haven't really looked into the problem yet, so I'm not fully aware of the advantages/disadvantages of his proposal. I don't think that the patches proposed on the comp.lang.tcl list create nasty side effects - you need a complicated setup with some naming conflicts to make this happen. So it should be save to include one or both of these patches in the Debian version if you are prepared to change this scheme again when Tcl7.6 comes out. I would be prepared to change it. If I rebuilt scotty to use Tcl7.6 instead of Tcl7.5, I would be changing it anyways. As things stand now, I don't think either solution (modifying files in an installed TCL package or requiring users to set TCLLIBPATH) is acceptible for Debian. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Linux binary? (Tkined)
Juergen Schoenwaelder writes: DavidWhat don't you like about it? I haven't really looked into the Davidproblem yet, so I'm not fully aware of the advantages/disadvantages of Davidhis proposal. John proposed to change the way the auto_path is used. He did not define an additional directory that is included in the auto_path and independent of the Tcl version. So every extension is automatically bound to the Tcl version - something I dislike. In an ideal world, I can see the advantage of making it independent of the Tcl version. However, since each new version of Tcl is usually incompatible at the binary level (and often even at the source level), I don't see this as a big problem. For example, it would be very unlikely that a libscotty.so built against Tcl7.5 would work with Tcl7.6. To make it work with Tcl7.6, you would have to rebuild scotty anyway. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Linux binary? (Tkined)
Mark Purcell writes: My Debian versions of 2.0.2/1.3.4 are available on ftp.debian.org. I have a huge backlog of things I am trying to work off and hope to get to version 2.1.1 in the next week or two. Please do put together the debian 2.1.1 package of scotty/tkined! I I'm still planning on packaging it, but I can't make any promises as to when. It hopefully won't be too long. Aside to Juergen. What do you think of Dr. Ousterhout's proposed improvement to package loading? I will probably add it, or something similar, to the Debian version of Tcl 7.5 before packaging Scotty 2.1.2. Thanks for Tkined and Debian. You're welcome, at least for Debian. Tkined is Juerhen's work. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: DANGER: installing ld.so libc5 with dftp
Brian C. White writes: No. This was a recent upgarde of a few packages in unstable on a right-up-to-date system. As I understand it, what happened was approximately this... dpkg --install .../ldso* .../libc5* .../other packages the results were unpack ldso unpack libc5(error perl: cannot find libdl.so.1) unpack others (error perl: cannot find libdl.so.1 occasionally) This is due to a problem in the ld.so packaging. It will be fixed in the next version. Until it's available, you should install ld.so by itself. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: tk-blt package released
Gordon Russell writes: I decided to call the package tk-blt, as there is another public domain package called blt which does something completely different! I did not make two packages for this one, as the difference between a user and a developer version would have been less than a KByte of disk space. Even though you've withdrawn your package, there is one thing that needs clarification. The reason for providing separate run-time and development packages is not to save disk space. The reason is so that multiple, possibly incompatible (at the API/ABI level), versions of a package can be installed simultaneously. Take a look at how this is done in the tcl74 and tcl75 packages. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: libg++-2.7.2.deb is old?
Bruce Perens writes: The libg++ doesn't compile with the ELF libc. That's why it's in that state. I'm not sure what's happening with this - whether there is a maintainer working on it or if we're just going to wait for GNU LIBC 6. What do you mean? libg++27-2.7.1-2 builds fine for me with libc5-5.2.18-10. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: elf-x11r6lib
Simon Martin writes: Whilst trying to install tk40 I get a dependency error on elf-x11r6lib. I've checked the Debian Web and ftp sites and haven't been able to find it. Can anyone tell me where to access this file? 'elf-x11r6lib' is provided by the xlib package. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: g77 and gcc-2.7.2.1
H. Manz writes: After this I have detected, that the Fortran Compiler g77 is not working longer since it was based on 2.7.2. I have than copied all the stuff from the 2.7.2 Version subdirectory to 2.7.2.1. Now the compiler was working but only half. Doing that could lead to dpkg/packaging inconsistencies in the future. Compiling you fortran files with 'gcc -c -V 2.7.2' should work. As son as I try to link an absolute module it crashes. I think this is because of some internal incompatibility between 2.7.2 and 2.7.2.1. I don't know. Can you make a small test case? David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Cross posting per request
On Wed, 4 Sep 1996, Philippe Troin wrote: On Wed, 04 Sep 1996 11:04:06 CDT David Engel ([EMAIL PROTECTED]) wrote: There is a curses-based version of Tk, but I don't have any idea of how well it works. Do you know where we could find this thing ? Might be interesting... It's at ftp.neosoft.com:/languages/tcl/alcatel/potpourri/ctk-4.0.tar.gz. I already took a look at it and, although it works, it's barely usable, IMO. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Cross posting per request
Bruce Perens writes: It's unfortunate that the printer config stuff (and other stuff from Red Hat) is written in TCL/TK. One thing we _don't_ assume is that the user has X (or even a VGA card - it might be a serial console). A shell/dialog solution would be much better. There is a curses-based version of Tk, but I don't have any idea of how well it works. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: need libbsd.so.1.0.0
joost witteveen writes: The current implementation of postgres95 for Debian requires libbsd.so.1.0.0, but the libc5 package includes only libbsd.a. (How) can I make libbsd.so.1.0.0 from the .a file? (Sorry to be asking what I imagine is such a naive question.) I have no idea where the libbsd.so.1.0.0 came from. But if you've got the .a file, cannot you make the .so file by: ... Guy already answered the part about building a shared library. (I may well be _very_ wrong, and please flame me if I am, but I just _think_ it would work, and it's easy to find out). My point was that somebody built a package that uses a shared library that we don't even provide. IMO, finding out, what it is, where it came from and why it was used is more important than making a shared version of the libbsd.a provided by libc. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: need libbsd.so.1.0.0
On Mon, 2 Sep 1996, Susan G. Kleinmann wrote: The current implementation of postgres95 for Debian requires libbsd.so.1.0.0, but the libc5 package includes only libbsd.a. (How) can I make libbsd.so.1.0.0 from the .a file? (Sorry to be asking what I imagine is such a naive question.) I have no idea where the libbsd.so.1.0.0 came from. David --- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: what is -ltk looking for?
Heiko Schlittermann writes: Since the libtk versions are incompatible (at API level, as far as I can guess -- I've tried (now successfully) to adopt source written for This is correct and is also why the author started putting the version number in the library file name. tk7.4 (?) and for every new libtk (7.4, 7.4) I encountered slight changes at API level) you should consider change your makefiles to the proper library (-ltk74 rsp. -ltk75). This is what should be done, except your examples should actually -ltcl7.4 and -ltcl7.5. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: clock -w writes bogus date
Bill Wohler writes: Package: util-linux Version: 2.5-5 clock -w is screwing up the CMOS clock. See: ... Is there anything here I'm missing (e.g., related programs/packages/settings)? The segmentation violation and hanging of clock -r is worth a bug report to be sure, but could the -w failure be a mis-configuration of the timezone or Universal time on my part? I had answered no to the Universal time question at install (due to [EMAIL PROTECTED] DOS). I just ran tzconfig to change /etc/timezone from Factory to US/Pacific thinking that was my problem. This has already been reported in Bug#4265. Here is a quick patch to clock.c in util-linux that fixed it for me: David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081 --- clock.c.origMon Aug 26 10:26:53 1996 +++ clock.c Mon Aug 26 10:31:38 1996 @@ -365,7 +365,7 @@ } /* Read the RTC time/date */ -retval = ioctl(cmos_fd, RTC_RD_TIME, tm); +retval = ioctl(cmos_fd, RTC_RD_TIME, tm); if (retval == -1) { perror(ioctl); exit(errno); @@ -463,7 +463,7 @@ int retval; -retval = ioctl(cmos_fd, RTC_SET_TIME, tmp); +retval = ioctl(cmos_fd, RTC_SET_TIME, tm); if (retval == -1) { perror(ioctl); exit(errno);
Re: gcc can't find termcap library
Bruce Perens writes: From: Larry 'Daffy' Daffner [EMAIL PROTECTED] This is not necessarily always possible. The one case I know about is when a program uses varargs instead of the newere, ANSI-standard stdarg. There is no varargs.h in libc5-dev_5.2.18-9 . Maybe older versions contain it. varargs.h is provided by gcc. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: gcc can't find termcap library
[EMAIL PROTECTED] writes: I recently installed Debian 1.1, including termcap.compat. Since then I've installed a custom version of the kernel (2.0.7) successfully, so I figure that gcc is reasonably well set up. However, I tried to compile kermit 5A(190) and the make failed, complaining that it couldn't find -ltermcap. The Debian termcap package is only intended for run-time use to allow old and non-Debian programs compiled with termcap to work. Any new programs should be compiled with ncurses. There is definitely a file of the form /lib/libtermcap.so.?, but I guess gcc is looking for a *.a file. How should I go about fixing this? You should convert the source and/or Makefile to use ncurses. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: ldso package and perl
Douglas Bates writes: When using dselect to install several upgraded packages today I suddenly started receiving error messages from perl. Here is part of the transcript. This due to a packaging problem in ldso. It will be fixed in a few days. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: where is wish ?
Hubert FAUQUE writes: I have tried make xconfig to generate a new config file for the kernel but it gives an error wish not found Could somebody tell me what is wish and in which package it is? Wish is the base-level Tk interpreter. Install either the tk40 or tk41 packages. if you install both, the wish from tk41 will be used by default. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: GCC frepo patch don't work.
I don't remember anyone responding to this. Jimen Ching writes: I've just installed debian linux 1.1 from Iconnect. I am working on a C++ project which depends heavily on the use of templates. Thus, I would like to use the -frepo patch from Cygnus. The installed gcc/g++ doesn't seem to have been patched correctly. At least the 'ld' is not working as described by Mumit Khan. I.e. the patched 'ld' was supposed to do a recompile/relink process. This was never done and I get lots of unresolved symbols. Mumit Khan sent me some binaries about a year ago for the Slackware distribution I used back then. I tried moving the files over to my current system, and things seem to work ok. But I am getting another error message from the patched 'ld'. This time, it says it can't find one of my header files. I.e. the header file which contained my class definition could not be found. This error is produced while in the recompile/relink phase (using Mumit's binaries). Has anyone ever tried the -frepo feature? It is actually patched in? Using the switch to g++ doesn't tell you anything since g++ ignores arguments it doesn't understand. This feature is very important. I can't continue my project without it. Please help, thanks in advance... No, Debian's gcc does not include the -frepo patch. Unless their is heavy demand for it by Debian users or H.J. Lu includes it in his general Linux versions, I have no plans to add it. Until that happens you'll have to get the Debian source and patch it yourself. It isn't that hard. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: No Distribution is 2.0.0 Current
Jeffery S. Coy Jr. writes: On Sat, 29 Jun 1996, David Engel wrote: Debian's version of libc 5.2.18 has the security fix plus a few other fixes. I really wish H.J. would continue to fix serious bugs in the last stable libc while the new one is still in alpha/beta testing. i thought libc 5.3.12 had been publicly released as stable some time ago. is this library being ignored due to doug lea's malloc, or is it something else? To clarify, I wasn't implying that libc 5.2.18 was the last stable release. I know that 5.3.12 was released some time ago. Debian isn't using it because we didn't want risk introducing any instabilities when we were trying to releas Debian 1.1. i think we are all a little more wary after libc 5.3.9's .rhost bug. but i agree, it would be nice if patches for important bugs were made available for proven libraries such as lib 5.2.18, rather than push the bleeding edge stuff so hard. Right. I've been saying for a long time now that Linux' huge growth makes stability a lot more important than it used to be. When an important problem is found in libc, we can't keep telling people to upgrade to the current beta version. It may fix the original problem but will several others. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: aout svgalib problem
Richard Kettlewell writes: Jeffery S. Coy, Jr. writes: i just installed the aout-svgalib-1.28-6.deb package, and noticed it installs to /usr/i486-linuxaout rather than /usr/lib/i486-linuxaout, so the system can't find it. No, /usr/i486-linuxaout/lib is the correct place; I'd expect the problem to be something else. I don't really have time to maintain the svgalib packages any more and it would be good if someone else would take over. This is partly my fault. In the last libc4 package, I moved libdb.so* from /usr/i486-linuxaout/lib to /usr/lib/i486-linuxaout. I did this so users who didn't install any a.out development packages (eg. libc4-dev, aout-gcc, etc.) wouldn't have a /usr/i486-linuxaout directory cluttering up /usr. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Problems doing compiles.
Evan Harris writes: reference to `_ctype' messages on the link step, although my hello, world test program will compile and link (but nothing else will). Please send the output from 'dpkg -l' and from compiling your program with 'gcc -v'. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Hard versus extended partitions
Bill Wohler writes: Is there a performance hit for using extended partitions, or should one be unconcerned about creating them willy-nilly? I don't know of any performance hits regarding extended partitions. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: pthreads and libc version for debian 1.1
Guy Maor writes: The latest pthread library is 1.60 beta4, released on 10/25/95. Compiled seperately it works moderately well. It's a lot of fun to play with. You can get more info at http://www.mit.edu:8001/people/proven/pthreads.html Actually, I saw two newer versions on the MIT ftp site, beta4_1 and beta5. Presumably libc 5.3.12 contains the newer version, so when 5.3 stabilizes and David upgrades, we'll have it. Until then, you might Both libc 5.3.12 and 5.4.1 include pthreads 1.60beta4. I will release libc 5.3.12 or 5.4.x (if it's released) after Debian 1.1 is released. want to get the pthread source from the above site and compile it seperately. It compiles out of the box under Linux. I am not familiar with pthreads at all, so I'm not comfortable with deviating from H.J. Lu's version. Unless someone steps forward to package pthreads separately, here is what I propose to do. I will be updating the kernel headers in libc 5.2.18-x to version 2.0.0 whenever Linus releases it. When I do that, I will include H.J.'s pthreads in the libc5 package. I will also have it provide a virtual package called pthreads1. Packages with threaded binaries should depend on pthreads1. This will make it easier to split out pthreads later if we want to. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: pthreads and libc version for debian 1.1
Michael Callahan writes: I would like to play with pthreads, and was wondering what I must do to get phtreads support built into my libc. Do I have to compile the latest version of libc myself, with pthreads specified somewhere? Or is it in the latest version already? Here is the relevant section from H.J. Lu's release notes: PTHREAD NOTES: The MIT pthread library 1.60 beta2 seems to work. It passed most of tests in the MIT pthread package. The pthread libraries are not installed by default, nor are binaries included. You may want to remove -g from the default CFLAGS for the pthread library. You need to install the libraries elfshared/libpthread.so.1.60.beta2 and elfstatic/libpthread.a in either /usr/local/lib or /usr/lib manually. You must run ldconfig after installing them. To compile pthread programs, you need to add -D_MIT_POSIX_THREADS to CFLAGS and -lpthread to LDFLAGS. The network _r functions are still missing. Due to the implementation of the pthread library, pthread only works with the shared libpthread. Would someone like to test the pthread library that is built? You'll need to get the libc5 source and recompile the entire package yourself. If libpthread proves to work, I'll include it in either the next libc5 package or a separate libpthread package. Also, why does debian 1.1 use libc-5.2.18 when libc-5.3.12 is available? Is it for stability reasons? For stability. Libc 5.3.12 was brand new when we first started preparing to release Debian 1.1. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: re-upload kernel-package-1.01
Manoj Srivastava writes: This _is_ standard practice for debian (including /usr/include/linux and /usr/include/asm in libc5-dev). This has been discussed at length and the decision reached after a long, drawn out debate. This is on it;s way to becoming a faq: Already started. There is now a /usr/doc/libc5/FAQ.gz file in the latest (5.2.18-8) libc5 packages. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: kernel headers
Manoj Srivastava [EMAIL PROTECTED] writes: The headers were included in libc5-dev after a rash of very buggy alpha kernel releases (1.3.7* or something like that) that proceeded to break compilations, etc. Kernel versions are changed far more rapidly than libc is, and there are higer chances that people install a custom kernel than they install custom libc. I strongly ask the ppl to think about this. This is a MAJOR difference from debian to the rest of the linux community. It even violates the recommedations from Linus himself (in the Kernels readme). Broken Headers due to instable Kernelinstalls are realy not a concern we should have. The System will just run fine, and those who install a new kernel and then compile a program are experienced enough to know how to get new version of the kernel if it wont work. I mean every single compiler run in the linux community is done with the actual kernelsources, why should we change this and act like we dont belong to linux? Programs which needs the linux/ and asm/ include-files have a reason to use them. A lot of programs compile dependingly of the kernelsource (for example the net-tools). Greetings Bernd This has already been debated enough. Debian will continue to include known-working kernel headers with libc unless and until that arrangement proves to be unworkable. As I have time, I will continue to encourage H.J. Lu and other Linux distributors to do the same. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: kernel headers
disagree on this issue. I still don't feel it is right to put kernel headers anywhere except with the kernel (or perhaps as their own package). If people So just think of them as libc headers instead of kernel headers. That's really how they are being used when referenced as /usr/include/*. things to break. Debian should concentrate on providing a complete, stable system. This is why the change was made. The new arrangement is more stable. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: linux include files arrangement
Does debian 1.1 observe the arrangement of the Linux kernel with regards to include files etc? ie should there be symlinks /usr/include to include and asm subdirectories in /usr/src/linux? No, Debian uses a more stable (IMHO) arrangement where libc provides its own version of the kernel headers that are known to be compatible. I previously only installed the kernel-header package from debian. But now I have just rebuilt the kernel from kernel-source-1.3.100-0.all.deb (because of module problems). Should I remove the /usr/include/linux files and make a symlink to the source directory? No, don't replace the headers in /usr/include/linux. The kernel does not use the headers in /usr/include. If you ever need to compile something with kernel headers that are newer than those procided by libc (which isn't likely unless you are doing systems level work), you should add -I/usr/src/linux/include to your compile line. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: tk tcl and dependecies
I'm copying this to debian-user since others may be confused as well. I found that installing tk41 in stead of tk40 will break several dependencies. Packages exist that depend on tk40 and won't be installed easily with tk41. Another thing is that those packages are compiled with links to /usr/lib/libtk4.0.so and can't work with /usr/lib/libtk4.1.so. (Because of this the dependencies are ok) That's the way it was intended. But if they had been compiled to /usr/lib/libtk4.so or even /usr/lib/libtk.so this problem wouldn't occur. Should tk40 and or tk41 provide a package named tk? and should these packages provide the mentioned library names? Tcl7.4/Tk4.0 and Tcl7.5/Tk4.1 are NOT fully compatible at the C level. Binaries built with Tcl7.4/Tk4.0 are not compatible with Tcl7.5/Tk4.1 and vice versa. To cope with this, the run-time packages are completely independent. You can have all of tcl74, tk40, tcl75 and tk41 installed if you like. Of course, you still only have one set of development packages installed at a time. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: unstable Packages file
I notice the Packages file for the unstable tree at ftp.debian.org is out of date. I though it was updated automatically? Perhaps not... We know about the problem and are trying to fix it. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: diald (Was: Re: Must pppd be run by root?)
The diald package removes all the hassle of starting and shutting down network connections over a transient link, incidentally - you might like to investuigate that. Is there a debianized version of diald 0.14 out there? (ELF). With the dctrl tk tool, too! (I can email it to the one making the package if needed). The debian package hasn't been upgraded to 0.14 yet. Would someone please do this. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: problems compiling with R0.93
When I try to compile my own hello world program I get these error messages: gcc -g -o hello -L/usr/X11R6/lib hello.o -lXaw -lXt -lX11 ld: Output file requires shared library `libc.so.4' gcc: Internal compiler error: program ld got fatal signal 6 make: *** [hello] Error 1 Compiling the program with: gcc -g -c hello.c works fine, but the linking part just won't work. It's funny because I've successfully compiled the 'xli' package on my machine. And I have checked, I do have the libc.so.4 library in the /lib directory. The problem is that the shared X libraries need the shared libc but the old, a.out gcc links in a static debug version of libc called libg when -g is used. Either don't link with -g so the shared libc will be used or add -static so the shared X libraries won't be used. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: 1.1 upg libc - long
Over the last few days, I have upgraded a (work) machine from 0.93r6 to 1.1 beta using Dale Scheetz's manual dpkg instructions. All went smoothly except for libc4. I upgraded dpkg from aout to elf tried again with the results below. Any ideas ? thanks cm nb - After above, i patched the 1.2.13 kernel to elf, recompiled and rebooted ok... From memory, /lib contains the libc4 contents as expected. linux96:/home/debian# dpkg -i --force-conflicts unstable/devel/libc4-4.6.27-14.deb dpkg: considering removing libc in favour of libc4 ... libc is not properly installed - ignoring any dependencies on it. dpkg: yes, will remove libc in favour of libc4. (Reading database ... 13471 files and directories currently installed.) Unpacking libc4 (from unstable/devel/libc4-4.6.27-14.deb) ... ln: libc.so.4.6.27.save: No such file or directory ln: libm.so.4.6.27.save: No such file or directory ln: libcurses.so.0.1.2.save: No such file or directory ln: libdb.so.1.85.1.save: No such file or directory rm: linux: is a directory dpkg: error processing unstable/devel/libc4-4.6.27-14.deb (--install): subprocess post-removal script returned error exit status 1 Errors were encountered while processing: unstable/devel/libc4-4.6.27-14.deb linux96:/home/debian# Could you please retry this with dpkg 1.1.6. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: upgrading from 0.93R6 to 1.1 beta
Our archive maintainer has not created the updates directory yet. When that happens, the a.out version will be placed there and the ELF version will be moved to the binary/base directory. The 1.1.5aout version of dpkg is necessary for upgrading the base packages. Once all the base libraries have been upgraded, then upgrading to the elf version will be useful/possible. Because of this, I think both versions should be left in an easy to get at location (together would be better than seperate) The problem has something to do with dselect not coping well with packages being listed twice in the Packages file. IMO, the upgrades directory idea makes a lot of sense. Just tell users to upgrade to the packages in that directory and then run dselect on the packages in the binary directories. Very simple to understand. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Permissions
is there a problem with the disks-i386 on master? -r--r--r-- 1 802 800 1474560 Apr 28 06:39 1440_boot_floppy -r--r- 1 802 800 1396095 May 5 01:50 1440_boot_floppy.gz I can't get to master at the moment to check this out, but I suspect th newer version, 1440_boot_floppy.gz, is the one that should be used. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Permissions
is there a problem with the disks-i386 on master? -r--r--r-- 1 802 800 1474560 Apr 28 06:39 1440_boot_floppy -r--r- 1 802 800 1396095 May 5 01:50 1440_boot_floppy.gz As best I can tell, 1440_boot_floppy.gz is the one that should be used. I renamed 1440_boot_floppy to 1440_boot_floppy.old and added world read permission to 1440_boot_floppy.gz. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: upgrading from 0.93R6 to 1.1 beta
Scott I assume you mean the dpkg in unstable/binary/base, since I Scott couldn't find any upgrades directory. BTW, is there an ELF version Scott of dpkg? I only saw an a.out version. Yes, in debian/project/experimental Our archive maintainer has not created the updates directory yet. When that happens, the a.out version will be placed there and the ELF version will be moved to the binary/base directory. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: dselect
I think I've got all the files now to upgrade to 1.1, but before I start I have a couple of questions. When I initially installed the stable version I managed to get it done, but I stumbled a bit with dselect. Does dselect expect the Packages file to reflect what files are available, or should it left intact? You should go through the first two menu items in dselect, Access and Update. This will update dselect with the current list of packages. Also, please update dpkg manually to the 1.1.5 or 1.1.6 a.out version before running dselect. Many of the new packages need the new dpkg to be installed. When removing packages with dselect are the selected packages removed or retained? I had a bad experience with this. Dselect evidently worked differently than I thought it would, and took a large bite out of my system before I could get to the control C. Only packages you have marked for removal, by using either '-' or '_' in the select window, should be removed. All other packages should be left alone. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Bug#2807: X - cpp dependency?
Does gcc now have 'Provides: cpp'? Yes. I had to install cpp 2.7.2-6 (even though i already had gcc 2.7.2-6 installed!) a week or so ago because netstd depended on cpp. Installing cpp shouldn't be necessary if you already have gcc. maybe netstd should depend on 'cpp | gcc'? or maybe gcc should provide cpp. It only needs to depend on cpp. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Unidentified subject!
Is anyone out there using a linux box as an NIS client? Does it work? Yes and yes. ypbind used to die occasionally with older versions but doesn't anymore with the current version. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081
Re: Bug#2807: X - cpp dependency?
Carlos The problem is that the links /usr/bin/cpp and /lib/cpp don't Carlos come with gcc. If gcc really gives everything, it should give these Carlos symlinks as well. I think they now do in the very latest version, released yesterday. That's right, though there is one scenario due to a bug in previous versions where /lib/cpp may be deleted. It is fixed by reinstalling. I'm going to fix it in the next version. David -- David EngelOptical Data Systems, Inc. [EMAIL PROTECTED] 1101 E. Arapaho Road (214) 234-6400 Richardson, TX 75081