Re: where can i get the sources from??
On Wed, May 20, 1998 at 01:28:27PM -0600, [EMAIL PROTECTED] wrote: > > is the S/Linux proyect dead? where can i get the latest sources of the > kernel, userland... i need to upgrade my gcc and compile other gnu > compilers, but i can't seem to find any info on the port to Sparc... General info about SparcLinux: http://geog.ubc.ca/s_linux.html http://vger.rutgers.edu/sparclinux.html ftp://vger.rutgers.edu/pub/linux/Sparc/ http://ultra.linux.cz/ mailto:sparclinux@vger.rutgers.edu X11 development: http://sunsite.mff.cuni.cz/OS/Linux/Sparc Debian distribution: ftp://ftp.debian.org/dists/frozen/main/disks-sparc ftp://ftp.debian.org/dists/frozen/main/binary-sparc mailto:debian-sparc@lists.debian.org "UltraPenguin" distribution: http://sunsite.mff.cuni.cz/linux/ultrapenguin-1.0/ RedHat distribution: ftp://ftp.redhat.com/pub/redhat/redhat-4.2/sparc mailto:[EMAIL PROTECTED] BTW: Alberto, en la revista "Linux Actual" número 2, que acaba de salir en España, hay un artículo sobre SparcLinux; no sé si esa revista llega a Costa Rica y, si lo hace, cuándo llegará... la dirección de la revista es <[EMAIL PROTECTED]> Un saludo, -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: compiling gpm
On Wed, May 20, 1998 at 06:45:23PM +0200, Davide Barbieri wrote: > Ciao, > > look at this error messages: > > test gpm-1.14-beta4# make > cc -MM *.c > .depend > In file included from /usr/include/sys/socket.h:37, > from gpm.c:34: > /usr/include/bits/sockunion.h:25: warning: No include path in which to find > netatalk/at.h [snip] > > is it a libc6-dev problem? Yes. Until I upload a new version, please remove the "#include " in line 37 of /usr/include/sys/socket.h -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Uploaded perl 5.004.04-5 (sparc) to chiark
On Sun, May 17, 1998 at 10:02:04AM +0200, Martin Schulze wrote: > On Wed, May 13, 1998 at 11:26:13AM +0200, Juan Cespedes wrote: > > Format: 1.5 > > Date: Tue, 10 Mar 1998 23:00:09 -0800 > > Source: perl > > Binary: perl-suid perl-debug perl-base perl > > Architecture: sparc > > Version: 5.004.04-5 > > Distribution: unstable > > This should go into frozen, too, right? I thought Debian/Sparc wasn't supposed to be released as hamm. At least, that's what Brian White said, and we're only working with unstable (slink) right now. Should we re-upload all the things that go to slink back to hamm? (BTW: I didn't change the "Distribution" or "Urgency" fields; they were that way in the souce package). -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: [SECURITY] New version of procps
On Fri, May 08, 1998 at 12:57:15PM +0200, Martin Schulze wrote:
>
> we have received a security bulletin concerning su from the
> secure-su / shadow package. A fixed version has been uploaded for
> all architectures except sparc and powerpc. (powerpc is not
> supported yet so it doesn't count.) Does someone of you feel
> responsible for recompiling the hamm version of shadow for sparc
> and uploading it? We'll postpone our announcement for this upload.
> I'd appreciate an upload today or tomorrow.
Same problem we had with bind: "procps" uses features of a
moder debhelper which needs perl5 which isn't available yet on
SparcLinux.
Sorry, we can't build it yet.
--
Juan Cespedes
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: bugs (?)
On Wed, May 06, 1998 at 10:57:39PM -0500, John Goerzen wrote: > Jonas Oberg <[EMAIL PROTECTED]> writes: > > > Ohh.. this must be my lucky day! I managed to get a programmer > > interested in helping out! Ahhh.. I got him! I caught him! > > Moahahahahah! Hrm.. Oh. I didn't say that out loud, did I? ;-) > > hehe, you must not have been paying attention :-) > > I'm already a Debian developer but nobody seemed to be able to tell me > the correct way to go about uploading packages for sparc. The > suggested way of doing it caused the packages to be rejected due to > lack of PGP-signature. > > So how do I go about building a sparc binary from an i386 source? dpkg-buildpackage -B -m"John Goerzen <[EMAIL PROTECTED]>" That should generate one or more .deb and one .changes file, ready to be uploaded to master. However: I don't know if "pgp" works well on Sparc; maybe you have to use flags "-uc -us" in dpkg-buildpackage and sign the .changes manually from an i386 box using: pgp -ast file_sparc.changes ; mv file_sparc.changes.dsc file_sparc.changes -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
New silo!
I've uploaded a glibc-based silo to master. Please test it (it works for me). Davide, I've made a NMU, but you are still the maintainer. Is this OK for you? Thanks. -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: dpkg 1.4.0.22
On Sun, May 03, 1998 at 12:27:17PM +0200, Eric Delaunay wrote: > > I found the problem: it comes from binutils 2.8.0.1.23. I upgraded to 2.9.1 > then now it works. I had a similar problem building glibc... I should have uploaded a newer binutils then :) > If not already done, I will upload, tomorrow morning, the following packages: > binutils2.9.1-0.1 > dpkg1.4.0.22 > lout3.10-1.1 > psutils 1.17-3 > sgml-tools 1.0.5-1 > (and maybe other packages I will successfully compile today). I'll upload a newer glibc... let's see if perl works with this one. -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: kernel-patch*sparc* and libc-sparc
On Mon, Apr 27, 1998 at 01:37:44PM +0100, James Troup wrote: > Hi, > > [ Just looking at some ftp.debian.org problems in relation to Quinn > Diff, and these are the sparc-related areas ] > > a) Why is kernel-patch*sparc* Architecture: sparc and not >Architecture: all? You don't need a sparc to install it and cross >compile a sparc kernel (I assume). m68k's kernel-patch is Arch: >all, and I use it on an i386 machine to cross compile the m68k >kernel images. You are right here. Eric, can you change that? > b) Is there a binary package produced from the libc-sparc source >package (in oldlibs)? I can't see one looking at >hamm/hamm/binary-sparc/Packages. If not, I'll ask Guy to remove >the source package. It should be the source for libc5_5.3.12-2_sparc.deb. I don't know why does it show "libc" as the source package instead of "libc-sparc". It's there only because libc5_5.3.12 is the only version or libc5 ported to Sparc, and we needed the source for all our packages. -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: libc 2.0.92-980301-1
On Thu, Apr 23, 1998 at 12:50:29PM +0200, Eric Delaunay wrote: > > I heard libm is broken in this libc6 release. Is a new release planned soon ? I'm working on a new version. I'm having a few problems compiling it, but I hope they will be fixed in two or three days. After then, I'll try to work on dpkg, perl and strace. -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: dpkg 1.4.0.22
On Tue, Apr 21, 1998 at 11:08:47PM +0200, Eric Delaunay wrote: > > Hello, > > I compiled the latest dpkg but it fails to run. dpkg & dselect hang before > reading the database. We have had many problems with C++; that could be the reason of the "dselect" failure. But C++ now works! Please try the latest egcs for sparc installed in master. However, I don't know why could dpkg fail. Do you know where does it fail? I will work on it after releasing a new version of glibc. -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: this is not a sudo bug
On Fri, Apr 10, 1998 at 02:31:55AM -0600, Bdale Garbee wrote: > Hi. > > I'm closing bug 20883, because it's not a bug in sudo, and I suspect that sudo > will get recompiled in due course by someone working actively on the sparc > port. The root cause of the problem is an incompatible change in the glibc > bits for sparc that requires many binary packages to be recompiled. Yes, it's not a bug in sudo, but it still needs to be recompiled for Sparc. We have two choices: a) you upload a new version of sudo with no changes (or almost none); b) we make a non-maintainer sparc-specific upload. I don't mind uploading a non-maintainer release of sudo, and I'll do that if you don't mind. Thanks. -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: new perl.deb... (finally)
On Tue, Apr 07, 1998 at 10:36:22AM -0700, Stephen Zander wrote: > > Ok, I expect to make good on my long awaited commitment to upload > a sparc/linux version of the latest perl today. > > Before I do, has anyone already done so & is just waiting for it to > come out of Incoming or something similar? I'm currently rebuilding > my slink mirror, so I can't check if it's already there but it wasn't > as of yesterday. I think all of us are waiting for your perl package :) -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Freeze of Hamm in other architectures
On Fri, Apr 03, 1998 at 02:56:05PM -0500, Brian White wrote:
>
> What is the general consensus about the stability of each of these
> architectures (powerpc, alpha, and sparc)? Should they be release with
> the i386 and m68k versions as 2.0 or should that wait until a later time
> (perhaps 2.1)?
sparc port is more or less stable, but it still lacks many
important things such as C++ or X.
I think we can wait for 2.1.
Thanks.
--
Juan Cespedes
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: gcc v egcs on sparc/linux
On Tue, Apr 07, 1998 at 12:33:10PM -0700, Stephen Zander wrote: > > I thought I'd seen a thread (on debian-devel or debian-private) discussing > the use of egcs on sparc as a replacement for gcc. on sparc? I haven't seen anything about sparc. > Is this planned/in-progress/necessary/useful/pointless/other? The use of egcs instead of gcc is mainly as a C++ compiler, but libstdc++ doesn't currently work on Sparc, so we cannot chose :) Latest egcs in slink still doesn't work compiled on sparc. -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Freeze of Hamm in other architectures
On Fri, Apr 03, 1998 at 02:56:05PM -0500, Brian White wrote: > > What is the general consensus about the stability of each of these > architectures (powerpc, alpha, and sparc)? To the Sparc developers: We still don't have X packages, we have no working C++, and we have some problems with glibc. And there are lots of packages left to be ported. I think we should wait for 2.1 for Sparc. Comments? -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: VERY outdated boot disks
On Tue, Mar 31, 1998 at 08:28:35PM -0600, John Goerzen wrote: > The boot disks on the ftp site (/debian/hamm/hamm/disks-sparc) are > over a year old. The disks on lix.polytechnique.fr are much more > recent and work a lot better. Why are they not in the mirror? It is > not efficient for lots of people to have to FTP all the way from > France to get them. Plus, if that site ever goes down, we're stuck. A newer version is in Incoming/ in master since March 8th. -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: SILO
On Mon, Mar 23, 1998 at 10:04:18AM +0100, Davide Barbieri wrote: > [From a private discussion about silo] > > Jakub Jelinek wrote: > > > I wasn't able to compile it. > > > One problem appears when compiling second/main.c > > > > > > PAGE_MASK not defined; I simply add a #include > > > > This is not a bug in SILO. I bet you either have completely braindamaged > > glibc headers or kernel headers. > > Yes, I really think that I have a messed up system. Anyway, I have > installed the latest version of glibc and kernel-headers. > I have also done, a fresh install, but the problem remains. > > > I think it is not the right way to fix bugs in > > other places than the sources they are in. I never use PAGE_MASK in SILO, > so > > I don't see why we should include it. > > Actually, what kernel headers are you using? If it is because some kernel > > header forgets to include , then it is a bug in kernel headers. > > But as far as I remember, this was fixed in kernel headers a long time ago. > > I use kernel-headers-2.0.33 It's a bug in kernel-headers-2.0.33, from sparclinux-2.0.33: you always have to include if you include . But it shouldn't hurt. -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: libc6-dev & libgdbmg1-dev conflicts
On Sun, Mar 22, 1998 at 09:45:56PM +0100, Eric Delaunay wrote: > > Last note: > libc6 does not provide the -doc package in which the info files reside > regarding the 2.0.7 release. Could this package be created for the next > upload (or these files added to the libc6 package) ? The documentation on > setting NIS is only provided in info format. libc6-doc is architecture: all, and if I upload it, it will replace the one from version 2.0.7 which is used in other architectures... But well, I think I can upload it as architecture: sparc. I'll do it in a few days. -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: g++, libstdc++ & I/O status
On Sun, Mar 22, 1998 at 07:25:42PM +0100, Eric Delaunay wrote: > > I further worked on egcc 1.0.2 (the latest release available including some > fixes for glibc 2.1). > Now my program hangs without displaying the message at all. > Under the debugger, it seems it loops indefinitely between _IO_file_seekoff > and > __register_frame_info in the _init function (called before main). > > I'm sorry, I have no more idea to provide a working g++ :-(( > In effect I don't know if the bug is in g++, libstdc++ or libc6 :-( > > Any new idea? I think we should try the latest snapshots of egcs (from egcs.cygnus.com). -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Patch for a group problem
On Wed, Mar 11, 1998 at 11:46:11PM +0100, Andreas Jaeger wrote: > > The following patch fix one group problem in glibc 2.1 that I had: > Displaying Nonsense :-) > > id aj gave the following output for me - due to corrupted memory :-(. > uid=207(aj) gid=100(users) > groups=0((¶p@(¶p@:root),4(adm),10(),14(),101(),102((¶p@([EMAIL > PROTECTED]:aj),104([EMAIL > PROTECTED]@4:aj),0(),4(),10(),14(),101(),102((¶p@([EMAIL > PROTECTED]:aj),104([EMAIL PROTECTED]@4:aj) > > This should be fixed. I can't reproduce the problem some of you > mentioned with root in the group list. It might be the same problem. > Please tell me if this helps or not. Even with that patch, it doesn't work at all. I'm trying to debug it... -- Juan Cespedes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Can't compile glibc with newer binutils :(
Unfortunately, binutils_2.8.1.0.23 fails to build glibc properly. It build everything OK (apparently), but when loading ld-linux.so.2 it creates something which doesn't work. It gets running all the time, without displaying anything... I'll try to get an older version of binutils. -- Juan Cespedes -- E-mail the word "unsubscribe" to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED]
Re: hamm freeze
On Thu, Mar 12, 1998 at 09:09:24AM +0100, Davide Barbieri wrote: > Juan Cespedes wrote: > > > Eric Delaunay wrote: > > > > - silo big troubles with libc6 :-(( > > > [...] > > > > - binutils 2.8.1.0.21 is out (maybe solve the silo problems ???) > > > > > > should I try the new binutils? Juan, are you going to > > > package them soon? > > > >Uploaded to master. > > binutils-dev_2.0.1.0.23 depends on binutils << 2.0.1.0.20 > and >> 2.0.1.0.19, which, AFAIK, isn't available as a debian > package Oops! I have just uploaded a new version which fixes this. > Instead, now I have a new problem: after installing libc6 > 2.0.92, /bin/login complains about not being able to load > shared libraries: symbol setutxent, version GLIBC_2.0 not > defined in file libc.so.6 with link time reference Hmm... I suffer the same thing, but I hadn't noticed it because I always use ssh instead of logging directly... Maybe we'll need to recompile login... -- Juan Cespedes -- E-mail the word "unsubscribe" to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED]
Re: hamm freeze
On Tue, Mar 10, 1998 at 10:18:04PM +0100, Eric Delaunay wrote:
>
> - glibca new release is available for testing
I don't know if it's a problem with initgroups() or
getgroups(), but I get "cespedes root" when using `groups' (it should
be just "cespedes"). Does anyone else suffer this?
I'm investigating it.
--
Juan Cespedes
--
E-mail the word "unsubscribe" to [EMAIL PROTECTED]
TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: hamm freeze
On Wed, Mar 11, 1998 at 09:18:00AM +0100, Davide Barbieri wrote:
> Eric Delaunay wrote:
> > - silo big troubles with libc6 :-((
> [...]
> > - binutils 2.8.1.0.21 is out (maybe solve the silo problems ???)
>
> should I try the new binutils? Juan, are you going to
> package them soon?
Uploaded to master.
--
Juan Cespedes
--
E-mail the word "unsubscribe" to [EMAIL PROTECTED]
TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: hamm freeze
On Tue, Mar 10, 1998 at 10:18:04PM +0100, Eric Delaunay wrote: > > Hello, > > hamm will be frozen in few days, but what should be made about the sparc > port? Are we ready for a release? > I did a quick check on the Packages file: > 616 packages were compiled for sparc (don't know how many are up-to-date) > 296 packages are arch indep. > (about 1500 packages are in the dist). > > If we want to release something, we need at least a core distribution (base & > net stuff, some admin tools, minimal X11 stuff). > > So, can we write a list of what should be (re)compiled first ? > It comes to mind some packages already in the process queue: > - silo big troubles with libc6 :-(( > - glibca new release is available for testing It's already in master's Incoming. I'm waiting for Guy Maor to install it. > - binutils 2.8.1.0.21 is out (maybe solve the silo problems ???) I'm building binutils-2.8.1.0.23 now. -- Juan Cespedes -- E-mail the word "unsubscribe" to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: libc6 progress
On Fri, Mar 06, 1998 at 08:37:43AM -0800, Joel Klecker wrote: > At 16:14 +0100 1998-03-06, Juan Cespedes wrote: > > I'm not going to upload it yet, I'm going to make sure which > >packages fail. If someone wants to have a look at it (warning: they > >may make your system unusable), I'll put it in > >master.debian.org:/home/cespedes. > > Would you mind putting the sources there too? (just the diff and dsc will > be fine) It's all there now. I haven't found any package failing; I'll test it a bit more and will upload it tomorrow if all goes OK. -- Juan Cespedes -- E-mail the word "unsubscribe" to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: How to help?
On Wed, Mar 04, 1998 at 01:16:40PM -0600, John Goerzen wrote: > Hi, > > We are installing Debian on a Sparc machine here. I have noticed that > there are a number of packages for i386 Debian that are either missing or > outdated for Sparc Debian. I am a maintainer of various packages under > i386, so I am familiar with how the Debian packaging system works. > > I am able to easily grab the source packages for various things and > generate the deb for sparc. My question is this: Is it possible for me to > upload these debs to master for inclusion in the sparc archive, even > though I am not a maintainer of those packages and don't want to be? Yes, of course. But before uploading one package, make sure that it works with the latest libraries... we are having some problems with glibc lately :) > And if so, what is the procedure for doing so? dpkg-buildpackage -B -rsudo -m"John Goerzen <[EMAIL PROTECTED]>" ... and upload the resulting .changes. If you have to modify the sources in any way, submit a bug report against the maintainer of the package telling what did you have to change. Note: `fakeroot' doesn't work yet (at least, on my system), so you will probably have to use `sudo'. > Does dupload/release generate the appropriate release messages? dupload works for me. I have never used `release'. > Do I need to note something in a change > log or create a new "debian" version of the package? You should try to upload it without any changes. If you need any changes in the *sources*, submit a bug report. You shouldn't change the copyright or changelog unless the current version is already uploaded and you have to change it for some reason. In this case, just make a non-maintainer release in the usual way. -- Juan Cespedes -- E-mail the word "unsubscribe" to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
libc6 progress
Well, I've been able to successfully compile the latest glibc snapshots. As Ulrich said, there are some binary incompatibilities; even some essential packages fail. I'm trying to make a list; all of them will have to be installed at the same time as the new glibc, preferably in single user mode. I'm not going to upload it yet, I'm going to make sure which packages fail. If someone wants to have a look at it (warning: they may make your system unusable), I'll put it in master.debian.org:/home/cespedes. -- Juan Cespedes -- E-mail the word "unsubscribe" to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: debian/sparc progress
On Mon, Mar 02, 1998 at 10:45:46PM +0100, Eric Delaunay wrote: > > Hello all developers, > > this is a quick summary of developments needed to continue on the sparc > port: Are you still working on kernel-heades-2.0.33_sparc? I would like to upload the next glibc soon, and I would like to know if I could make it depend on kernel-headers :) Thank you. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: debian/sparc progress
On Tue, Mar 03, 1998 at 09:28:06AM +0100, Andreas Jaeger wrote: > >>>>> Juan Cespedes writes: > > Juan> Yes, I'm working on it. It compiles more or less cleanly, but > Juan> it fails to run *any* program. I'm trying to find the reason... > Is even make check failing? That would be really strange. The current > snapshot works on ix86 (and AFAIK alpha also) but I didn't see anybody > complaining. Yes, I've also compiled them on i386 and they work without any patches... but I have to apply a few for Sparc. I'll send them to Uli in a few days. (BTW: if you want, I can send them to you now, but I'm still not sure if all of them are `correct'). > P.S. Richard checked in the following sparc related changes. I don't > know if they would help. Yes, I think some of them are from the Debian sources :) But they aren't yet in the latest diff in alpha.gnu.org. Thanks for your help. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: debian/sparc progress
On Mon, Mar 02, 1998 at 03:44:47PM -0800, Stephen Zander wrote: > >>>>> "Juan" == Juan Cespedes <[EMAIL PROTECTED]> writes: > Juan> Yes, I'm working on it. It compiles more or less > Juan> cleanly, but it fails to run *any* program. I'm trying to > Juan> find the reason... > > Has this version been released at all? No, not yet. You can find the upstream sources in ftp://alpha.gnu.org/gnu/libc ftp://ftp.funet.fi/pub/gnu/alpha/gnu/libc ftp://ftp.kernel.org/pub/linux/libs/glibc I'm working on it... -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: debian/sparc progress
On Mon, Mar 02, 1998 at 10:45:46PM +0100, Eric Delaunay wrote: > > Hello all developers, > > this is a quick summary of developments needed to continue on the sparc > port: > > 1/ there is a new, binary incompatible, glibc snapshot release by the GLIBC >team (see a previous post from Ulrich Drepper <[EMAIL PROTECTED]>). >I think we *should* release it ASAP to base our new ports on it. >Juan do you work on it ? Yes, I'm working on it. It compiles more or less cleanly, but it fails to run *any* program. I'm trying to find the reason... I hope we will not have to change many programs after this new version is released. I'll try to have it working at the end of this week. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: LIBC 6 changes
On Tue, Feb 17, 1998 at 12:42:00PM -0800, Bruce Perens wrote: > Dear Debian SPARC developers, > > Ulrich Drepper [EMAIL PROTECTED] is the head GNU LIBC developer. > He would like to be kept in the loop when you make changes to LIBC. > Please make sure to send him patches, etc. Hi. I'm the glibc maintainer for Sparc. Sorry, I didn't send our patches to Uli yet mostly because I was a bit busy lately... but I'll begin doing it as soon as I manage to compile the latest release (in one or two days). -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: kernel-source-2.0.33 for sparc
On Tue, Feb 17, 1998 at 09:37:59AM +0100, Eric Delaunay wrote: > > BTW, how to handle the kernel-patches package ? > I have to upload a binary package for it (kernel-patches-*_sparc.deb) but > should I also upload its source counterpart ? Well, I think so... every binary package should have an associated source package. But how to call the *source* package? `kernel-patches-2.0.33' won't be a good idea; I would use `kernel-patches-sparc'. I think James was also working on kernel-patches for m68k. James? Have you uploaded it? What name for the source package are you using? -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: kernel-source-2.0.33 for sparc
On Wed, Jan 28, 1998 at 03:40:57PM +0100, Eric Delaunay wrote: > > I'm willing to create a kernel-image package based on 2.0.33 for sparc. Are you still working on it? I would like to upload a new glibc which would depend on kernel-headers-2.0.33, to avoid including many files in the package and to make some people happier :) -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: New name for glibc-sparc?
On Mon, Feb 16, 1998 at 01:15:41PM +0100, Andreas Jaeger wrote: > > I forgot one question: > Are you going to put the kernel headers in a separate package as it > has been done for glibc-2.0.6? Yes, I will, but I can't do it with the current kernel-headers package. I'm waiting for a specific kernel-headers for Sparc. BTW: Eric? Are you there? Will we have a kernel-headers-2.0.33_*_sparc.deb soon? -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: New name for glibc-sparc?
On Sun, Feb 15, 1998 at 02:01:09PM +0100, Andreas Jaeger wrote:
> >>>>> Juan Cespedes writes:
>
> >The `glibc-sparc' *source* package is what we are using now as
> > the sources for libc6 in Sparc; however, there's nothing
> > sparc-specific on it, it's just a newer version of the `glibc' used in
> > other architectures. We need another source package name so that the
> > two versions can coexist.
>
> Are you adding the actual date of the snapshot? In that case you might
> consider calling it just e.g. glibc-snapshot-980202.
I'm adding the date only in the *version*, not in the package
name.
Thanks for your suggestions, I think I'll use `glibc-pre2.1'.
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
New name for glibc-sparc?
The `glibc-sparc' *source* package is what we are using now as the sources for libc6 in Sparc; however, there's nothing sparc-specific on it, it's just a newer version of the `glibc' used in other architectures. We need another source package name so that the two versions can coexist. It's used mainly on Sparc, but it may be used in other archs sooner or later. The `powerpc' people, for example, is willing to use it, and we may try to use it for i386 after Debian 2.0 is released. However, I would like to change its name: `glibc-sparc' seems very sparc-specific, and I would like to hear your suggestions on this issue. How about `glibc-pre2.1'? (it's based on pre-alpha versions of glibc-2.1). (BTW: I would like to use `glibc-alpha', but that could confuse the AXP people :)) Other alternatives would be `glibc-beta', `glibc-exp', `glibc-devel'... Comments? -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: LD_PRELOAD used with setuid programs (was Re: Fakeroot security problem)
On Mon, Feb 09, 1998 at 10:36:26PM +, James Troup wrote: > James Troup <[EMAIL PROTECTED]> writes: > > [ lots of stuff that was intended for debian-private and not > debian-sparc ] > > Excuse me, but which genius decided to use a *public* mailing list as > a Maintainer: address? Sorry, my fault. The libc6 used in Sparc (pre2.1) is too buggy and there are many people working on it, so I thought it could be useful if debian-sparc is the maintainer of the new `glibc-sparc' source package. I didn't want that address to receive mails addressed to [EMAIL PROTECTED]'. I'll change the Maintainer: field in the next upload (in one or two days). > James - snarked off. Sorry :( -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: LD_PRELOAD used with setuid programs (was Re: Fakeroot security problem)
On Mon, Feb 09, 1998 at 09:30:15PM +, James Troup wrote: > Juan Cespedes <[EMAIL PROTECTED]> writes: > > > Yes, both ld-linux.so.2 and ld-linux.so.1 should be fixed; nobody > > should be able to run a setuid program in a LD_PRELOAD environment. > > At least, I can't find any reason to allow it, and many people could > > use it to try to find exploits. > > But there _are_ reasons to do allow it (see below, and also add > libnfslock to the list). If there weren't any someone would have > presented these patches much earlier. I disagree. That sort of things shouldn't be done by an ordinary user, but by the sysadmin (writing wrappers for the programs involved, for example). The package `libnfslock' adds one line to /etc/ld.so.preload. That file has always worked, even with setuid programs, and will not stop working because of this patch. > > --- Start of forwarded message --- > Message-ID: <[EMAIL PROTECTED]> > Date: Sun, 8 Feb 1998 15:39:10 -0600 > Reply-To: Aleph One <[EMAIL PROTECTED]> > From: Aleph One <[EMAIL PROTECTED]> > Subject: Re: Another ld-linux.so problem > To: [EMAIL PROTECTED] > > On Sat, 7 Feb 1998 [EMAIL PROTECTED] wrote: > > > Yes. SOCKSifying stupid protocols that require binding ports <1024, for > > example. Assuming you install libsocks5_sh.so in /usr/lib, you can do: > > > > $ (export LD_PRELOAD=/usr/lib/libsocks5_sh.so; rsh machine.outside.firewall > > pwd) > > > > and have it work. This is basically what the runsocks script does. If that is needed, then that `runsocks' script should be a setuid binary that does it. > Another example: installing a library that overides mktemp, tempnam and > other dangerous library functions with more secure ones. So the feature > is indeed useful. The correct behavior should be for the dynamic linker > to give up at the first error. Alternatively you should be able to > configure such libraries via the configuration file instead of an > environment variable. You cant do so now as far as I can tell. Suppose that some library corrects one important security fix in `/bin/su' replacing a `dangerous library function', and the sysadmin adds that library to /etc/ld.so.preload. If we still can use LD_PRELOAD with setuid binaries, an ordinary user could do: $ LD_PRELOAD=libc.so.6 /bin/su in order to use the symbols from libc.so.6 instead of the ones in your library (yes, LD_PRELOAD is evaluated before /etc/ld.so.preload). Additionaly, it's very difficult for me to trust all the libraries in /lib, /usr/lib, and all the directories contained in /etc/ld.so.conf... for example, do you really think that it's safe to use LD_PRELOAD with the libraries `libc.so.4' or `libc.so.5' in a libc6-compiled setuid binary? -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: libc6 status
On Mon, Feb 09, 1998 at 02:41:45PM +0100, Andreas Jaeger wrote:
>
> Don't make your life too difficult. Just apply the following patch
> which should be in the next glibc snapshot.
> Index: libc.map
> ===
> RCS file: /egcs/carton/cvsfiles/libc/libc.map,v
> retrieving revision 1.38
> diff -u -r1.38 libc.map
> --- libc.map 1998/02/01 16:18:18 1.38
> +++ libc.map 1998/02/09 07:52:01
> @@ -20,6 +20,9 @@
> # This is for ix86 only.
> _fp_hw;
>
> +# This is for sparc only.
> +.div; .mul; .rem; .udiv; .umul; .urem;
> +
Sorry, but that wouldn't work. Current binutils don't work
if you include symbols beginning with `.' in the libc.map. This is
(IMHO) a bug in binutils... I'm CC'ing H.J. Lu.
> Juan> The glibc development people are a bit quiet... there aren't
> Juan> new versions yet, but I'm going to release a new version including a
> Juan> patch for /etc/ld.so.preload handling (it doesn't work now), and that
> Juan> library including .{u,}{div,mul,rem}
>
> Whom are you contacting? I'm with the glibc team and asked already
> some times for patches and didn't receive any answer at all:-(. As
> I've told you already Ulrich Drepper who's mainting glibc is moving at
> the moment and therefore not available (but should be soonish again).
Sorry, I didn't want to disturb you... all of you are doing a
very good job :)
I sent those suggestions to Miguel de Icaza, who is
(was?) the one who used to apply patches for Sparc on glibc, and
didn't get any answer.
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: libc6 status
On Sun, Feb 08, 1998 at 01:47:56AM +0100, Eric Delaunay wrote:
>
> 1/ rootdisk.sh (from boot-floppies-2.0.0) searchs for
>/usr/lib/libc_pic/soinit.so, /usr/lib/libc_pic/sofini.so &
>/usr/lib/libc_pic/interp.so files that are missing from our libc6 package.
>I don't know much about creating the subset libc library used in rootdisk,
>but I guess it follows the same method as of libc6 for Intel.
>Can you check the Debian rules file from libc6 2.0 to create a similar
>libc6-pic package ?
Hmmm... I can't find those files in libc6-pic_2.0.6-3_i386.deb
either. Are you sure they are needed?
In my glibc tree, I find `elf/soinit.os' and `elf/sofini.os'.
Do you know if those are the same files? However, I can't find
`interp.so'.
> 2/ the .rem symbol is defined in libslang (0.99.38) but not in the the
> libslang
>generated for miniroot by processing the libslang-pic package.
I'm going to define all the `.{u,}{div,mul,rem}' symbols in a
new shared library to be included in /etc/ld.so.preload, until the
glibc development team decides if they should be defined in libc.so or
not. (BTW: which name would you use for that library? `sparc-arith.so'?)
> Before trying to fix the latter problem, I would want to know if a new libc6
> upload is planned?
The glibc development people are a bit quiet... there aren't
new versions yet, but I'm going to release a new version including a
patch for /etc/ld.so.preload handling (it doesn't work now), and that
library including .{u,}{div,mul,rem}
Do you need any other change in libc6?
> PS: In the meantime of having a libc6 which could be reduced cleanly from
> the pic lib, I included a full version in the ramdisk. However, I
> cannot
> create a rescue disk because this ramdisk is too large. A complete
> libc6
> takes about 960 KB !
Maybe we could compile other libc6 stripping uneeded things
for the bootdisk (such as threads, nis, db...)
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: boot-floppies progress
On Sun, Feb 08, 1998 at 01:20:16AM +0100, Eric Delaunay wrote: > > Hello all, > here is my progress in this development of boot disks for Sparc. Fine :) > BTW, I've created 2 new packages only useful in the sparc world: > .fdisk-sparc is the missing program from util-linux > .sparc-tools contains elftoaout & piggyback utilities to create tftpboot >images with initrd. > Can I upload them directly to master or is there a policy about announcing new > packages before ? The policy says you should announce them to debian-devel, but I think it's not necessary here as they are sparc-specific and you already announced them on debian-sparc... > Notice: I don't know if sparc-tools is the right name for this package. > I'm thinking about sparc-utils or tftpboot-utils, ... > Any suggestion is welcome. I would suggest `sparc-utils'... -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
LD_PRELOAD used with setuid programs (was Re: Fakeroot security problem)
On Wed, Jan 28, 1998 at 11:52:45PM +, Mark Baker wrote:
> On Wed, Jan 28, 1998 at 11:46:46PM -, [EMAIL PROTECTED] wrote:
> > Fakeroot is not the only library in a privileged directory that should
> > not be run with a setuid executable by an unprivileged user.
> > Although I don't know of an exploit, my Electric Fence library has not
> > been audited for that kind of security. If there is a shared library for
> > Checker, it probably has the same status.
>
> Do you think debian should, whatever the upstream author does, fix ld.so so
> this isn't possible (it would actually make the code simpler)? I can't think
> of any good reason not to, even if the security risk is negligible.
Yes, both ld-linux.so.2 and ld-linux.so.1 should be fixed;
nobody should be able to run a setuid program in a LD_PRELOAD
environment. At least, I can't find any reason to allow it, and many
people could use it to try to find exploits.
The fix is very easy, both in libc6 and ld.so:
=== patch for glibc_2.0.6-2 ==
--- elf/rtld.c. Sun Feb 8 22:55:45 1998
+++ elf/rtld.c Sun Feb 8 22:57:02 1998
@@ -356,7 +356,7 @@
char *list = strdupa (preloadlist);
char *p;
while ((p = strsep (&list, " ")) != NULL)
- if (! __libc_enable_secure || strchr (p, '/') == NULL)
+ if (! __libc_enable_secure)
{
struct link_map *new_map = _dl_map_object (NULL, p, lt_library, 0);
if (new_map->l_opencount == 1)
==
=== patch for ld.so-1.9.6
--- d-link/boot1.c. Mon Jul 21 21:45:35 1997
+++ d-link/boot1.c Sun Feb 8 22:59:26 1998
@@ -561,7 +561,7 @@
str2++;
c = *str2;
*str2 = '\0';
- if (!_dl_secure || _dl_strchr(str, '/') == NULL) {
+ if (!_dl_secure) {
tpnt1 = _dl_load_shared_library(NULL, str);
if (!tpnt1) {
if (_dl_trace_loaded_objects)
======
Should I submit a bug report against these packages?
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
kernel-headers-* for different archs (was Re: What's Debian's /usr/src policy)
On Thu, Jan 22, 1998 at 11:05:25PM -0600, Manoj Srivastava wrote:
>
> Should I change the architecture of the kernel-headers package
> from all to any?
I think so.
Debian supports at least 4 architectures: i386, alpha, m68k
and powerpc. How many of them can use `kernel-headers-2.0.32'?
At least, `sparc' doesn't.
I'm CC'ing to debian-{alpha,m68k,powerpc,sparc}.
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: FD_SET
On Thu, Feb 05, 1998 at 04:58:21PM +0100, Davide Barbieri wrote: > Ciao, > someone could explain to me, why file descriptors are now > defined both in sys/select.h and in linux/time.h? > Is this correct? glibc-2 tries to be self-contained; you shouldn't include in any application (excepting those very related to the kernel); so, many things that used to be in are now in ; and that's where the programs should look for them. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: debian and sparc...
On Wed, Jan 28, 1998 at 04:22:23PM +0100, Marcus Riska wrote: > Its a simple question... > When is sparc debian planned to be released? what is the current status of > the project? It may be ready within one or two months... we even may have it available when Debian 2.0 is released. The kernel works ok, the glibc is beginning to stabilize, and we already have lots of packages. The main lack is the install disks, but some people are working on it. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: new uploads
On Mon, Jan 26, 1998 at 12:30:18AM +0100, Eric Delaunay wrote: > > I uploaded full libc5 package (sources + binaries) patched for utmp-wrapper. > I also uploaded new netbase (3.03-1) & netstd (3.02-1) to fix troubles with > the > glibc upgrade. I'm afraid it won't work. From your .changes: -- Files: bb506e259dfd8a8ce5ad0908916ed3e0 657 oldlibs required libc_5.3.12-2.dsc 7ddd2fdf228cf65515e2ac26b8ada975 1537477 oldlibs required libc_5.3.12.orig.tar.gz 0b830cfca73ba7dedc9ae157b30b292e 1161054 oldlibs required libc_5.3.12-2.diff.gz -- That won't be installed because a newer version of the source package `libc' already exists in /debian/hamm/hamm/source. If you want that sources to be included, you'll have to change the source package name, as I did with `glibc-sparc'. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
New releases of glibc-2.1?
Does anyone know what is happening to glibc-2.1? There aren't any changes in the source tree since December 29 (there used to be one or two releases every week). -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: bad experiences with libc6 2.0.91-980111-3
On Fri, Jan 23, 1998 at 02:52:50PM +0100, Eric Delaunay wrote: > > I get the following error from the rlogin process: > > usr/sbin/in.rlogind: Symbol `_sys_errlist' has different size in shared > object, consider re-linking > > (Hopefully I can still login from consoles or telnet). > > It seems to come from tcpd in netbase package. > (sed is another package which show me the same warning). According to the glibc FAQ, `_sys_errlist' shouldn't be used. `strerror()' should always be used instead. List of binaries (in my machine) which use `_sys_errlist': $ grep -l _sys_errlist /bin/* /usr/bin/* /sbin/* /usr/sbin/* /bin/sed /usr/bin/fping /usr/bin/fromport /usr/bin/last /usr/bin/lastb /usr/bin/namei /usr/bin/skill /usr/bin/slogin /usr/bin/snice /usr/bin/ssh /usr/bin/strace /usr/bin/toport /usr/bin/vrfy /usr/bin/w /sbin/getty /sbin/init /sbin/rmt /sbin/syslogd /sbin/telinit /usr/sbin/portmap /usr/sbin/rmt /usr/sbin/rpc.ugidd /usr/sbin/sshd /usr/sbin/tcpd /usr/sbin/tcpdchk /usr/sbin/tcpdmatch /usr/sbin/try-from > Now I'm trying to build the latest netbase package (3.03-1) but it fails due > to some conflicts between bits/socket.h and asm/socket.h. Why does it need `asm/socket.h'? -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: .rem .div .mul ... in glibc-2.1
On Fri, Jan 23, 1998 at 02:44:54PM +0100, Eric Delaunay wrote:
> Juan Cespedes wrote:
> > Why aren't .[u]{div,rem,mul} defined as global in glibc-2.1?
> > Is there a good reason for this?
> >
> > I think they should be added to `libc.map' in the glibc
> > sources to declare them as global. If not, some programs may not work
> > (in fact, they don't work now).
>
> My first attempt (a while ago) to build glibc was to add such entries in the
> libc.map because I got some undefined symbols at link time. It was wrong. I
> didn't realized that libc.so should not be a soft link to libc.so.6 but have
> to be a script that tells the linker to use the dynamic library first, then
> the static one to resolve undefined symbols in the former.
Why was it wrong?
I see two problems if we leave those symbols undefined:
1) Some of them are defined in ld-linux.so.2, because they are used there:
> $ nm -D /lib/ld-linux.so.2 | grep ' \.'
> 000127bc T .div
> 00012a64 T .udiv
> 000126c4 T .umul
> 00012cdc T .urem
And, see this:
> $ cat /usr/lib/libc.so
> /* GNU ld script
>Use the shared library, but some functions are only in
>the static library, so try that secondarily.
>The dynamic linker defines some functions used by libc.so.6,
> but ld uses definitions from libc.a before examining the
>dependencies of libc.so.6 to find ld-linux.so.2. */
> GROUP ( /lib/libc.so.6 /lib/ld-linux.so.2 /usr/lib/libc.a )
So, if my program (`hello') uses `.div', at build time gcc searches
first in libc.so.6; it isn't there. Then, it finds in ld-linux.so.2
and volia! it's there, and that symbol is declared as `undefined'.
But, if two releases of the glibc later, ld-linux doesn't need `.div'
anymore, it would be out of that library and my program would say
`undefined symbol'.
2) Let's suppose you write a program which uses some symbol from libm,
which in fact uses some of these symbols (eg, `.umul'). At build time,
that symbol will be found in libc.a and it will be included in the
executable. But what would happen if two releases of the glibc later,
that symbol from libm does not only use `.umul', but also `.mul'? You
will have one undefined symbol that the interpreter won't be able to
resolve.
So, I see one of these solutions:
1) Define them as global in libc.so.6 (adding them to libc.map)
2) Leave them only in libc.a, *and* make them local in ld-linux.so.2
*and* include them as local in libm.so.6
I think the former is the best one.
> Can you try to replace libm.so link with this script:
>
> GROUP ( /lib/libm.so.6 /usr/lib/libc.a )
I think that won't work. Think about it.
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: What's Debian's /usr/src policy
On Fri, Jan 16, 1998 at 03:46:11PM -0800, Joel Klecker wrote: > > powerpc, and sparc both use glibc 2.1, but I expect that the others will > move in that direction eventually, once glibc 2.1 goes final. I didn't know that powerpc was already using glibc-2.1. Is this the case? If so, we might work together and have the same source tree. The source package we use with sparc is called `glibc-sparc', but if powerpc also uses pre2.1, I might rename it to `glibc-pre2.1'. All the binary packages have the same name they have in other architectures (libc6, libc6-dev, etc). -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
.rem .div .mul ... in glibc-2.1
Why aren't .[u]{div,rem,mul} defined as global in glibc-2.1?
Is there a good reason for this?
I think they should be added to `libc.map' in the glibc
sources to declare them as global. If not, some programs may not work
(in fact, they don't work now).
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Uploaded glibc-sparc 2.0.91-980111-3 (source sparc) to master
-BEGIN PGP SIGNED MESSAGE- Format: 1.5 Date: Thu, 22 Jan 1998 09:32:10 +0100 Source: glibc-sparc Binary: timezones libc6 libc6-pic libc6-dbg locales libc6-doc libc6-dev Architecture: source sparc Version: 2.0.91-980111-3 Distribution: unstable Urgency: low Maintainer: Juan Cespedes <[EMAIL PROTECTED]> Description: libc6 - The GNU C library version 2 (run-time files). libc6-dbg - The GNU C library version 2 (debugging/profiling files). libc6-dev - The GNU C library version 2 (development files). libc6-pic - The GNU C library version 2 (PIC library) locales- Locale data files and utilities. timezones - Time zone data files and utilities. Changes: glibc-sparc (2.0.91-980111-3) unstable; urgency=low . * Really removed `_IO_2_1_stdin_' warning now * Include missing linux/version.h and linux/autoconf.h * Replace /usr/include/asm-sparc with /usr/include/asm Files: 1ade8c9cf314bcb162273132116ca282 430393 libs required glibc-sparc_2.0.91-980111-3.diff.gz 2761de6522c912a4c5f019c94bbcf5e6 738 libs required glibc-sparc_2.0.91-980111-3.dsc d3881962dc1f4e92ef625b37b1895e7d 1363748 base required libc6_2.0.91-980111-3_sparc.deb da8038e341bd694b951c97699aab5dd6 1631272 devel standard libc6-dev_2.0.91-980111-3_sparc.deb 230bfa6e951c99049d6add4ce8b4a749 6727456 devel optional libc6-dbg_2.0.91-980111-3_sparc.deb ff8dce4a2363116b52c077feb3415ace 2041174 devel optional libc6-pic_2.0.91-980111-3_sparc.deb 1cbe1300e6b7b02412d06abcb632710e 1225840 admin standard locales_2.0.91-980111-3_sparc.deb 22ef294cb0e8a694aebba902bea2bfd5 258424 base required timezones_2.0.91-980111-3_sparc.deb -BEGIN PGP SIGNATURE- Version: 2.6.3ia Charset: noconv iQCVAwUBNMc+c12FmLxCq1O1AQGSugP/QCn5JZp+gimXlH2vBNSC907RjK7Zbozt bGfSTOfyFWZatUhZ1HimmhYOCnME0KUSh+ZL8X41G305xGqmzzKg4saw8EIdrIWH iD5mkaJKyj6qqwq6CTdWpNeaxw33KPgNjEJnP+IHMbZf5RtJF9Bw7xtpzbLZzjtr VPsJOIQG8nI= =/0O5 -END PGP SIGNATURE- -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Uploaded glibc-sparc 2.0.91-980111-2 (source sparc) to master
On Wed, Jan 21, 1998 at 11:57:12PM +0100, Eric Delaunay wrote: > > I suggest the next upload of libc6 to completely squeeze this message. Done. I'm uploading it now (but it's a bit slow...). It will be in Incoming/ tonight (CST time). If you want to begin compiling things, a copy of the fixed ld-2.0.21.so is in my home directory in master. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Uploaded glibc-sparc 2.0.91-980111-2 (source sparc) to master
On Wed, Jan 21, 1998 at 06:52:43PM +0100, Gergely Madarasz wrote: > Would it be possible to just provide a dynamic linker separately from the > original package somewhere in your home directory ? OK, I'm recompiling it now; I'll leave it in my home dir tomorrow. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: poor old paci :-(
On Wed, Jan 21, 1998 at 03:16:12PM +0100, Davide Barbieri wrote: > Unfortunately I have to upgrade almost everything. I started > with glibc (the last one), and then I upgraded the login > package. Now I'm no more able to login. What's wrong? You should install before the last base-files, to install /etc/nsswitch.conf. The next version of glibc will depend on base-files (>= whatever). -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Uploaded glibc-sparc 2.0.91-980111-2 (source sparc) to master
On Tue, Jan 20, 1998 at 08:30:29AM -0500, Johnie Ingram wrote: > > I think the easiest thing to do is ask Guy to rm $(find */binary-sparc > -type f), then upload correct versions of the packages which must be > compiled by hand. The rest of the distribution can be rebuilt in > about 4 days. No, I don't think so. There shouldn't be different packages with the same version; someone could have installed them and he would think there aren't new ones. Proposed solutions: 1.- Do non-maintainer uploads of all of them. 2.- Patch glibc for not displaying any warning, and only upload the ones that fail (if there are any). -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Uploaded glibc-sparc 2.0.91-980111-2 (source sparc) to master
On Tue, Jan 20, 1998 at 09:43:27AM +0100, Juan Cespedes wrote: > Please tell me if you want me to upload a new version with a > little patch to avoid displaying this warning (the code is already > written; I only need to re-compile and re-upload the whole thing). Note: maybe it's a good idea to leave the warning as it is now; this way, we know which programs still need to be recompiled with this new glibc. Note 2: maybe there are some programs which fail to run, because of this little incompatibility. I haven't found any yet, but be warned. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Uploaded glibc-sparc 2.0.91-980111-2 (source sparc) to master
Note: this libc has some incompatibilities with the older version (2.0.90-971126-1). I think all the programs should still work fine, but most of them will display a warning on stderr, similar to these ones: cp: Symbol `_IO_2_1_stdin_' has different size in shared object, consider re-linking rm: Symbol `_IO_2_1_stdin_' has different size in shared object, consider re-linking ln: Symbol `_IO_2_1_stdin_' has different size in shared object, consider re-linking Please tell me if you want me to upload a new version with a little patch to avoid displaying this warning (the code is already written; I only need to re-compile and re-upload the whole thing). -BEGIN PGP SIGNED MESSAGE- Format: 1.5 Date: Mon, 19 Jan 1998 13:44:04 +0100 Source: glibc-sparc Binary: timezones libc6 libc6-pic libc6-dbg locales libc6-doc libc6-dev Architecture: source sparc Version: 2.0.91-980111-2 Distribution: unstable Urgency: low Maintainer: Juan Cespedes <[EMAIL PROTECTED]> Description: libc6 - The GNU C library version 2 (run-time files). libc6-dbg - The GNU C library version 2 (debugging/profiling files). libc6-dev - The GNU C library version 2 (development files). libc6-pic - The GNU C library version 2 (PIC library) locales- Locale data files and utilities. timezones - Time zone data files and utilities. Changes: glibc-sparc (2.0.91-980111-2) unstable; urgency=low . * Workaround enabled fixing nasty `_IO_2_1_stdin_' bug * Removed warning when using with a binary compiled with glibc-2.0.90 * Added shlibs entry: `Depends: libc6 (>= 2.0.91)' Files: 277723867da1d0aeb66d4cb75295197c 5662027 libs required glibc-sparc_2.0.91-980111.orig.tar.gz c28dfd2750a3b3381a971b82d86b041d 428964 libs required glibc-sparc_2.0.91-980111-2.diff.gz 4eb449010dc209dad4f05bc2ea90cde0 738 libs required glibc-sparc_2.0.91-980111-2.dsc 48e1c85f8c64d61721541e7ab95592c8 1363216 base required libc6_2.0.91-980111-2_sparc.deb 49a6fab7ddeb0a70861c9818f9c9140c 1629866 devel standard libc6-dev_2.0.91-980111-2_sparc.deb 3fd5266a7b545d4456f6e854658a6f31 6727836 devel optional libc6-dbg_2.0.91-980111-2_sparc.deb 9750781356fec7f8855d6c8cbba1c879 2041422 devel optional libc6-pic_2.0.91-980111-2_sparc.deb 30f19359490588d70f4f14362a3466e8 1225798 admin standard locales_2.0.91-980111-2_sparc.deb 175b3f7a940698aa8d4173c9d5abc616 258466 base required timezones_2.0.91-980111-2_sparc.deb -BEGIN PGP SIGNATURE- Version: 2.6.3ia Charset: noconv iQCVAwUBNMOk1l2FmLxCq1O1AQEJWwQAn+HFP668CDkkWrRUf/8MRgqP5k85dMgH vP6/lhpc9VLNVKRn64xZGeZeebgeFsbeokXbryCOxt8o5gBErZ4NZhm4DfLKYkFx 7DO5Vjj0P1w4mYJINM814OQJULYEqLxyKCJDyTOVAVuGRl/0iL7UpRF48jhCACf3 etehaotP2Mg= =tBXM -----END PGP SIGNATURE- -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: login failure (was: Re: List of Uncompilable Sparc Packages)
On Mon, Jan 19, 1998 at 11:54:50PM +0100, Eric Delaunay wrote: > > Is it ok to libc6 to depend of base-files (>= 1.3.5) ? If so, can you add it? OK, but not on this release... I'm already uploading it. But I may upload a new one soon. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Status of the new glibc
On Mon, Jan 19, 1998 at 11:17:13AM -0500, Mark W. Eichin wrote: > > It's not incompatible with glibc-2.0; it is with older pre-2.1 > > Oh, I see, and 2.0.90 is really 2.1-- rather than 2.0++. > > > Note: I think all the programs will still work. They will > > only display 2 warnings when executing them. If you want, that > > I guess I misinterpreted -- I thought you were saying the new libc > made older programs things dump core? The warnings I can live with > (as long as they go to stderr, and even then, a way to turn them off > (temporary environment variable?) will probably make some autoconf-based > tests happier.) I'll upload it without displaying anything. I still think that all the programs will work, but be careful: I may be wrong. Anyway, binaries compiled with *this* library will probably *not* work with previous releases (I'm adding a propper `shlibs' file). -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Status of the new glibc
On 19 Jan 1998, Mark W. Eichin wrote: > Is there any reason not to change the soname, if it's that > incompatible? It's not incompatible with glibc-2.0; it is with older pre-2.1 libraries. > it's going to be *hard* to upgrade if we have to > basically redo the last month worth of bootstrapping from scratch. If > we can use a new soname, I can deal with moving forward with this > release... but if we can't do that, or the equivalent, I'd have to > recommand *not* using this libc release until it does change enough > (and stabilize) that an soname change gets made. H.J. Lu says it won't change that way again... I hope he'll be right. Note: I think all the programs will still work. They will only display 2 warnings when executing them. If you want, that warnings can be avoided... -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Status of the new glibc
I've (finally) managed to fix the problems with the new glibc; I'm testing it, and I'll upload it in one or two days... However: almost *all* of our binaries will have to be recompiled. Some vital structures have changed and all the programs segfaulted, but I managed to add a work-around in ld.so; however, I don't know if that will fix the problem with all the executables, so a warning is displayed when a binary compiled with the old libc is executed, this way: $ echo Hello, world. echo: WARNING: linked with old glibc-2.0.90. Workaround enabled echo: Symbol `_IO_2_1_stdin_' has different size in shared object, consider re-linking Hello, world. $ _ I've seen that some of you have uploaded *many* sparc packages; that's a pity, because all of them will have to be uploaded again (and most of them will need a non-maintainer version). I have mailed Ulrich to know if this will happen again with the glibc. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: login failure (was: Re: List of Uncompilable Sparc Packages)
On Sun, Jan 18, 1998 at 12:44:22AM +0100, Eric Delaunay wrote: > > I added the default /etc/nsswitch.conf coming with glibc sources and now it > works fine. > Juan, don't forget to provide this file in your next libc package. But it's already provided by the `base-files' package!: $ ls -l /etc/nsswitch.conf -rw-r--r-- 1 root root 347 Apr 19 1997 /etc/nsswitch.conf $ dpkg -S /etc/nsswitch.conf base-files: /etc/nsswitch.conf $ dpkg -l base-files Desired=Unknown/Install/Remove/Purge | Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) ||/ NameVersionDescription +++-===-==- ii base-files 1.3.5 Debian Base System Miscellaneous Files So, if you don't have it on your system, you should upgrade your `base-files' package. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
New glibc...
Well, I'm still having many problems with the latest glibc snapshots... I think I'll fix them, but it's not easy. I have a newer libc available at master.debian.org:/home/cespedes, but WARNING: It doesn't work with binaries compiled with previous glibc's. So, DON'T INSTALL IT if you have libc6 compiled binaries: they won't work. However, I would like some of you to try to help me on this issue... -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
New glibc and versioned symbols
I'm having *many* problems with the new glibc... Many of the programs linked with the current glibc in hamm don't work with the latest version I was about to upload. Example: $ rm rm: Symbol `_IO_2_1_stdin_' has different size in shared object, consider re-linking rm: too few arguments Try `rm --help' for more information. Bus error $ gzip gzip: Symbol `_IO_2_1_stdin_' has different size in shared object, consider re-linking gzip: compressed data not written to a terminal. Use -f to force compression. For help, type: gzip -h Bus error $ sync sync: Symbol `_IO_2_1_stdin_' has different size in shared object, consider re-linking Segmentation fault (well, there may be a problem with _IO_2_1_stdin_ :-)) I'm still trying to see if that symbol has changed... Any hint on how to solve this? Thanks. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
-altdev packages (was Re: Bug#16987: xpm4g: attempts libc5 build on sparc)
On 12 Jan 1998, Mark W. Eichin wrote: > I didn't realize there were that many altdev libraries. Since there > never *were* libc5 X libs, I'd just as well not build altdev ones > either -- it's not like we have a "bo" sparc release to support, > right? Unless I hear some *really* good arguments, I'm not going to > add an X altdev package, either... Maybe we should get rid of all the -altdev packages, and `altgcc'? I don't know what's the best alternative. I only thougth it was easy to build all the altdev stuff, and I did it. Comments? -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: New glibc
On 12 Jan 1998, Mark W. Eichin wrote: > As I posted earlier: > > > a major bug in libc6-dev (linux/socket.h and > > bits/socket.h have *totally conflicting* definitions of SOCK_*, MSG_*, This is very common in glibc. They want the programmer to include *only* , and not . That way, there won't be any conflict. Doesn't `strace' work without including ? Note: I can't use `#if 0', because the kernel and utilities very related to the kernel wouldn't compile. I could use `#ifndef _SYS_SOCKET_H', but IMHO that's too dirty. `strace' (and almost all the programs) should be able to compile without including anything from ``linux/''. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: pb compiling with gcc-2.7.2.3-3
On 11 Jan 1998, Mark W. Eichin wrote:
> Problem is, some programs are hard to change like that (some stuff in
> tetex-bin, for example, does that initialization for an extern in a
> library, so there isn't a main() into which to move the
> initialization...)
You can initialize variables in a library using a constructor.
Example:
==
#include
FILE * stream;
void init_function(void) __attribute__ ((constructor));
void init_function(void)
{
stream = stdin;
}
==
That way, `my_function()' is executed even before `main()'.
> Note, also, that as far as I can tell from the X build, the *only*
> platform ever to have this problem is LynxOS, and that was about as
> marginal a Unix you could get (it's a realtime unix, just this side of
> the unix vs. realtime line from vxworks...) Even if there isn't a
> standard to back me up [if there is, then this is a bug, plain and
> simple] I'd still argue that it should be fixed since it's going to
> break so much code [as we've learned on this project...]
Yes, I agree with you. We should tell this to the glibc
developers.
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
New glibc
I'm about to upload a new release of glibc, with some bugs fixed: strcmp(), and getdomainname() (and other unimplemented syscalls). Is there any other bug you would like to get fixed? -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: pb compiling with gcc-2.7.2.3-3
On Sat, 10 Jan 1998, Eric Delaunay wrote:
> I've tracked the problem down to the following source I cannot compile:
> ---cut here
> #include
> FILE *out = stdout;
> main()
> {
> fprintf( out, "hello\n" );
> }
> ---cut here
Should that program work? You are initializing a local
variable with the contents of an extern one. It's like this one:
---cut here---
extern int a;
int b = a;
main() {}
---cut here---
...which doesn't work on sparc, and doesn't work on i386.
In libc5 and glibc-2.0, std{in,out,err} are defined to... I
don't know what, but they aren't strictly `extern FILE *', as ANSI C
says they should be.
> I'd appreciate any help to determine where is the problem (gcc, libc, anything
> else, ...).
I think the problem is in that program. But I may be wrong...
I don't have the ANSI C specs here.
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: glibc-2.1 & string optimizations
On Fri, 9 Jan 1998, Eric Delaunay wrote: > I discovered a problem with string optimizations in glibc 2.1 because each > str... function is overriden by a macro which uses its arguments multiple > times (see bits/string2.h). For instance, you cannot write something like > strcmp( *++argv, "foo" ) > anymore. However, many programs are based on such constructs :-( Known bug. (at least, I found it and I mailed debian-sparc and Miguel de Icaza, without any answer yet). I think it fails only with `strcmp', and only if one of the two arguments is a string constant. Can anyone confirm that? I haven't tested all the functions, but the rest of the file /usr/include/bits/string2.h (where that macro is defined) seems OK. I've just mailed Ulrich Drepper... waiting for his answer :) BTW: I'm building a new glibc now, which will fix the getdomainname() bug (ie, "Unimplemented SPARC system call 162"), and some other things... I'll see if I can fix that too. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
ld.so: FIXED
I think I've fixed the problem with ld.so_1.9.6 and `ldd'. I'm uploading a new version now. It was an incorrect use of the `udiv' and `urem' functions on Sparc... But it still fails with some libraries... I don't know why. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: libc6 select problems?
On Sun, 4 Jan 1998, Eric Delaunay wrote: > I've put the source a long time ago on my ftp site (under > lix.polytechnique.fr:/pub/Linux/debian/sparc/bo/source/libs). > I'm now willing to work on libc5 to add some Debian specific stuff, like > the utmp wrapper and move the -dev pkg to -altdev. That's already done (-dev -> -altdev), and it works. But I think we shouldn't spend much time on libc5... there's plenty of work to do with libc6, with X... -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: libc6 select problems?
On 2 Jan 1998, Mark W. Eichin wrote: > Well, at least I've figured out that it isn't a kernel problem. I > built xntpd and rsync using libc5-altdev and alt-gcc, and aside from a > typo (already reported) in the alt sched.h, they both work fine. It's a kernel problem. I'll explain myself: There are two syscalls for select: $ grep select /usr/include/asm/unistd.h #define __NR_select 93 /* Common */ #define __NR__newselect 230 /* Linux Specific */ `libc5' uses the first one (93), but `libc6' uses the second (230). Both of them should work OK, but there's a little bug in the kernel sources which makes it execute `llseek' instead of `select' when the second syscall is used. So, `select(n,...)' is interpreted by the kernel as `llseek(n,...)'. `llseek' returns EBADFD if `n' isn't a valid file descriptor (and this is normally the case when we use select). See my previous mail to debian-devel for a kernel patch. > I'm not having much luck in figuring out where in glibc-sparc select > itself is built, though (the Makefiles are a little too magic.) I'm > about to just build it and read the logs and see hack on it from that > direction... Yes, it's real ugly... it's defined from sysdeps/unix/Makefile, looking at file syscalls.list which is in sysdeps/unix/sysv/linux. It basically translates the call directly to a system call. > Is anyone working on gdb, or strace? Looks like we're stuck at the > "printf" level of debugging right now, and that's kind of > sad... strace needs actual porting (I got as far a build that did a > fake_exec and ran the program, but didn't see any *real* syscalls), > gdb uses the bfd it ships with, and that would need to be updated to a > current sparc-linux-aware libbfd. It shouldn't be too hard to port, > but still... I'm working on `strace'... there is a working version (in .rpm format) in ftp.nuclecu.unam.mx:/pub/Linux/Sparc-miguel (well, I cannot remember exactly the path, but it was something like that). It compiles and runs OK as libc5... I'm working on porting it to libc6. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: libc6 select problems?
On 3 Jan 1998, Mark W. Eichin wrote: > Now at least I've got someplace else to look. Question: where are the > sources to the *libc5* that we're using? It's 5.3.12-mumble, not the > debian standard 5.4.x, for which sources are on the mirrors, which > doesn't have *any* sparc support that I can find... I think they are exactly the sources from RedHat, which have a lot of patches from David S. Miller and Miguel de Icaza. I have never built it (I think Davide Barbieri and Eric Delaunay did). -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
libc6 select problem SOLVED
I think I've discovered the problem... I have two Sparcs handy, with exactly the same config, and the same packages installed; one of them showed the problem but the other didn't. The only difference between them is the kernel, so I thought it might be kernel related. I've investigated a bit, and I think this patch to the Linux Kernel should fix this problem: == --- linux/arch/sparc/kernel/systbls.S~ Sun Dec 14 19:37:14 1997 +++ linux/arch/sparc/kernel/systbls.S Sun Jan 4 00:30:16 1998 @@ -122,7 +122,7 @@ .long C_LABEL(sys_create_module), C_LABEL(sys_delete_module) .long C_LABEL(sys_get_kernel_syms), C_LABEL(sys_getpgid), C_LABEL(sys_bdflush) .long C_LABEL(sys_sysfs), C_LABEL(sys_nis_syscall), C_LABEL(sys_setfsuid) - .long C_LABEL(sys_setfsgid), C_LABEL(sys_llseek), C_LABEL(sys_time) + .long C_LABEL(sys_setfsgid), C_LABEL(sys_select), C_LABEL(sys_time) .long C_LABEL(sys_nis_syscall), C_LABEL(sys_stime), C_LABEL(sys_nis_syscall) .long C_LABEL(sys_nis_syscall), C_LABEL(sys_llseek) /* "We are the Knights of the Forest of Ni!!" */ == It's against sparclinux-2.0.32-971208 from vger, but it should apply cleanly to almost any modern version of the Linux kernel. Please tell me if this solves your problems... it worked for me. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Bug in glibc-2.1 for Sparc
Hi. I've discovered a nasty bug in glibc 2.1: arguments to
strcmp() (and possibly other functions) are evaluated more than once:
$ cat kk.c
#include
#include
static char * kk(void)
{
printf("Got here!\n");
return "yeah";
}
main()
{
strcmp(kk(), "yeah");
}
$ gcc -O2 -o kk kk.c
$ ./kk
Got here!
Got here!
Got here!
$ _
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: got new hardware
On Tue, 16 Dec 1997, Eric Delaunay wrote: > PS: Did you see an announce about glibc-2.0 available for sparc? > Anyone tried it ? > Can we adopt it to replace our glibc snapshot ? I've seen it, but I think we shouldn't switch back to glibc-2.0 now that 2.1 works (more or less). IIRC, Ulrich didn't want to include that patches in the official 2.0, and he is going to release an alpha version of glibc 2.1 soon. Also, there are now many packages compiled against glibc 2.1 which probably wouldn't work with glibc 2.0. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: possible bug in sparc libc6-dev
On Thu, 11 Dec 1997, Joey Hess wrote:
> phantglobs.h:67: macro `strcpy' used without args
> phantglobs.h:68: macro `strncpy' used without args
> make[2]: *** [main.o] Error 1
Yes, I've had that problem with some programs, too.
> Fixes for this include -
>
> 1 commenting out the function prototypes (shouldn't hurt anything), but will
> probably impact a fair number of packages.
> 2 compiling without -O2, which disables the macros.
> 3 fixing the include files somehow (one way is to just remove the macros, of
> course, but that may slow things down, they appear to be optimized).
>
> I guess option 3 is best, if it's doable.
I think we should do 1). A program should not have any
function prototype which is in the libc; it should just include
instead. We had a similar problem with `errno', do you remember?
And... we have a similar problem with many other functions.
For example, this program fails to compile with glibc-2.0 (the current
one used with i386), even without -O2:
--
#include
extern int isalnum(int);
int main() {}
--
So I think we can't do 2) or 3). 1) is the best choice. And
I think you should would submit a bug report against the package with
that prototypes.
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: Makin' packages...
On Tue, 9 Dec 1997, Michael Shuey wrote:
> 1) Where should I upload these .deb files? lix? master?
If you are an official Debian Developer, I think you should
upload them to master. If not, I would consider the posibility of
becoming a developer. I don't want two distributions: one in lix
and other in master (sorry Eric; I think your machine has been very
useful, but we should upload things to master).
> 2) What do I do about packages that use altgcc? I assume that anything I
> build should be built against libc6 (correct me if I'm wrong), but what
> do I do about things (like the latest version of bash) that build some
> libraries with libc5? Currently I'm just not building the libc5 portion.
> Any suggestions (other than making altgcc work; you guys might be able
> to do that but I can't :-)
I'm working on altgcc right now. If everything goes OK, I
may have a working version this week.
> 3) I want X to work. Is anyone out there working on Xsun and, if so, how
> could I help? If not, anyone have any pointers for someone who's never
> built a large Debian package but would like to try?
Pointers? The Debian Policy and the Debian Packaging Manual,
available from ftp://ftp.debian.org/debian/doc/package-developer.
I'm afraid I can't help you with X, but you could have a look
at the RedHat sources for that.
> 4) I'm starting to build up a small list of very minor problems with source
> packages as I build things. Should I, at some point, file all these as
> bug reports?
If they are minor building problems, yes, I think you should
submit bug reports.
> By the way, nice job Juan Cespedes. Those .debs of libc6 & binutils seem to
> be working fine. :-)
Thank you :) Please tell me any problem you could find with
those packages.
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: New glibc and new binutils
On Tue, 2 Dec 1997, Johnie Ingram wrote: > "Juan" == Juan Cespedes <[EMAIL PROTECTED]> writes: > > Juan> I've just uploaded a new release of binutils (2.8.1.0.17) and > Juan> libc6 (971126) to master. The former will be in the FTP area in > > I see the source code there, but not the binary packages for gcc and > libc6. Are they also available from an unofficial site? ftp://ftp.de.debian.org/pub/debian-incoming/glibc-sparc_2.0.90-971126-1.changes ftp://ftp.de.debian.org/pub/debian-incoming/glibc-sparc_2.0.90-971126-1.diff.gz ftp://ftp.de.debian.org/pub/debian-incoming/glibc-sparc_2.0.90-971126-1.dsc ftp://ftp.de.debian.org/pub/debian-incoming/glibc-sparc_2.0.90-971126.orig.tar.gz ftp://ftp.de.debian.org/pub/debian-incoming/libc6-dbg_2.0.90-971126-1_sparc.deb ftp://ftp.de.debian.org/pub/debian-incoming/libc6-dev_2.0.90-971126-1_sparc.deb ftp://ftp.de.debian.org/pub/debian-incoming/libc6-pic_2.0.90-971126-1_sparc.deb ftp://ftp.de.debian.org/pub/debian-incoming/libc6_2.0.90-971126-1_sparc.deb ftp://ftp.de.debian.org/pub/debian-incoming/timezones_2.0.90-971126-1_sparc.deb ftp://ftp.de.debian.org/pub/debian-incoming/binutils_2.8.1.0.17-0.1_sparc.changes ftp://ftp.de.debian.org/pub/debian-incoming/binutils_2.8.1.0.17-0.1_sparc.deb ftp://ftp.de.debian.org/pub/debian-incoming/libbfd2.8.1.0.17-dev_2.8.1.0.17-0.1_sparc.deb ftp://ftp.de.debian.org/pub/debian-incoming/libbfd2.8.1.0.17_2.8.1.0.17-0.1_sparc.deb (ftp.de.debian.org:/pub/debian-incoming/ is a hourly mirror of master.debian.org:/home/Debian/ftp/private/project/Incoming) They are also available from ftp://ftp.etsit.upm.es/pub/Linux/local/debian -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Incomplete upload found in Debian upload queue
Sorry for all that junk: it was me. I built the new glibc using the address [EMAIL PROTECTED]' as the maintainer address, and the Upload Queue in ftp.uni-erlangen.de seems to be very verbose :) Well, I couldn't upload it to erlangen, but I'm uploading it to master now. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
New glibc and new binutils
I've just uploaded a new release of binutils (2.8.1.0.17) and libc6 (971126) to master. The former will be in the FTP area in a few hours; the latter will probably be in a few days, because it's a new package. It will be available in the Incoming/ until then, which is mirrored in some sites, such as ftp.de.debian.org:/pub/debian-incoming Please tell me if it works for you; it certainly solves the sockets problem for me. Happy Hacking, -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
binutils-2.8.1.0.17
Has anyone been able to compile binutils 2.8.1.0.17 for Sparc? If haven't been able to do it yet, and I need it to try latest versions of glibc. Any hints? Thank you. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
binutils-2.8.1.0.17 ??
Trying to build the latest snapshot of glibc, I'm having some problems with versioned symbols. In the glibc sources it's documented the need of binutils 2.8.1.0.17, which I don't have. I've mailed Ulrich Drepper and H.J. Lu asking for it... does anyone know if binutils_2.8.1.0.17 exist, and where can I get it? Thanks for your help... -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: glibc and the sockets problem
On Wed, 19 Nov 1997, Juan Cespedes wrote: > Apparently, RedHat is going to release > version 5.0, which comes with glibc 2 for Sparc. > They have patches for the socket problem in their > .src.rpm; IIRC, it's in > ftp://ftp.redhat.com/pub/redhat/devel/rhl/SRPMS/glibc-* I tried to apply the socket patch to a recent version of glibc, but it has already been included :) I'm building it now... if everything goes OK, it will finish the compilation in about 5 or 6 hours, and I may be uploading it to master tomorrow, after testing it a bit. BTW: I think it's better if we upload everything to master instead of some things to master, and some things to lix. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
glibc and the sockets problem
Apparently, RedHat is going to release version 5.0, which comes with glibc 2 for Sparc. They have patches for the socket problem in their .src.rpm; IIRC, it's in ftp://ftp.redhat.com/pub/redhat/devel/rhl/SRPMS/glibc-* I'll try to build a new version of glibc with it, but you are free to do it before me :) I would like to release Debian 2.0 for Sparc; maybe it's not too late to do it after all. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Problems executing /lib/ld-linux.so.2
Hi. I have been trying to compile a new glibc many, many times, wasting a lot of time :), and it always failed in a point where the dynamic linker (ld-linux.so.2) was executed with one program as the argument. IMHO, that is one of the best things the new libc has, but it never worked for me in sparc. Example: [i386] $ /lib/ld-linux.so.2 /bin/ls -al total 2 drwxr-xr-x 2 root root 1024 Nov 17 11:43 . drwxr-xr-x 8 root root 1024 Nov 17 11:43 .. [sparc] $ /lib/ld-linux.so.2 /bin/ls -al [it hangs] Well, it sometimes hangs, but sometimes dies with SIGSEGV, sometimes with SIGBUS, sometimes with SIGILL... Stracing it, this is the last syscall: [000123b4] mmap(0x1, 24576, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_FIXED, 3, 0) = 0x1 Can you see what happens? I didn't see it until two weeks trying to know what happened. Hint: it is mmap'ing something, _overwriting_ the code that is being executed! It happens because the text addr in elf32_sparc starts at offset 0x1, overwriting part of the loader. In i386 systems, it starts at 0x08048000, so there isn't any problem here. The default text start offset is defined in file ld/emulparams/elf32_sparc.sh, in the binutils source. Does anyone know if it can be changed? If not, maybe we'll have to change the text address of ld-linux.so.2 in glibc. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Does sparc need/want libc5 (i.e. non-g) versions of libraries?
On 4 Nov 1997, James Troup wrote: > I'm finally getting round to updating my packages which include 4 > shared libraries and I need to know if I should exclude the sparc from > the building of the libc5 versions of the libraries? sparc has a working libc5, but we still don't have altgcc, so it's difficult to build both libc5 and libc6 versions of the libraries. But I hope we'll have altgcc some time soon :) So, please, don't treat sparc specially at all. Thanks. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Unidentified subject!
On Tue, 4 Nov 1997, Alexey Pialkin wrote: > Yesteday i'v installed Debian on my lovely SPARC Classic and during > installation i have a bunch of problems : > > 0) fdisk > 1) /dev/nfs > 2) /dev/loop & MSDOS fs > > Yes,these problems are well known(& known for a long time), so may be it shoud > be fixed in standard distribution ? Before fixing things as the booting phase, we need a working libc6... and it's not easy! But I hope I will have a modern one working in one week or so. > Now about packages. First of all - where shoud i get a most recently version > of them ? As far I undestand this place is not ftp.debian.org :) just because > i'v got some packages from it depending on libc6 and no libc6 at all :)). Try `ftp://lix.polytechnique.fr/pub/Linux/debian/sparc'. > The second - what are needing/whanting packages ? I am ready to take care > about some of them.. Well, all of them :) The best thing would be to start with the required packages, then the important ones, and so on... Thanks for your interest in Debian/Sparc! -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: ld seg faults
On Tue, 28 Oct 1997, Davide Barbieri wrote: > Ciao, > ld segfaults very often. Too often. > This affects results from './configure', and prevents > a lot of programs to be compiled fine. > > my ldso is 1.9.2-1 `ld' is from binutils. Can you tell us which binutils are you using, and an example where `ld' fails? BTW, I'm not able to reproduce your problem compiling glibc, but I'm trying to compile a newer one (971026), with gcc_2.7.2.3-2. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: my attempt to compile glibc
On Mon, 27 Oct 1997, Davide Barbieri wrote: > gcc: Internal compiler error: program cc1 got fatal signal 11 > make[3]: *** [/home/paci/libc-970928/builddir/math/s_ldexp.os] Error 1 > make[3]: Leaving directory `/home/paci/libc-970928/math' > make[2]: *** [math/subdir_lib] Error 2 > make[2]: Leaving directory `/home/paci/libc-970928' > make[1]: *** [all] Error 2 > make[1]: Leaving directory `/home/paci/libc-970928/builddir' > make: *** [build] Error 2 Signal 11? Then, there isn't any problem with that glibc sources. signal 11 in gcc means one of these things: 1. Bug in gcc 2. Bug in the libc you have installed 3. Bug in the kernel 4. Hardware problem > I have tried, also, the old gcc 2.7.1. With the same results? Failing exactly in the same point? This is important. > Sorry, but I can't compile glibc. I'm having some problems, too, but mine are regarding to newest binutils... -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: more glibc problems...
On Tue, 14 Oct 1997, Eric Delaunay wrote: > I would want to know why there are no crtbegin[S].o & crtend[S].o for > sparc32 architecture? These files are needed to compile with ctor/dtor. > It seems they are only available in alpha & sparc64 architecture! > Where are they now? They are now provided by gcc. Latest gcc for Debian/Sparc in master.debian.org, version 2.7.2.3-2, already have them, and they work ok. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
glibc: Bug in libm.so? (was Re: new packages...)
On Fri, 17 Oct 1997, Miguel de Icaza wrote: > > BTW, there is a bug in libm.so.6: > > > > $ nm --dynamic /lib/libm.so.6 | head > >U .div > >U .rem > >U .umul > > A GLIBC_2.0 > > Why should be this the case? Those should be defined in libc. Well, I don't know why, but they aren't defined in libc.so or in ld-linux, only in libc.a: $ nm --dynamic /lib/ld-linux.so.2 | grep \\.div $ nm --dynamic /lib/libc.so.6 | grep \\.div $ nm --defined /usr/lib/libc.a | grep -B2 \\.div sdiv.o: a *ABS* T .div So I still think this is a bug. libm.so should not have an undefined symbol which is not defined in ld-linux.so.2 or libc.so. PS: This is with 970928; I'm just now trying to build 971001, the latest release (AFAIK). BTW, where are now the latest snapshots? alpha.gnu.ai.mit.edu does not seem to work... -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
glibc: Bug in include file?
/usr/include/bits/ioctls.h are different in i386 and Sparc;
but the one from Sparc doesn't work: it refers to
which doesn't exist.
I'm using the one from i386 without troubles...
These are the differences between the i386 and the Sparc
version:
[EMAIL PROTECTED]:/debian/glibc-2.0.90-970928/sysdeps/unix/sysv/linux# diff -u
bits/ioctls.h sparc/bits/ioctls.h
--- bits/ioctls.h Sat Jun 21 06:48:56 1997
+++ sparc/bits/ioctls.h Thu Sep 11 20:21:46 1997
@@ -1,4 +1,4 @@
-/* Copyright (C) 1996 Free Software Foundation, Inc.
+/* Copyright (C) 1996, 1997 Free Software Foundation, Inc.
This file is part of the GNU C Library.
The GNU C Library is free software; you can redistribute it and/or
@@ -25,6 +25,19 @@
/* Use the definitions from the kernel header files. */
#include
+#include
+
+/* Oh well, this is necessary since the kernel data structure is
+ different from the user-level version. */
+#undef TCGETS
+#undef TCSETS
+#undef TCSETSW
+#undef TCSETSF
+#define TCGETS _IOR ('T', 8, struct __kernel_termios)
+#define TCSETS _IOW ('T', 9, struct __kernel_termios)
+#define TCSETSW_IOW ('T', 10, struct __kernel_termios)
+#define TCSETSF_IOW ('T', 11, struct __kernel_termios)
+
#include
#endif /* bits/ioctls.h */
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: new packages...
On Thu, 16 Oct 1997, Davide Barbieri wrote: > pgp, rsa and ssh are in non-us section; master can hold them? Oops, sorry, I didn't noticed it... No, they have to be uploaded to nonus.debian.org, using anonymous ftp, to directory /debian-non-US/Incoming. AFAIK, this is the only way; I think chiark and erlangen can't be used to upload non-US packages. -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: new packages...
On Thu, 16 Oct 1997, Davide Barbieri wrote: > Now I'm creating a new pgp debian package, signed with my pgp key. > Then I have to upload it to chiark, and not to master, right? Why not to master? You can upload it to master, to chiark, or to erlangen. If you upload to master, you'll have to do it using ssh, rsh or user ftp; if you upload to chiark or to erlangen, it's enough to do anonymous ftp (I always upload to erlangen; the others are too slow from here). -- Juan Cespedes -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: new packages...
On Thu, 16 Oct 1997, Davide Barbieri wrote:
> My pgp key was generated on a i386; now I need to port on sparc
Hmmm... aren't they binary-compatible?
> porting the public key is easy (pgp -kxa on i386, ftp the ascii key,
> pgp -ka on sparc); but I need also the secret key...
On the i386, type:
$ pgp -kxa [EMAIL PROTECTED] whatever ~/.pgp/secring.pgp
This will save a copy or your PGP secret key, coded in ASCII,
in the file `whatever.asc'.
Move the file `whatever.asc' to the Sparc, and then, in the Sparc:
$ pgp -ka whatever.asc ~/.pgp/secring.pgp
This will create the file ~/.pgp/secring.pgp with your secret key.
Gook luck.
--
Juan Cespedes
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
