Hurd Traffic #69 For 29 Nov 2000

Editor: Zack Brown

By Paul Emsley  and  Zack Brown

Mach 4 ( | Hurd Servers ( | Debian Hurd Home ( | Debian Hurd FAQ ( | debian-hurd List Archives ( | bug-hurd List Archives ( | Hurd Reference Manual ( | Hurd Installation Guide ( | Cross-Compiling GNUMach ( | Hurd Hardware Compatibility Guide (

Table Of Contents


Want to help write KC Debian Hurd? See the KC Authorship page (../author.html) the KC Debian Hurd homepage (index.html) , and the Thread Summary FAQ (../summaryfaq.html) . Send any questions to the KCDevel mailing list. (

Mailing List Stats For This Week

We looked at 145 posts in 651K.

There were 53 different contributors. 22 posted more than once. 18 posted last week too.

The top posters of the week were:


1. X With Microsoft Protocol
13 Nov 2000 - 20 Nov 2000 (6 posts) Archive Link: "X with Protocol != "osmouse" anyone?"
Summary By Paul Emsley
People: David CoquilPaul EmsleyMarcus Brinkmann

Marcus Brinkmann wanted to know whether X worked with mouse protocols other than "osmouse", particularly "IntelliMouse" and "Microsoft", since UCHIYAMA Yasushi has said (over a year ago) that they had been working.

David Coquil tried "Microsoft" but found that it didn't work, Marcus replied that he needed to use the latest gnumach. David did so and confirmed that "Microsoft" was still functioning.

(ed. [Paul Emsley] For now...)



2. Status Of PPP
20 Nov 2000 (3 posts) Archive Link: "Offer for help"
Summary By Zack Brown
People: Daniel E. BaumannNeal H Walfield

Robert Marlow had a three month vacation from university, and wanted to hack on the Hurd. He remembered that PPP had needed work awhile before, and asked if that would still be a good place to start. Daniel E. Baumann replied, "I am currently porting BSD user space PPP. I should be done by the end of the week." And Neal H Walfield gave links to an explanation of implementation details ( and a basic plan ( from the list archives.


3. Debian's KDE Maintainer To Work On Hurd Port
21 Nov 2000 - 22 Nov 2000 (4 posts) Archive Link: "question abou KDE,icewm"
Summary By Zack Brown
People: Marcus BrinkmannIvan E. Moore IIMartin StenzelPetros Sidiropoulos

Petros Sidiropoulos asked if it was possible to run KDE on the Hurd, and Marcus Brinkmann replied, "Nobody knows. It should be, but there might be any number of porting problems." He suggested that Petros get the source packages, compile them, test it out, and report his findings. Ivan E. Moore II, as maintainer of most of the Debian packages involved, also added:

please clue me in on any problems as well (as maintainer of most of those packages).

Once I get a little more free time I'll be setting up a hurd box and will start working on this...I haven't touched hurd in over a year so there will be a learning curve again. :)

Martin Stenzel also said:

I successfully compiled and tested the qt libraries (version 2.2.1), which KDE depends on (get it at I did not compile KDE ( itself, though.

What you certainly need is a lot of memory (128 MB, part of it may be virtual) and disk space (KDE is a memory and disk hog).

Since KDE incorportes system tools I go with Marcus in that there is most probably some porting work to be done.


4. X Twice
21 Nov 2000 - 23 Nov 2000 (4 posts) Archive Link: "Can't start X twice in the same Hurd session"
Summary By Paul Emsley
People: David CoquilMarcus Brinkmann

David Coquil noted that when he killed the (3.3.6) X server, he couldn't start it again - it was not possible to connect to an X socket (_X11TransSocketUNIXConnect: Can't connect:). Marcus Brinkmann told him that something had gotten stuck in pfinet, and advised David to kill pfinet and restart X. David tried this and found that it fixed the problem; but the underlying issue of why pfinet got in a tangle still remains unresolved.


5. Hurd Parallelization
22 Nov 2000 (5 posts) Archive Link: "Perl 5.6"
Summary By Paul Emsley
People: Neal WalfieldRoland McGrathNeil WalfieldPaul Emsley

Robin Putters announced that he had successfully compiled Perl 5.6, and that in order to get the .deb files, folks should add the following to sources.list:

deb unstable main

He explained that he'd compiled the package (taking 4 hours) on a Dual PIII 733 machine. Neil Walfield noted: " There is no smp support under the hurd. Did you cross compile? " .

Roland McGrath clarified the SMP issue with the Hurd, saying, "The Hurd has (in theory) excellent support for multiprocessors, i.e. it is highly multithreaded. However, neither gnumach nor oskit-mach currently has working SMP support. The Mach code base supports SMP in general, and there have been working SMP versions in the past; there should not be an unreasonable amount of work in making SMP work in either kernel."

(ed. [Paul Emsley] Sound like an excellent project. Who knows? It could be a one-liner.)


6. 4th Set Of Permission Bits
24 Nov 2000 - 25 Nov 2000 (6 posts) Archive Link: "4th set of permission bits?"
Summary By Paul Emsley
People: Marcus Brinkmann

Robert Marlow was curious if anyone was working on the 4th set of permission bits provided by the Hurd. Marcus Brinkmann replied that all the Hurd implementation was in place, but that things like chmod and other fileutil programs had to be extended to take advantage of these bits. He added that it was still undecided what the default should be for these bits: " active or use "other" bits " ?

Robert replied, wondering as to the reasons against having the 4th set of permission bits active. Marcus gave two reasons, revealing an extra bit of information:

1. No tools support manipulation of the fourth set of permission bits.

2. There is one more bit. If 0, for not-logged-in user the permission set of "other" is used. If 1, the fourth set is used. It is undecided if the this should probably be reversed (so the permissions for not logged in users are empty by default).

He went on to say that the main difficulty was that the bit manipulation had not been added to the fileutil programs and that the extra bit should toggle the behaviour of the 4th permission bits.


7. PostgreSQL: Not There Yet
24 Nov 2000 - 25 Nov 2000 (2 posts) Archive Link: "Compiling PostgreSQL and System V IPC"
Summary By Zack Brown
People: Dibyendu MajumdarMarcus Brinkmann

Dibyendu Majumdar tried to build PostgreSQL version 7.0.2 on the Hurd, and reported, "The build failed ultimately because the System V IPC functions aren't supported on Hurd." Marcus Brinkmann gave a pointer to a bug report ( and said, "I don't know what good alternatives for postgresql exist in the Hurd."


8. Free Software Developer's Meeting
25 Nov 2000 (1 post) Archive Link: "Off topic? Free software developers meeting"
Summary By Zack Brown
Topics: Emulators: plex86
People: Raphael Bauduin

Raphael Bauduin announced:

I'm organising an open source and free software developers meeting, and it would be interesting to have some of you present (there are gGNU project I think should be represented as GPG, Hurd, and others I forgot but that are as important as those two....). It will happen in Brussels the 3rd and 4th february 2001 just after the linux-expo in Paris, so if you are there, it shouldn't be a problem to come over to Brussels and if you don't go to Paris, we would be happy to have you in Brussels anyway :-)

A very basic website is at with a mailing list, and an apply form (this is not organised by a corporation, the .com was easier to remember for my friends.... :-)

If you need more info, you can contact me at ( so I can mail you a document explaining what will hapen and who's coming (confirmed are: Jeremy Allison-Samba, David Faure-KDE, Kevin Lawton-Plex86,Mathieu LAcage-Gnome, Peter De Schrijver-Embedded Linux, Ami Gangula-Midgard, Phil Thomson-Veepee, fyodor told me he would come too)

There was no reply.


9. OpenLDAP For The Hurd
25 Nov 2000 - 26 Nov 2000 (2 posts) Archive Link: "OpenLDAP patch"
Summary By Zack Brown
People: Igor Khavkine

Igor Khavkine announced:

Hi, I just happen to have some time between assignments and exams, and so I finished my work on OpenLDAP. The patch is attached and you can also find it on my new site

Just put the patch into the debian/patches directory, the only glitch is that you have to regenerate the configure script with autoconf.

A couple of notes. OpenLDAP includes support for CThreads however it looks for the header in <mach/cthreads.h> while gnumach puts it in <cthreads.h>. Shouldn't we follow the general Mach convention and put it in <mach/cthreads.h> too? Also, is it usually necessary to free dynamically allocated memory right before an exit() statement? On UNIX I'd say no, but apparently OpenLDAP should also work on Windows, whose memory management is notoriously poor.

Half a day later he added, "openldap needs libsasl7-dev to compile, that package wasn't in the Debian archive so I had to patch it too."







We Hope You Enjoy Hurd Traffic

Kernel Traffic is grateful to be developed on a computer donated by Professor Greg Benson and Professor Allan Cruse in the Department of Computer Science at the University of San Francisco. This is the same department that invented FlashMob Computing. Kernel Traffic is hosted by the generous folks at All pages on this site are copyright their original authors, and distributed under the terms of the GNU General Public License, version 2.0.