Re: Books on File Systems and File System Programming

2009-08-13 Thread Joerg Sonnenberger
On Thu, Aug 13, 2009 at 10:56:26PM +0530, Siju George wrote:
 Could Some one recommend good books on  File Systems and File System
 Programming Please?

The only real book available is the BFS one. Google a bit, there is a
PDF version of it somewhere. If you want it on dead tree, ask a local
printer.

Joerg


Re: Call for help with firefox 3.5

2009-07-02 Thread Joerg Sonnenberger
On Thu, Jul 02, 2009 at 12:23:51PM +0300, Hasso Tepper wrote:
 My guess is that it's related OSS backend because nobody actually uses it 
 in the upstream (I don't think that any OS relevant today except FreeBSD 
 and DragonFly use OSS by default at all). There are several commits into 
 other backends to solve sync problems, but none into OSS code. Therefore 
 call for help - if you can and understand what OSS is (I don't ;), please 
 look at it.

NetBSD and OpenBSD use either the SunOS interface or OSS compatibility.
I think the situation for Solaris is somewhat similar... In other words:
only Mac OS X, Windows and Linux use something other than OSS :)

Joerg


Re: Call for help with firefox 3.5

2009-07-02 Thread Joerg Sonnenberger
On Thu, Jul 02, 2009 at 11:08:38AM -0700, Matthew Dillon wrote:
 Where is the wip/firefox?  I've been using:
 
 cvs -d anon...@anoncvs.netbsd.org:/cvsroot update -Pd pkgsrc

Externally on pkgsrc-wip.sf.net.

Joerg


Re: Instant crash with Linux OpenOffice

2009-06-02 Thread Joerg Sonnenberger
On Tue, Jun 02, 2009 at 11:46:53PM +0200, Francois Tigeot wrote:
 Pure madness.

Come on, it has gotten a lot better since the OpenOffice 1.x days.
I think they no longer use Pascal in the build system, not sure about
Java and Python though.

Joerg


Re: pkg_dry on DragonFlyBSD

2009-05-09 Thread Joerg Sonnenberger
On Fri, May 08, 2009 at 10:27:16PM +0100, Steve O'Hara-Smith wrote:
   There may well come a time for either of these where there are two
 incompatible versions extant supporting two actively used versions of their
 client (think around a major version bump). It would be much easier to
 maintain multiple active versions in the pkgsrc framework than in the
 kernel source tree.

What is more likely? The kernel interface changing enough to require
fixing the module or an incompatible version to appear?

I am against pushing kernel modules into pkgsrc because I strongly
believe based on experience from watching different kernels and external
modules for years, that the former happens a lot more often.

Why does it make it better to keep them in base? Because that way the
exposure is much greater. Heck, if it is part of the normal buildworld,
people do care about breaking it...

Joerg


Re: HPET cputimer

2009-04-30 Thread Joerg Sonnenberger
On Thu, Apr 30, 2009 at 06:55:47PM +0800, Sepherosa Ziehau wrote:
 Looks like yours doesn't have the resource template.  May be BIOS
 disabled it by default?

I'd bet it is not using the WinXP modus...

Joerg


Re: Can't start mysqld

2009-04-22 Thread Joerg Sonnenberger
On Wed, Apr 22, 2009 at 08:29:46PM +0100, Colin Adams wrote:
 I can access mysql if i start it with the --skip-grant-tables option,
 and then i can set the root user's password, but I still can't login
 when i restart it.

Did you even bother to follow the instructions from installation method?
Consider running pkg_info -D mysql-server again...

Joerg


Re: Consequences of major libc changes

2009-04-13 Thread Joerg Sonnenberger
On Mon, Apr 13, 2009 at 03:28:23PM +0300, Hasso Tepper wrote:
 The problem is that there is too many packages with common problem (for 
 DragonFly) - they don't include unistd.h, but compile regardless of that 
 on Linux, NetBSD and FreeBSD, seems. Many of such packages are C++. Maybe 
 there is something wrong with our compiler? Examples of such packages are 
 ham/fldigi, x11/nucleo and editors/notecase.

I don't think it is the compiler. I also can't find any other *system*
header on NetBSD that pulls in unistd.h (except getopt.h, but that does
so on DragonFly as well).

Joerg


Re: Consequences of major libc changes

2009-04-13 Thread Joerg Sonnenberger
On Mon, Apr 13, 2009 at 04:57:55PM +0200, Erik Wikstr?m wrote:
 On 2009-04-13 15:10, Joerg Sonnenberger wrote:
  On Mon, Apr 13, 2009 at 03:28:23PM +0300, Hasso Tepper wrote:
  The problem is that there is too many packages with common problem (for 
  DragonFly) - they don't include unistd.h, but compile regardless of that 
  on Linux, NetBSD and FreeBSD, seems. Many of such packages are C++. Maybe 
  there is something wrong with our compiler? Examples of such packages are 
  ham/fldigi, x11/nucleo and editors/notecase.
  
  I don't think it is the compiler. I also can't find any other *system*
  header on NetBSD that pulls in unistd.h (except getopt.h, but that does
  so on DragonFly as well).
 
 What I can't understand is why C++ programs would have problems with
 unistd.h more than C programs. Might it be that one of the C++ headers
 pulls it in on Linux and NetBSD but not on DragonFly (they should not
 since unistd.h is not part of ISO C or C++)?

Missing prototypes are generally ignored in C mode.

As I said, for the system headers I can't find a difference, which is
what is puzzling me.

Joerg


Re: HEADS UP: iwi(4) firmware upgraded.

2009-03-03 Thread Joerg Sonnenberger
On Tue, Mar 03, 2009 at 06:58:37PM +0800, Sepherosa Ziehau wrote:
 Thanks to Johannes Hofmann's work, iwi(4)'s firmware is upgraded to 3.0.

sysutils/iwi-firmware3?

Joerg


Re: The state of DragonFly and pkgsrc 2008Q4

2009-01-03 Thread Joerg Sonnenberger
On Sat, Jan 03, 2009 at 03:35:55PM +0100, Simon 'corecode' Schubert wrote:
 -100, neg, horrible solution.  can't let changes in world be dictated
 by whatever in pkgsrc.

Right, just lets randomly change ABIs of system libraries. Oh wait.

Joerg


Re: My personal pkgsrc FAQ

2008-12-17 Thread Joerg Sonnenberger
On Wed, Dec 17, 2008 at 09:07:23AM +0100, Matthias Schmidt wrote:
 - IIRC it should be possible to build pkgsrc packages in parallel.  Does that
   mode works well?  If yes, we could distribute the builds between multiple
   machines to reduce the time needed to run a full build.  And yes, a have a
   machine to offer :)

It works well, but if you want to distribute over a non-local network
you get some interesting issues with redistribution of the packages
during the build. I don't think it is worth the effort.

 - How about renting a colocation machine?  Renting a fast machine with a fast
   traffic flat is not that expensive nowadays.  We even have some SoC money
   left.  Donating some money is not a big problem either.

Last time I check, SMP scalability in fork/exec heavy work load was not
that well, so it was more effective to use vmware than native SMP.
Otherwise it can bring down the build time a lot.

Joerg


Re: Question about pkgsrc installation dir

2008-12-14 Thread Joerg Sonnenberger
On Sun, Dec 14, 2008 at 11:27:40PM +1100, Petr Janda wrote:
 Why are we by default installing into /usr/pkg as opposed 
 to standard /usr/local for 3rd party apps. This always bites me when Im 
 installing manually various things that are not in pkgsrc that only look 
 at /usr/local.

Because /usr/local is for the administrator to install its own stuff.
You should strongly avoid mixing managed and unmanaged files, so it
would mean either installing manual stuff somewhere else or put the
managed things somewhere else. The pkgsrc approach is the latter as sane
default.

Joerg


Re: Serious Postfix weirdness

2008-11-15 Thread Joerg Sonnenberger
On Sat, Nov 15, 2008 at 07:21:19PM +1100, Petr Janda wrote:
 How is it possible that the sender uses wscale 2^7 and the receiver 2^0? Is 
 this a problem?

One side says it can do window scaling and the other can't. That is not
necessarily a problem, it can happen depending on the maximum TCP window
sizes as the BSD stack will not advertise window scaling if it doesn't
need it (IIRC). Do you have a firewall in between? At least old PF
(not sure which is in src right now) had issues with window scaling.

Joerg


Re: Serious Postfix weirdness

2008-11-14 Thread Joerg Sonnenberger
On Fri, Nov 14, 2008 at 08:15:04PM +1100, Petr Janda wrote:
 Supposedly the problem here is that the sending machine has got a firewall 
 in front of it thats blocking ICMP MUST FRAGMENT.

Is net.inet.tcp.path_mtu_discovery=1?

Joerg


Re: Serious Postfix weirdness

2008-11-14 Thread Joerg Sonnenberger
On Fri, Nov 14, 2008 at 11:05:00PM +0200, Jordan Gordeev wrote:
 Is there a technical reason (e.g. related to where the Path MTU is  
 stored), for having it off till now?

Stupid network admistrators that consider all ICMP traffic evil and
block it. But IMO it should be active by default.

Joerg


Re: about C getline function

2008-11-12 Thread Joerg Sonnenberger
On Wed, Nov 12, 2008 at 12:54:28PM +0100, dark0s Optik wrote:
 Where is defined getline C function in DragonFlyBSD ?

There is no getline function in C. There is fgets and the more useful,
but less portable fgetln.

Joerg


Re: Sony PCG-R505EL

2008-10-07 Thread Joerg Sonnenberger
On Tue, Oct 07, 2008 at 12:56:39AM -0400, Justin C. Sherrill wrote:
 Is there a firewire block device that could correspond to the CD drive?

They are supposed to use SBP (SCSI over IEEE1394) and just appear as
normal SCSI CD-ROM. Are you sure the firewire controller itself is
supported?

Joerg


Re: mutt's buffy-list

2008-08-28 Thread Joerg Sonnenberger
On Wed, Aug 27, 2008 at 12:49:13PM -0700, Matthew Dillon wrote:
 I could fake it I guess, but programs really have no business making
 assumptions about directories based on their st_nlinks or st_size
 fields.

At least nlinks == 2 is a hard-coded assumption for leaf directories in
various programs like find(1). You don't have to change the link count
for every file create/destroy, only directories.

Joerg


Re: mutt's buffy-list

2008-08-27 Thread Joerg Sonnenberger
On Thu, Aug 28, 2008 at 12:01:50AM +0900, YONETANI Tomokazu wrote:
 My questions are, is checking st_size of a directory is usual on other
 platforms/filesystems?  Does anyone know of other filesystems on which
 a directory can have st_size of zero?

I don't know any filesystem where the directories have a fixed size of
0. I don't think that should be done, it is helpful for other things.

Nevertheless, it should likely check the link count...

Joerg


Re: Hammer: Transactional file updates

2008-08-20 Thread Joerg Sonnenberger
On Tue, Aug 19, 2008 at 07:00:23PM -0700, Matthew Dillon wrote:
 * Unix in general does not guarantee atomicy between mmap-read or
   mmap-written blocks and read() or write() ops.  This is because
   it has no way to know what the user program actualy wants when,
   since all memory accesses are page-based.

Unix in practise *does* guarantee this. The only exception of an
Operating System still in use is OpenBSD. It is non-trivial to do, e.g.
in NetBSD it was done with the switch to the Unified Buffer Cache,
FreeBSD did as you very likely know did it by changes to the Buffer
Cache etc. Otherwise I agree.

Joerg


Re: Hammer: Transactional file updates

2008-08-19 Thread Joerg Sonnenberger
On Sun, Aug 17, 2008 at 08:40:20PM +1000, Dmitri Nikulin wrote:
 I personally believe that Unix should have had a transactional file IO
 API from the start, so that all modern file systems would implement it
 and atomicity would be the standard, not a rare feature.

I am not exactly sure what you mean with atomicity, but can you
demonstrate even *one* filesystem where writes of two processes are
atomic relative to each other? I don't know any.

There are also very good reasons why Unix filesystem IO never was
transactional. It is way too expensive and complex to allow that.

Joerg


Re: [ANNOUNCEMENT] The pkgsrc-2008Q2 Branch

2008-07-28 Thread Joerg Sonnenberger
On Sun, Jul 27, 2008 at 07:21:21PM -0400, Constantine A. Murenin wrote:
 Or like `pkg_add -u` on OpenBSD? :-)

When did you last read the pkg_add man page coming with pkgsrc? :)

Joerg


Re: Hammer on snapshot cd's

2008-07-16 Thread Joerg Sonnenberger
On Wed, Jul 16, 2008 at 05:12:02PM +0800, Sepherosa Ziehau wrote:
  There seems to be a traditional, irrational fear of -O2 in the FreeBSD
  community, which I can't explain.  I've heard something about -O2 and inline
  assembly, but that's probably old as well.

Like the horrible mess they do for the ip checksumming...
Frankly, I believe most of that is due to bad inline assembly specs...

 I think it something about accessing (or dereferencing) unaligned
 fields in a union.

-O2 turns on -fstrict-aliasing by default, which means that two types
can only overlap, if they one is char or share definition. That's a bit
simplified. It is something you really want to explicitly turn off for a
kernel...

Joerg


Re: Hammer on snapshot cd's

2008-07-15 Thread Joerg Sonnenberger
On Tue, Jul 15, 2008 at 03:22:34PM -0500, Vincent Stemen wrote:
 Compiling with gcc -o query -g -Wall main.c produced no warnings at
 all.  But when I added #include query.c to the main code module after
 it was working stand alone,  I got a number of compiler warnings about
 possible use of initialized variables in functions that are in query.c.

This is depending a lot on the call graph analyser and without
optimising it simply isn't done.

Joerg


Re: Call for pkgsrc help - Firefox broken

2008-07-06 Thread Joerg Sonnenberger
On Sat, Jul 05, 2008 at 07:51:22PM -0400, Justin C. Sherrill wrote:
 I've been working on a new build, though I've updated from pbulk 0.28 to
 0.32 and I'm not making it past the Resolving... stage.  I was hoping
 getting a complete build soon means we (I) can do another rebuild and have
 it go much faster because less packages will need to be built.

If it still doesn't work, please give me the presolve.log.

Joerg


Re: Call for pkgsrc help - Firefox broken

2008-07-05 Thread Joerg Sonnenberger
On Thu, Jul 03, 2008 at 04:27:40PM +0300, Hasso Tepper wrote:
 Recent Firefox (security) update to 2.0.0.15 in pkgsrc broke it for 
 DragonFly:

I've seen this in the past and it was always a linker error.
Check if there's some new conditional somewhere missing the DragonFly
case.

Did I say I hate how upstream hasn't cared about including the patches
since Firefox 1.0.7 at least?

Joerg


Re: copying utf8 to msdos

2008-07-03 Thread Joerg Sonnenberger
On Thu, Jul 03, 2008 at 08:06:04PM +0200, Simon 'corecode' Schubert wrote:
 does anybody know how to copy files with UTF8 filenames to a msdos fs?  
 On accented letters and Umlauts cp/rsync complain Invalid argument and  
 don't create/copy the file.

Specify UTF8 as locale when mounting the filesystem? See -L.

Joerg


Re: EHCI working?

2008-07-02 Thread Joerg Sonnenberger
On Wed, Jul 02, 2008 at 04:12:44PM +0100, Steve O'Hara-Smith wrote:
   It seems to be rather better here - no error messages in the logs
 and writing to my mp3 player goes at 1.5-1.8 MB/sec while reading runs at
 4-5MB/sec.

Those sound like normal data rates for slower flash modules, no
surprise.

Joerg


Re: Library path for compiling/linking

2008-06-17 Thread Joerg Sonnenberger
On Tue, Jun 17, 2008 at 01:39:59AM -0500, Vincent wrote:
 Previously, setting the library path in the environment has allowed a lot of
 packages to compile without special configure options on systems such as
 NetBSD.  Is this deliberately disabled for GCC on Dragonfly?  Or is it
 something that is temporarily broken?

Never used that and all I tried failed very badly when I needed it.
Consider using a wrapper script for cc/c++ instead.

Joerg


Re: Path MTU discovery question

2008-05-22 Thread Joerg Sonnenberger
On Thu, May 22, 2008 at 10:24:39PM +0300, Jordan Gordeev wrote:
 Is there a reason that Path MTU discovery is not turned on by default?

At least in the past it created issues with lots of broken firewalls
that block all kinds of ICMP.

Joerg


Re: problem with ELF

2008-05-12 Thread Joerg Sonnenberger
On Mon, May 12, 2008 at 07:37:32PM +0200, dark0s Optik wrote:
 I've installed opera, but when I lanch it, console outputs

Load linux emulation.

Joerg


Re: pkgsrc -current on DragonFly 1.13.0/i386

2008-05-11 Thread Joerg Sonnenberger
On Sun, May 11, 2008 at 05:16:51PM +0200, Robert Luciani wrote:
 Also, packages with special licenses are removed.

Just use the normal upload script and it will reduce the list
automatically according to the NO_BIN_ON_FTP attribute.

Joerg


Re: 7-Zip / Bzip2

2008-05-11 Thread Joerg Sonnenberger
On Sun, May 11, 2008 at 10:46:38AM -0700, Freddie Cash wrote:
 Instead of pulling in another app, consider pulling in libarchive and
 friends from FreeBSD 6+, and then adding 7z support to that.

You know that one of the two reasons I wrote the compression_program
support in libarchive was 7z's lack of a proper library interface?

Joerg


Re: Announcing the pkgsrc-2008Q1 Branch

2008-04-14 Thread Joerg Sonnenberger
On Mon, Apr 14, 2008 at 02:51:32PM +0300, Hasso Tepper wrote:
 Note that devel/m4 is now broken in DragonFly because of security fix 
 pulled up from HEAD with two days during release. Fixing it is beyond my 
 knowledge, so please someone have a look.

The older version should have all the patches. You are welcome to complain
to the gnulib folks for not including them...

Joerg


Re: Ruby crash on DragonFly

2008-03-31 Thread Joerg Sonnenberger
On Tue, Apr 01, 2008 at 02:53:52AM +0200, Michael Neumann wrote:
 I don't know why, but my application raises an illegal instruction (signal 
 4) exception if it is linked with -pthread.

Main binary must be linked against libpthread if any DSO mgiht depend on
that.

Joerg


Re: Pkgsrc problems [ was: lang/python24 build problems]

2008-03-27 Thread Joerg Sonnenberger
On Thu, Mar 27, 2008 at 10:08:45AM +0100, Robert Luciani wrote:
 Didn't you mention that the build system could be distributed?

I don't think it helps that much.

 If CPU is a bottleneck, perhaps some of us could lend out hardware?
 Also perhaps it would ease the burden you when checking for failures
 etc.. Perhaps we should stir something up.

Let me give you some different numbers. The last NetBSD build I did on a
1.8Ghz Core 2 took 3 days and 2h. It was using a disk per CPU and
journaling UFS, but async mounts should give a very similiar effect.
It should be noted that the results for my other build machine (not
building for unrelated reasons) used a vmware setup for DragonFly and
was doing bulk builds in less than 4 days (one machine / core).
Something like that would cost  1000 EUR today.

Joerg


Re: what components of X must to install

2008-03-27 Thread Joerg Sonnenberger
On Fri, Mar 28, 2008 at 06:23:44AM +0100, dark0s Optik wrote:
 Hi all, I would like a minimal window manager over DragonFly, like
 Blackbox, but I don't understand how to install X server.
 In xorg 6, we had x server, but now, what components I must to install over 
 OS?

Pich xf86-input-mouse, xf86-input-kbd and your video driver. You
normally will need at least xauth and xinit from the application package
as well.

Joerg


Re: Building modular xorg

2008-03-25 Thread Joerg Sonnenberger
On Mon, Mar 24, 2008 at 01:16:36AM -0700, Dave Hayes wrote:
 This actually worked. However building what I'm trying to build seems to
 require all of X, since the build stopped at:
 
   configure: error: X is required, but it was either disabled or not found.

This is generally a sign that it tries to link with libXt for testing.
Whether or not it actually *needs* libXt is a separate issue.

Joerg


Re: Building modular xorg

2008-03-25 Thread Joerg Sonnenberger
On Mon, Mar 24, 2008 at 01:12:35PM -0700, Dave Hayes wrote:
 The package x11/Xrender is in meta-pkgs/modular-xorg-libs. I'll 
 poke this information over on pkgsrc-users. Thanks for the response.

No, it isn't. It never was.

Joerg


Re: Digest users.v001.n1342

2008-03-21 Thread Joerg Sonnenberger
On Fri, Mar 21, 2008 at 03:59:04AM +0100, Klaus Heinz wrote:
 Is there any reason to keep
 
   #if defined(_SC_NPROCESSORS_ONLN)  !defined(__DragonFly__)
 
 and _not_ to simplify this to
 
   #if defined(_SC_NPROCESSORS_ONLN)

You mean beside the possibility that _SC_NPROCESSORS_ONLN is a enum?
No, not really. The original patch just hard-coded 1, which is good
enough.

Joerg


Re: _SC_NPROCESSORS_ONLN on Dragonfly

2008-03-20 Thread Joerg Sonnenberger
On Thu, Mar 20, 2008 at 12:07:00AM +0100, Klaus Heinz wrote:
 I am feeding back patches to the authors of audio/gtkpod and found this
 one (in patch-bj) concerning Dragonfly:

That code should be nuked completely, it is just stupid. Noone has yet
found a reason why it should use more than one thread by default.

Joerg


Re: [GSoC] the AMD64 port

2008-03-18 Thread Joerg Sonnenberger
On Tue, Mar 18, 2008 at 10:36:32PM +0200, Jordan Gordeev wrote:
  1. Does porting require messing with the source code of gcc/binutils/gdb?

Not much, if at all.

  2. Do I need to have an AMD64 machine with more than 4 GB RAM to be able 
 to fully test my work?

No.

  3. Would I be required to get 32-bit binaries working on 64-bit kernels or 
 would being able to run 64-bit binaries suffice?

The latter.

  4. Is SMP part of the equation?

SMP is almost a non-issue. The only code that is not shared with i386 in
this area is the assembly bootstrap for the second CPU, the rest is
identical.

Joerg


Re: Installation on Macbook Pro

2008-03-10 Thread Joerg Sonnenberger
On Mon, Mar 10, 2008 at 06:31:20PM +, Christopher Rawnsley wrote:

 On 9 Mar 2008, at 19:22, Joerg Sonnenberger wrote:
 I suppose this is the same problem I have seen in NetBSD. Basically,
 some firmware images reenable interrupts when the legacy support is
 turned off. Fix can be found in NetBSD's UHCI driver.

 That would wouldn't happen to be related to another problem I am having? 
 When I boot the live disc I get a prompt asking which kernel options I 
 want. Now the keyboard has always worked in this situation. When the live 
 disc has booted, however, the keyboard occasionally works. I have to reboot 
 again and cross my fingers...

That could be an unacknowledged interrupt. Not sure.

Joerg


Re: howto use /usr/include/iconv.h

2008-01-25 Thread Joerg Sonnenberger
On Fri, Jan 25, 2008 at 05:58:34PM +, Johannes Hofmann wrote:
 what is the correct way to use iconv.h from /usr/include instead of 
 /usr/pkg/include ?

Symlink it in yet another directory that is included before
/usr/pkg/include in the search path.

Joerg


Re: cvsup

2008-01-16 Thread Joerg Sonnenberger
On Wed, Jan 16, 2008 at 02:35:48AM +, Vincent Stemen wrote:
 I am not clear how to access it.

Try something like the attached script.

Joerg


update.sh
Description: Bourne shell script


Re: cvsup

2008-01-15 Thread Joerg Sonnenberger
On Tue, Jan 15, 2008 at 10:23:27PM +, Vincent Stemen wrote:
 Is the repository available via rsync and, if not, I am curious as to
 why not?  Rsync is robust, popular, portable, and written in C.

Yes. Check the mirror list, e.g. allbsd has it.

Joerg


Re: Fetch pkgsrc index with pkg_search

2007-12-13 Thread Joerg Sonnenberger
On Tue, Dec 11, 2007 at 05:19:27PM +0100, Matthias Schmidt wrote:
 Thanks, I wasn't aware of this file.  The file is not very parser
 friendly, compared with the INDEX file :)

Huh? That file is easy to process with awk(1), that makes it by
definition parser friendly :-)

Joerg


Re: Samba LDAP

2007-12-13 Thread Joerg Sonnenberger
On Wed, Dec 12, 2007 at 01:32:07PM -0600, Thomas Donnelly wrote:
 I have been watching the DragonFlyBSD project for a few months now and so 
 far it seems very promising. We are looking at doing a 2 server setup of 
 samba and ldap as a cluster/redundant pair. After looking around, it seemed 
 that DragonFlyBSD would be a good choice.

After using such a setup for over 2 years -- avoid it if you can. The
LDAP part I mean.

Joerg


Re: where is xfs?

2007-11-26 Thread Joerg Sonnenberger
On Mon, Nov 26, 2007 at 01:26:08AM -0500, Stephane Russell wrote:
 Does anyone knows where is the xfs program on the latest pkgsrc package?
 It's normally included in the xorg package, but I can't find it with the
 latest version (pkgsrc 2007Q3).

The X Font Server hasn't been packaged, because frankly I haven't seen
the need for it. Do you have a setup where it is helpful?

Joerg


Re: va_copy() problem

2007-11-13 Thread Joerg Sonnenberger
On Mon, Nov 12, 2007 at 08:53:46PM +, Johannes Hofmann wrote:
 Joerg Sonnenberger [EMAIL PROTECTED] wrote:
  On Mon, Nov 12, 2007 at 06:58:30PM +, Johannes Hofmann wrote:
va_copy(va1, va);
n = vsnprintf(s-str + s-len, s-size - s-len, format, va);
va_end(va1);
  
  va - va1 in the call to vsnprintf. Also consider using vasprintf.
 
 Ooops, you are right of course.
 But unfortunately it still crashes badly...

Works fine after the change on both NetBSD current and DragonFly 1.8.

Joerg


Re: va_copy() problem

2007-11-12 Thread Joerg Sonnenberger
On Mon, Nov 12, 2007 at 06:58:30PM +, Johannes Hofmann wrote:
   va_copy(va1, va);
   n = vsnprintf(s-str + s-len, s-size - s-len, format, va);
   va_end(va1);

va - va1 in the call to vsnprintf. Also consider using vasprintf.

Joerg


Re: gcc update (was: Re: va_copy() problem)

2007-11-12 Thread Joerg Sonnenberger
On Mon, Nov 12, 2007 at 09:27:17PM +0100, Simon 'corecode' Schubert wrote:
 No, the problem is that gcc uses %ebx after a function call, which it is
 not allowed to do:

It is allowed to. From the i386 ABI:

If necessary, a function saves the values of %edi, %esi, and %ebx in the
positions shown and restores their values before returning to the caller.

Joerg


Re: moused + modular xorg 'sticking' ?

2007-11-08 Thread Joerg Sonnenberger
On Thu, Nov 08, 2007 at 02:07:55PM +0800, Sepherosa Ziehau wrote:
 My problem with /dev/sysmouse is that my mouse pointer will go to
 upper left corner and can't move anymore.

Sounds like the FPU-used-in-signal handler issue.

Joerg


Re: moused + modular xorg 'sticking' ?

2007-11-08 Thread Joerg Sonnenberger
On Thu, Nov 08, 2007 at 11:43:46AM +0100, Simon 'corecode' Schubert wrote:
 First, I am surprised that we don't know about this bug.  Second, maybe
 somebody could elaborate on the manifestation of this bug and if the use
 of floating point arithmetic is allowed in signal handlers.

At least about the second: there's nothing in ISO C suggesting anything
else.

Joerg


Re: moused + modular xorg 'sticking' ?

2007-11-08 Thread Joerg Sonnenberger
On Thu, Nov 08, 2007 at 02:17:21PM +0100, Simon 'corecode' Schubert wrote:
 Joerg Sonnenberger wrote:
  On Thu, Nov 08, 2007 at 02:07:55PM +0800, Sepherosa Ziehau wrote:
  My problem with /dev/sysmouse is that my mouse pointer will go to
  upper left corner and can't move anymore.
  
  Sounds like the FPU-used-in-signal handler issue.
 
 do you know what is going on there or do you only know that the fpu has
 to do something with this?

I know that various signal handlers used floating point and the issues
reported fall into the category of likely overwritten FPU state.

Joerg


Re: moused + modular xorg 'sticking' ?

2007-11-08 Thread Joerg Sonnenberger
On Thu, Nov 08, 2007 at 12:05:25PM -0800, Matthew Dillon wrote:
 :I know that various signal handlers used floating point and the issues
 :reported fall into the category of likely overwritten FPU state.
 :
 :Joerg
 
 Should we start saving and restoring the FP context?  The ucontext
 structure does have enough space reserved for it.  During the LWP
 work we expanded the FP save space to 512 bytes.

Yes, please. This is a bug after all :-)

Joerg


Re: Install JRE

2007-10-10 Thread Joerg Sonnenberger
On Wed, Oct 10, 2007 at 11:25:33PM +1000, Dmitri Nikulin wrote:
 On 10/7/07, Justin Chan [EMAIL PROTECTED] wrote:
  Hi, i am using dfBSD 1.10.0.
 
  I have tried to use the usr/pkgsrc/lang/sun-jre15 ,jre14,jdk14,jdk15...all
  couldn't work.
 
  Will appreciate, if i can be pointed to a previous
  mail ,if existed, which shows the actual steps in
  installing JRE.
 
  Thanks alot.
 
  Regards,
  Justin Chan
 
 I think it's about time somebody checked out OpenJDK to port it to
 DragonFly. Being GPL, it'd be a(n almost) complete Java solution.

Thanks for volunteering to do it.

Joerg


Re: /etc/rc.d/localdaemons

2007-09-11 Thread Joerg Sonnenberger
On Mon, Sep 10, 2007 at 09:31:18PM -0700, Dave Hayes wrote:
 Joerg Sonnenberger [EMAIL PROTECTED] writes:
  No, it is a complete mess. If you really want to keep them separate, use
  a second directory like /etc/rc.d.pkg and change /etc/rc to scan that as
  well.
 
 Is the code itself a mess or is the idea bad? If the idea is bad, I'd be
 interested in why, so I can understand this comment better. :)

The idea is bad. It makes assumption about what the orders the second
set of scripts is allowd to take. It makes *no* sense at all if the
second set of scripts is already on the root filesystem. etc pp

Joerg


Re: /etc/rc.d/localdaemons

2007-09-10 Thread Joerg Sonnenberger
On Mon, Sep 10, 2007 at 10:52:02AM -0700, Dave Hayes wrote:
 I believe pkgsrc puts rc.d style scripts in /usr/pkg/etc/rc.d. I've also
 put a few in there myself. 

It hasn't done that for ages.

 There's two reasons I personally like this idea, one is to keep the
 startup script with the prefix where it's software installed, and two is
 that I tend to keep as much local software modification out of /etc as
 possible.

Copy / symlink them by hand and use a prefix.

  By the way, FreeBSD's startup scripts check the rc.d scripts for the a
  rcorder tag to then use rcorder to order them.
 
 This sounds like a good thing to bring to DragonFly. :)

No, it is a complete mess. If you really want to keep them separate, use
a second directory like /etc/rc.d.pkg and change /etc/rc to scan that as
well.

Joerg


Re: Naive question on brightness keys

2007-08-29 Thread Joerg Sonnenberger
On Wed, Aug 29, 2007 at 01:44:03PM -0400, Radu-Cristian FOTESCU wrote:
 --- Hans-Werner Hilse [EMAIL PROTECTED] wrote:
  
  Because those aren't handled by any software. Nothing is eaten. 
  But nothing is provided either.
 
 I am sorry to be that stubborn, but nothing is provided either when at the
 boot loader's prompt, yet the key are working then!

At the boot prompt, the BIOS is still using the legacy configuration.

 But (wrt to the graphical table), note that *all* the Linux distros I tested,
 and also PC-BSD (I am not sure on NetBSD) were using ACPI too, and they were
 also (Linux) able to suspend to disk! In short, not using APM, yet working
 stuff.

It might have a specific ACPI EC controller. E.g. on Thinkpad's the
volume keys are handled via the EC as well. The suspend keys are now
translated into proper ACPI events, some others are handled in hardware.
Which are handled and which not is BIOS specific now.

Joerg


Re: modular xorg build fails

2007-08-26 Thread Joerg Sonnenberger
On Tue, Aug 21, 2007 at 09:53:55AM -0500, [EMAIL PROTECTED] wrote:
 I'm having trouble building modular xorg on a new install of DragonFly 1.8.1;
 modular-xorg-apps and -fonts build okay but modular-xorg-drivers fails
 when it tries to build modular-xorg-server.

Gt modular-xorg-server 1.3.0nb2 and try with that. A merge fault removed
from essential patches and this is one of the side effects.

Joerg


Re: Metapackage kde-3.5.7 absent from 1.10

2007-08-23 Thread Joerg Sonnenberger
On Thu, Aug 23, 2007 at 11:33:01PM +0200, Simon 'corecode' Schubert wrote:
 Radu-Cristian FOTESCU wrote:
 kde-3.5.7.tgz is present in:
 ftp://chlamydia.fs.ei.tum.de/pub/DragonFly/packages/stable/DragonFly-1.8/kde/
 but it is absent from:
 ftp://chlamydia.fs.ei.tum.de/pub/DragonFly/packages/stable/DragonFly-1.10/kde/
 Should this be considered as a bug? :-)

 A pkgsrc problem.  kdenetwork didn't build.

s/pkgsrc/KDE/

Joerg


Re: libgtop2 and xorg break

2007-08-14 Thread Joerg Sonnenberger
On Mon, Aug 13, 2007 at 05:56:33PM -0400, Justin C. Sherrill wrote:
 Technically, we can remove PKG_DBDIR, LOCALBASE VARBASE, and X11_TYPE, as
 the values we have in /etc/mk.conf are the defaults on DragonFly anyway. 
 Joerg, can you confirm this?

For X11_TYPE yes, for the others: objected. It makes it harder to figure
out the values of the configuration files, esp. for someone not running
DragonFly as well.

Joerg


Re: libgtop2 and xorg break

2007-08-13 Thread Joerg Sonnenberger
On Mon, Aug 13, 2007 at 08:28:32AM +0200, Pieter Dumon wrote:
 however, /etc/mk.conf contains
 X11_TYPE=xorg
 (within the .ifdef BSD_PKG_MK  .endif)

Just remove that one, there should be /usr/pkg/etc/mk.conf which should
be used instead.

Joerg


Re: dfly mail archive not working

2007-08-11 Thread Joerg Sonnenberger
On Sat, Aug 11, 2007 at 03:16:44PM -0400, Chris Turner wrote:
  ${ctl_command} ${apache_flags} -k ${action}

Why the -k?

Joerg


Re: DragonFly installation experience / suggestions

2007-08-11 Thread Joerg Sonnenberger
On Sat, Aug 11, 2007 at 09:33:36PM +0200, Michael Neumann wrote:
 Sorry, you're right! It was a different version.

 === Installing for libxslt-1.1.21
 ERROR: libxslt-1.1.20 is already installed - perhaps an older version?
 ERROR: If so, you may use either of:
 ERROR: - pkg_delete libxslt-1.1.20 and /usr/pkg/bin/bmake

...

The only package I can find that explicitly requests 1.1.21 is
py-libxslt. And that makes sense as they are quite right together.

Joerg


Re: Looking for ap22-subversion

2007-08-08 Thread Joerg Sonnenberger
On Wed, Aug 08, 2007 at 02:04:57PM -0300, Sd?vtaker wrote:
 Did someone port ap22-subversion to DFBSD?

You have to build subversion-base with option apache22.
This is related to the incompatibility of apr versions between
Apache 2.0 and Apache 2.2.

Joerg


Re: Trying to setup PF and spamd...hell!

2007-08-08 Thread Joerg Sonnenberger
On Thu, Aug 09, 2007 at 01:22:27AM +1000, [EMAIL PROTECTED] wrote:
 I decided to try spamd/greylisting so I configured pf.conf the following way:

Don't rdr into the 127/8, use a different subnet.

Joerg


Re: Problems with installing xorg under dragonfly.

2007-08-05 Thread Joerg Sonnenberger
On Sun, Aug 05, 2007 at 06:57:47PM +0200, Gergo Szakal wrote:
 Clean your source tree with pkgtools/pkgclean first, the mk.conf change
 still applies.

No, it doesn't.

Joerg


Re: Release 1.10 update - 31-Jul-07

2007-08-02 Thread Joerg Sonnenberger
On Thu, Aug 02, 2007 at 12:36:32AM -0400, Justin C. Sherrill wrote:
 On Wed, August 1, 2007 11:34 pm, Matt Emmerton wrote:
 
 
  For pkgsrc, why couldn't we just include a recent snapshot of /usr/pkgsrc
  on the CD (much like how FreeBSD provides a snapshot of /usr/ports).  The
  tree that I updated last night is 387MB uncompressed, 195MB compressed.
 
 I agree with this - not including pkgsrc shrinks the size of the download,
 but it's still 80-something meg, and you end up using the bandwidth anyway
 pulling pkgsrc down through anoncvs, which also requires a bunch of extra
 steps which are only documented in a few places.

ftp.netbsd.org has tarballs of anoncvs checkouts. That makes the initial
setup much faster.

Joerg


Re: FF varsym

2007-07-29 Thread Joerg Sonnenberger
On Sat, Jul 28, 2007 at 12:55:21PM -0700, Matthew Dillon wrote:
 Why firefox is doing that I have no idea.

Let me make a bet: realpath(3).

Joerg


Re: default Sendmail plus Cyrus-SASL

2007-07-27 Thread Joerg Sonnenberger
On Fri, Jul 27, 2007 at 07:00:21PM -0700, Joseph Garcia wrote:
 SENDMAIL_LDFLAGS=-L/usr/pkg/lib

Add -R/usr/pkg/lib here as well.

Joerg


Re: CMake support for DragonFly BSD

2007-07-13 Thread Joerg Sonnenberger
On Fri, Jul 13, 2007 at 09:58:05AM -0400, Alexander Neundorf wrote:
 the build system CMake (http://www.cmake.org), which is used by several free 
 software projects, e.g. KDE4, doesn't officially support DragonFly BSD, but 
 we'd like to change this. AFAIK DragonFly is quite similar to FreeBSD. 

Why do you want to change anything? It worked fine so far. A number of
packages in pkgsrc use it and work fine AFAICT.

 What does uname report ? DragonFly or dragonfly or does it depend on the 
 version of DragonFly ?

From dfly18.pkgsrc-box.org:

uname:
DragonFly
uname -v:
DragonFly 1.8.0-RELEASE #0: Mon Jan 29 18:17:49 PST 2007
root@:/usr/obj/usr/src/sys/GENERIC 
name -m:
i386
uname -r:
1.8.0-RELEASE

Joerg


Re: CMake support for DragonFly BSD

2007-07-13 Thread Joerg Sonnenberger
On Fri, Jul 13, 2007 at 10:21:57AM -0400, Alexander Neundorf wrote:
 On Friday 13 July 2007 10:11, Joerg Sonnenberger wrote:
  On Fri, Jul 13, 2007 at 09:58:05AM -0400, Alexander Neundorf wrote:
   the build system CMake (http://www.cmake.org), which is used by several
   free software projects, e.g. KDE4, doesn't officially support DragonFly
   BSD, but we'd like to change this. AFAIK DragonFly is quite similar to
   FreeBSD.
 
  Why do you want to change anything? It worked fine so far. A number of
  packages in pkgsrc use it and work fine AFAICT.
 
 Do you have a DragonFly.cmake in Modules/Platform/ ?
 CMake doesn't have one. If you have one, could you please post it or send a 
 link to the cmake sources used by DragonFly ?
 If there is no DragonFly.cmake, cmake should complain that your system is not 
 supported.

We are just using a copy of the FreeBSD file. I take it your a cmake
developer? Can you make it easier to override a number of hard-coded
default locations? See the pkgsrc patch for what I mean.

Joerg


Re: CMake support for DragonFly BSD

2007-07-13 Thread Joerg Sonnenberger
On Fri, Jul 13, 2007 at 01:34:10PM -0400, Alexander Neundorf wrote:
 On Friday 13 July 2007 13:22, Joerg Sonnenberger wrote:
  On Fri, Jul 13, 2007 at 12:53:49PM -0400, Alexander Neundorf wrote:
The prefix is fully user-settable and the patch ensures
that cmake works with dependencies installed via pkgsrc.
  
   Can you please explain, I know basically nothing about pkgsrc.
 
  What I mean is: when I build cmake to install to /some/arbitrary/path
  via pkgsrc, I expect it to look for normal libraries in the same
 
 So would it help if cmake also checks the bin/, lib/ and include/ directories 
 of its own install location ?
 This would be easy to add.

Yes, that is basically what the patch tries when adding @PREFIX@/...
This is IMO preferable to hard-coding any non-standard locations (/usr),
/usr/local makes a lot less sense with this. If I don't install into
/usr/local, I normally wouldn't want software to scan it either.

  location. I would also prefer to keep the list of hard-coded directories
  down, but that conflicts with the idea of work-out-of-the-box on
  insane platforms. For X11, pkgsrc knows where to look or which one to
  pick up. That can make a difference when you have both native headers
  and modular Xorg installed (think SunOS and other older Unix versions).
 
 There are still the environment variables which can be set up. What do you 
 think about this ?

The environmental variables are not set by the normal user, I think.
When cmake is built by pkgsrc, the system knows what location it should
look for.

A good compromise would a package-system switch be that removes most
additional path names and add a few options to specify where to look for
header files and libraries. The X11 stuff as second option is logical as
well.

Joerg


Re: [ANNOUNCEMENT] The pkgsrc-2007Q2 Branch

2007-07-03 Thread Joerg Sonnenberger
On Tue, Jul 03, 2007 at 10:51:36AM -0700, Matthew Dillon wrote:
 :Binary packages are available from
 :http://www.pkgsrc-box.org/packages/2007Q2/DragonFly-1.8/All
 :and the stable link been adjusted accordingly.
 :
 :Joerg
 
  Good timing!

It would be nice if someone could look at the breakage of jdk-1.4+. I
think it is either linprocfs or someone mincore(2) or so, but I'm not
sure. The latter should be fixed anyway.

Joerg


Re: 1.10 release will be mid-July tentatively.

2007-07-03 Thread Joerg Sonnenberger
On Tue, Jul 03, 2007 at 01:18:00PM -0700, Matthew Dillon wrote:
 
 :I just committed the last piece of usb_port.h macros removal cleanup to
 :the USB code. Unfortunately I will be away from home next weeks
 :travelling around the Europe and enjoying my holidays. So, if it turns
 :out that I broke something in progress, I can't take care of fixing it
 :myself, sorry ;).
 :
 :About release number I have one concern though. A lot of code I have
 :written during last year checks for __DragonFly_version = 19. It
 :would be nice to see it working after 1.10 release as well ;).
 :
 :-- 
 :Hasso Tepper
 
 Good point.  I'll call the release 1.10 but I'll use 195000 for 
 DragonFly_version.

Why not simply continue counting and use 1A...

Joerg


Re: [ANNOUNCEMENT] The pkgsrc-2007Q2 Branch

2007-07-02 Thread Joerg Sonnenberger
On Sun, Jul 01, 2007 at 11:25:50AM +0100, Alistair Crooks wrote:
 The pkgsrc developers are very proud to announce the new pkgsrc-2007Q2
 branch, which has support for more packages than previous branches. 
 As well as updated versions of many packages, the infrastructure of
 pkgsrc itself has been improved for better platform and compiler
 support.

...

Binary packages are available from
http://www.pkgsrc-box.org/packages/2007Q2/DragonFly-1.8/All
and the stable link been adjusted accordingly.

Joerg



Re: [ANNOUNCEMENT] The pkgsrc-2007Q2 Branch

2007-07-02 Thread Joerg Sonnenberger
On Mon, Jul 02, 2007 at 08:15:06AM +0200, Joerg Sonnenberger wrote:
 Binary packages are available from
 http://www.pkgsrc-box.org/packages/2007Q2/DragonFly-1.8/All
 and the stable link been adjusted accordingly.

On the same note: I'll be running regression bulk builds with
CHECK_PERMISSIONS=yes for a while. This is expected to break packages,
so you might consider using the stable packages for a while :-)

Joerg


Re: Apache in pkgsrc has largefile support on?

2007-06-26 Thread Joerg Sonnenberger
On Tue, Jun 26, 2007 at 09:12:12AM -0700, Dave Hayes wrote:
 From /usr/pkgsrc/www/apache/Makefile:
 
   # Explicitly turn on large file support
   APACHE_CUSTOM_CFLAGS+= -D_LARGEFILE_SOURCE
   APACHE_CUSTOM_CFLAGS+= -D_FILE_OFFSET_BITS=64
 
 DFly does not have largefile support yet, correct?  So this should be
 turned off for DFly, yes? 

All BSDs have *only* largefile support except for kernel ABI compat.

Joerg


Re: Using PKGSRC to install to non-standard LOCALBASE

2007-06-25 Thread Joerg Sonnenberger
On Sun, Jun 24, 2007 at 09:07:07AM -0700, Dave Hayes wrote:
 If I wanted to build Apache, for example, out of pkgsrc but I wanted
 it to be installed to a different prefix than LOCALBASE, how would I
 do that?

The recommented way is to bootstrap to the new prefix and use the bmake
from that. You can with some care alter the LOCALBASE variable together
with PKG_DBDIR to get a similiar effect, but IMO it is more likely to
hit problems like which combination am I currently running with.

Joerg


Re: structure has no member named `kp_eproc'

2007-05-15 Thread Joerg Sonnenberger
On Tue, May 15, 2007 at 01:48:18AM +1000, Petr Janda wrote:
 I'm trying to compile net-snmp from SVN applying the pkgsrc patches and 
 I can't seem to figure out why its failing on this error. What's the  
 meaning of this error and how to fix it?

Patches in pkgsrc only apply to DragonFly 1.6 and 1.8. Well, they work
on 1.8 and did work on 1.6 at one point. Please don't break that
upstream...

Joerg


Re: rebuild DragonFly using gcc 4.1

2007-05-07 Thread Joerg Sonnenberger
On Sat, May 05, 2007 at 10:14:47PM +0200, Erik Wikström wrote:
 For third party application don't follow what's written under Installing 
 software, look at http://wiki.dragonflybsd.org/index.cgi/HowToPkgsrc and 
 follow the steps under the section Pre-built pkgsrc packages, but use 
 the addresses found on
 
 ftp://packages.stura.uni-rostock.de/pkgsrc-current/DragonFly/RELEASE/i386/All
 
 if a package you want can't be found there take a look in
 
 ftp://packages.stura.uni-rostock.de/pkgsrc-current/DragonFly/RELEASE/i386/vulnerable

Note that those packages are no longer updated. The current sets can be
found under
http://www.pkgsrc-box.org/packages/current/DragonFly-1.8/All/
for pkgsrc HEAD and
http://www.pkgsrc-box.org/packages/stable/DragonFly-1.8/All/
for the release branch.

Before starting to mirror those, please contact me off-list.

Joerg


Re: 2 NIC's on the same subnet.

2007-05-04 Thread Joerg Sonnenberger
On Fri, May 04, 2007 at 09:49:42PM +1000, Petr Janda wrote:
 Is it wrong? What are the possible implications?

It can be done with some care. The problem is that the current ARP
behaviour has some issues, e.g. it complains about receiving ARP answers
on the wrong interface etc.

Joerg


Re: Gnome: Couldn't load icon... bug

2007-05-01 Thread Joerg Sonnenberger
On Tue, May 01, 2007 at 08:49:23AM -0700, walt wrote:
 libgtop hasn't compiled on -DEVEL for quite awhile.  I think Joerg is
 waiting for the kernel data structures to stabilize.  That's why I'm
 not nagging him the way I used to ;o)

It doesn't work on anything post 1.7.X. I've run out of free beer ^W^W
desire to fix it...

Joerg


Re: va_copy() needs sys/cdefs.h?

2007-04-22 Thread Joerg Sonnenberger
On Sun, Apr 22, 2007 at 02:56:33PM +0200, Klaus Heinz wrote:
 Is this the correct behaviour and should such a patch be fed back to the
 software's authors?

No, it isn't. It is a bug that was fixed a while after 1.4 and pulled up
later. Are you running the latest version of 1.4?

Joerg


Re: [HEADS UP] Binary package changes

2007-04-12 Thread Joerg Sonnenberger
On Thu, Apr 12, 2007 at 02:36:39PM +0200, Grzegorz B?ach wrote:
 When I try install pkg from vulnerable dir and its depends are it All
 dir, dependences can't be installed, but when I try install pkg where
 pkg and every depends are in All dir everything is OK.

Yeah, that's a known bug. I want to fix it, but it is a lot harder.

Joerg


Re: [HEADS UP] Binary package changes

2007-04-10 Thread Joerg Sonnenberger
On Thu, Apr 05, 2007 at 05:32:59PM +0200, Grzegorz B?ach wrote:
 I can't install any packages from pkgsrc-box.org, but from
 pkgsrc.dragonflybsd.org it works OK.

Can you rebuild pkg_install with the attached patch and see if that
works?

Joerg
Index: ftpio.c
===
RCS file: /home/joerg/repo/netbsd/src/usr.sbin/pkg_install/lib/ftpio.c,v
retrieving revision 1.73
diff -u -r1.73 ftpio.c
--- ftpio.c 11 May 2006 23:50:15 -  1.73
+++ ftpio.c 10 Apr 2007 13:32:34 -
@@ -655,7 +655,10 @@
char *s, buf[MaxPathSize];
char tmpname[MaxPathSize];
char best[MaxPathSize];
-   int rc, tfd;
+   int rc, tfd, retry_tbz;
+
+   retry_tbz = 0;
+   best[0]='\0';
 
rc = ftp_start(base);
if (rc == -1) {
@@ -684,8 +687,16 @@
 * we can't use the pkg wildcards here as dewey compare
 * and alternates won't be handled by ftp(1); sort
 * out later, using pmatch() */
-   (void) snprintf(buf,  sizeof(buf), nlist %.*s*.t[bg]z %s\n,
-   (int)(s-pattern), pattern, tmpname);
+   if (retry_tbz) {
+retry_with_tbz:
+   (void) snprintf(buf,  sizeof(buf), nlist %.*s*.tbz 
%s\n,
+   (int)(s-pattern), pattern, tmpname);
+   retry_tbz = 0;
+   } else {
+   (void) snprintf(buf,  sizeof(buf), nlist %.*s*.tgz 
%s\n,
+   (int)(s-pattern), pattern, tmpname);
+   retry_tbz = 1;
+   }
}
 
rc = ftp_cmd(buf, \n(550|450|226).*\n); /* catch errors */
@@ -704,7 +715,6 @@
return NULL;
}

-   best[0]='\0';
if (access(tmpname, R_OK)==0) {
int matches;
FILE *f;
@@ -747,6 +757,9 @@
warnx(nothing appropriate found);
}
 
+   if (retry_tbz)
+   goto retry_with_tbz;
+
unlink(tmpname);
 
if (best[0] == '\0')


[HEADS UP] Binary package changes

2007-04-05 Thread Joerg Sonnenberger
Hi all,
due to the dead of my old server I've moved the pkgsrc build system to
colocation. A few other things apply as well, so the following changes
should be of general interest:
(1) Binary packages for pkgsrc-current are available from
http://www.pkgsrc-box.org/packages/current/DragonFly-1.8/
and FTP respectively. This is built with X11_TYPE=modular.
(2) Binary packages for pkgsrc-2007Q1 will follow after Easter when the
branch is cut.
(3) rsync is available as well, if you want to mirror the packages
please contact me first. Due to bandwidth contraints (which might get
sorted out soon), I'd like to keep the list short.

I'm also looking for a donation of another KVR667D2D4F5K2/4G, as the
virtual machine configuration is running at the memory limit of the box.

Joerg


Re: To be a new DFly commiter

2007-03-17 Thread Joerg Sonnenberger
On Sat, Mar 17, 2007 at 01:26:21PM +0100, Grzegorz B?ach wrote:
 Brute-force algoritm with collision can take password 100 time faster
 than brute-force without brute-force.

Again, password hashes are *not* simple MD5 hashes. They are not even
simple salted MD5 hashes. That doesn't mean that a collision is not
possible against them, but at least not the stnadard versions everyone
is talking about.

Joerg


Re: To be a new DFly commiter

2007-03-17 Thread Joerg Sonnenberger
On Sat, Mar 17, 2007 at 03:30:11PM +0100, Michel Talon wrote:
 Another excellent statement! Maintaining a decent ports system is a task for
 hundred people. FreeBSD has aroud 200 people doing that, Debian, around
 1000.

To be fair, Debian *needs* the thousand people because the approach to
packagement they use doesn't scale.

 One has to be totally unaware of realities to suggest tools from
 obscure Linux distributions, wether they are good or bad, when such
 distribution may collapse at any moment. Already the move  to NetBSD pkgsrc
 has cost DFLY division by 3 of the number of available ports with respect
 to FreeBSD for an advantage that i have hard time to even discern.

Package counting like comparing penis length. There are more important
parameters... I've spoken with at least one member of FreeBSD's portmgr
who cursed the current size of the tree, making it very hard to
maintain or move forward. A friend also just reminded me that ports has
over 8700 (!) Perl modules in the list, factoring that out reducing the
divisor a lot.

You don't know the advance? Check out a pkgsrc tree and try building
random stuff on DragonFly and do the same with ports. Any other question
needed? As the person responsible for 2/3 of that I decided to use the
way of least resistence and the way more appealing for technical
reasons.

 The NetBSD people have replaced the horrible mess which is the 4000 lines 
 bsd.port.mk by a similar horrible mess except it is scattered over many
 5 lines files.

You have to start somewhere. Moving logic out of the make scripts is an
on-going task, which is work if you don't want to break something. Or
want to keep the breakage down to strange packages doing bad things.

 Like in many cases it is OpenBSD which is doing the good
 work, and in particular they have understood the obvious, that is a ports
 system must be centered about binary packages, not recompiling source. 

I simply refuse to comment that, but it is somewhat ironic...

Joerg


Re: To be a new DFly commiter

2007-03-17 Thread Joerg Sonnenberger
On Fri, Mar 16, 2007 at 06:58:58PM -0700, Matthew Dillon wrote:
 What does this mean?  I would dearly like to integrate portions of
 pkgsrc managed packages into our buildworld and installworld 
 system, that is have the buildworld create a little package building
 jail and build and install selected packages, with appropriate defaults,
 as part of the base system build.  Then we would not have to import or
 maintain the sources for at least the larger integrated pieces (such
 as sendmail/postfix, bind, etc).

I have some not-yet-commited patches to make unprivileged building a lot
more useful. I also plan to work on cross-compilation support for parts
of the pkgsrc tree. Both will help a lot for tighter integration of
binary packages into distributions.

Joerg 


Re: To be a new DFly commiter

2007-03-16 Thread Joerg Sonnenberger
On Fri, Mar 16, 2007 at 06:07:07PM +0100, Grzegorz B?ach wrote:
 When you do buffer-overflow in passwd you can exec any code with root 
 priviledges,
 but with tcb you must change root password to run code with root priviledges,
 and administrator will see this faster.

Who said that I want to change the root password? I can easily just
create a new user with uid 0, login remotely as that and change the
entry back. Very little log pollution and that can be easily taken care
of.

Joerg


Re: Re: To be a new DFly commiter

2007-03-16 Thread Joerg Sonnenberger
Can you please fix your MUA to follow mailing list etiquettes with
regard to line length? Thanks.

On Fri, Mar 16, 2007 at 08:47:47PM +0100, [EMAIL PROTECTED] wrote:
 I don't like pkgsrc because this limitations:
 1. Many packages in pkgsrc are obsolete, and there are no development
 version of almost all packages (i wan't to see new version of dbmail,
 xorg and enlightenment 0.17 in packages system).

Not running behind every development version is a feature. Even the
stable releases of many programs are bad enough...

I don't know about enlightenment. Ask the maintainer for dbmail, he
might or might not have a reason. Keep in mind that many of us use this
software on our server, so we are somewhat reluctant to randomly update
working code.

Modular Xorg is work in progress. Don't know which newer versions you
want to see as the supported parts are almost all up-to-date.

 2. Many packages in pkgsrc are configured with '--without-threading'
 option, because not all systems which pkgsrc work on support threading.
 I want omit this limitation.

I find only xerces-c, where it is disabled by default. The rest use
whatever the packages does by default. I don't know what you base your
opinion on, it certainly isn't because some of the platforms lack thread
support (in which case pth would be used automatically).

Joerg


Re: wiki log of #dragonfly irc channel

2007-03-09 Thread Joerg Sonnenberger
On Fri, Mar 09, 2007 at 01:28:10AM +, Helge Rohde wrote:
 Yeah, i would have thought so too. But apparently they do bend their rules 
 when the see the need, atleast in Germany they *can* put you into jail until 
 you tell them the passphrase and i have heard similar from other european 
 countries.

That's only partly true. First of all, without the consent of a judge,
you can be put into jail only until the next day. Second, to keep you in
prison they have to have enough evidence to warrant it. Third, you can
not be forced to answer the questions.

Joerg


Re: wiki log of #dragonfly irc channel

2007-03-07 Thread Joerg Sonnenberger
On Tue, Mar 06, 2007 at 06:45:27AM +1100, Dmitri Nikulin wrote:
 And if you've upset the government enough that a warrant is issued,
 any amount of their impressive technology will completely side-step
 the cryptography anyway - e.g. tempest emissions or a surprise seizure
 of belongings.

Sorry, but this is complete bull shit. The average policy agency
*anywhere* does have no fucking chance to deal with cryptography. Even
the secret services have no chance dealing with it from the stored data
alone. It is somewhat different when you can actively monitor the
encryption process, but in that case you have no reason to deal with the
cryptography itself anyway because you can just watch the plain text.

Joerg


Re: wiki log of #dragonfly irc channel

2007-03-05 Thread Joerg Sonnenberger
On Mon, Mar 05, 2007 at 10:50:23AM +, B. Estrade wrote:
 
 To be honest, I was halfway serious.  I generally don't hide my
 identity when on irc because I don't go on to just 'hang out', but I
 do like to monitor the #dragonflybsd channel.

You know that most IRC users also log the channels they are in? Many of
those are 24/7 online, So this literally doesn't help much...

Joerg


Re: fonts problem

2007-02-26 Thread Joerg Sonnenberger
On Mon, Feb 26, 2007 at 06:58:34AM +0100, Huub wrote:
 I just installed OpenOffice.org 2.1 and RealPlayer, but they report to 
 have problems with fonts.

Add some font pathes to emul/linux/etc/fontconfig/local.conf and/or
fonts.conf (the former more correct, the latter easier)

Joerg


Re: DragonFly testing

2007-02-22 Thread Joerg Sonnenberger
On Thu, Feb 22, 2007 at 02:43:18AM +0100, Gergo Szakal wrote:
 (I admit that mplayer in pkgsrc tree is quite outdated.)

Actuallly, no. It is the latest release version. Complain to the devs
for not creating releases more often.

Joerg


  1   2   3   >