Re: Gcc-ada errata?

[Chris Turner]

there was some recent commit to CURRENT about needing to enable GCC4 via
make.conf ..

but if that still doesn't work:

Might be possible to bootstrap and/or port via FreeBSD emulation ..

I've got a DragonFly support patch for CMUCL that I keep meaning
to submit that was developed via the easily-obtained
FreeBSD executables..

not sure how that would hook in with the GCC build procedure,
but :

"This is left as an exercise for the reader" :)




Bill Hacker wrote:
> walt wrote:
>> On Tue, 20 Feb 2007, Bill Hacker wrote:
>>
>>> Date: Tue, 20 Feb 2007 00:34:37 +0800
>>> From: Bill Hacker <[EMAIL PROTECTED]>
>>> Newsgroups: dragonfly.users
>>> Subject: Gcc-ada errata?
>>>
>>> Anyone seen this (DFLY 1.8 REL, Celeron 1 GHz, 512MB SDRAM,
>>> IBM/Hitachi PATA
>>> 20GB HDD):
>>>
>>> =
>>> # cd /usr/pkgsrc/lang/gcc34-ada
>>>
>>> # bmake install -DUSE_GCC34
>>
>> Oops -- ignore my other answer.  I read the README ;o)  The gcc34
>> Makefile
>> clearly assumes that you've already build gcc34 *with* ada.  AFAICT, the
>> only way you can do that is to install gcc3 first, because that includes
>> ada by default.  You can then bootstrap to gcc34, or maybe even directly
>> to gcc34-ada without building gcc34, not sure.
>>
> 
> Let's see...
> 
> ===
> # cd gcc3-ada
> # bmake install clean
> ERROR: This package is not available for these platforms: Darwin-*-*
> Interix-*-* DragonFly-*-*.
> ERROR: This package has set PKG_FAIL_REASON:
> ERROR: gcc3-ada-3.3.6 is not available for DragonFly-1.8.0-i386
> ERROR: This package has set PKG_SKIP_REASON:
> ERROR: Sorry, need an Ada compiler
> *** Error code 1
> 
> Stop.
> bmake: stopped in /usr/pkgsrc/lang/gcc3-ada
> 
> ===
> 
> # cd ..
> # cd gcc3
> # bmake install clean
> WARNING: [depends.mk] Unknown object format for installed package
> digest-20050731
> => Required installed package digest>=20010302: digest-20050731 found
> ===> Checking for vulnerabilities in gcc3-3.3.6
> work -> /usr/obj/pkgsrc/lang/gcc3/work
> ===> Installing dependencies for gcc3-3.3.6
> => Required installed package gcc3-c-3.3.6{,nb*}: NOT found
> => Verifying reinstall for ../../lang/gcc3-c
> ERROR: This package is not available for these platforms: Darwin-*-*
> Interix-*-* DragonFly-*-*.
> ERROR: This package has set PKG_FAIL_REASON:
> ERROR: gcc3-c-3.3.6 is not available for DragonFly-1.8.0-i386
> *** Error code 1
> 
> Stop.
> bmake: stopped in /usr/pkgsrc/lang/gcc3-c
> *** Error code 1
> 
> Stop.
> bmake: stopped in /usr/pkgsrc/lang/gcc3
> 
> ===
> 
> 'GNATS' as the cartoon character said
> 
> Hmm.
> 
> (ssh's to a FreeBSD 6.X box.)
> 
> +++
> triligon# cd /usr/ports/lang/gnats
> triligon# less pkg-descr
> The GPL version of GNU Ada compiler system built on GCC 3.4.6
> 
> Binaries built with this version of compiler are covered by the
> GPL license.  Use the Ada compiler bundled with GCC 4.0 and
> subsequent for LGPL licensing.
> 
> 
> 
> Joy!  gcc 4.0 and later should be OK as-is?
> 
> ===
> 
> From the DFLY 1.8 release notes:
> 
> "Bring in GCC-4.1 (setenv CCVER GCC41 to use)."
> 
> Let's see.
> 
> =
> 
> # setenv CCVER GCC41
> # gcc -v
> gcc: in path []/usr/libexec/GCC41/gcc: No such file or directory
> 
> =
> 
> And even with case changed, the subdir:
> 
> /usr/libexec/gcc41 is indeed empty...
> 
> Perhaps the ISO image is not enough. But I do not see gcc41 on
> Chlamydia, either.
> 
> Bill
> 
> 

Re: dd vs. truncate for creating vkernel root images

[Chris Turner]

Nuno Antunes wrote:
> Hi all,
> 
> Is there any advantage in using
> 'dd if=/dev/zero of=/var/vkernel/rootimg.01 bs=1m count=2048'
> over the much quicker
> 'truncate -s 2G /var/vkernel/rootimg.01'
> other than getting a root image filled with zeros?

not sure if the 'filled with zeros' counts as the same thing,
but the former will pre-allocate the storage whereas the latter
will not:

  $ truncate -s 100M ./foocate
  $ dd if=/dev/zero of=./ddcate bs=1m count=100
  $ ls -l *cate
  -rw-r--r--  1 foonicator  staff  104857600 Apr 11 19:26 ddcate
  -rw-r--r--  1 foonicator  staff  104857600 Apr 11 19:25 foocate
  $ du -sk *cate
  102448  ddcate
  48  foocate

so it depends on your preference if you want to pre-allocate or
defer allocation.

Re: dd vs. truncate for creating vkernel root images

[Chris Turner]

Chris Turner wrote:
> so it depends on your preference if you want to pre-allocate or
> defer allocation.

although I just remembered the 'skip' argument to 'dd' does the same..

libm update plans?

[Chris Turner]

Cheers all..

The masochist in me was trying to get some mildly glibc specific
sound apps to run on df, and discovered that the C99? 'trunc'
function has not yet been ported over from NetBSD's math library

This leaves me with 4 options:

1) whine here to ask if anyone has any update plans:

   "waah! Does anyone have any update plans?"

   from digging around it looks like trunc() is in NetBSD
   (checked only v1.46 of math.h)

   http://cvsweb.netbsd.org/bsdweb.cgi/src/include/math.h?rev=1.46

2) Hack in a private copy to the particular package in question

   (sheer lazyness!)

3) port over myself (see #1, #2 :)

4) find something else to do for kicks

will probably put this lower in my ever-growing todo list for the moment
if not #1, but just thought I'd ask..

Thanks,

- Chris

Re: libm update plans?

[Chris Turner]

Matthew Dillon wrote:

> 
> I looked for various pieces of example code.  FreeBSD has a really
> old implementation in (libmsun) which is not suitable, and NetBSD uses
> the same implementation.  Gnu seems to have a clean 4-line implementation
> done in 2005:
> 
> http://gcc.gnu.org/ml/fortran/2005-05/msg00227.html

1) haven't read this yet, to keep ye-olde eyes clean :)
2) just assumed that since DF was rebasing things from netbsd
   (based on df commit messages), that this was added new after the fact

good summary of netbsd's changes here:

http://cvsweb.netbsd.org/bsdweb.cgi/src/lib/libm/shlib_version

although based on a quick glance at:

http://cvsweb.netbsd.org/bsdweb.cgi/src/lib/libm/src/s_floor.c
($NetBSD$ v1.11)

http://cvsweb.netbsd.org/bsdweb.cgi/src/lib/libm/src/s_trunc.c

($NetBSD$ v1.2)

http://www.netlib.org/fdlibm/fdlibm.h

( /* @(#)fdlibm.h 1.5 04/04/22 */ )

and

http://www.freebsd.org/cgi/cvsweb.cgi/src/lib/msun/src/s_trunc.c

($FreeBSD$ v1.1)

It looks like this was a fresh implementeation for FBSD 5 using
a similar structure (and misnamed header) from the s_floor.c
brought in to netbsd since 3.x release

aside from all the various conditionals, length, and so on,
what do you find unsuitable about this version?

gak!

the whole point of mentioning it was so I didn't spend hours doing
this ^ .. oh well. that's what I get for being lazy - hopefully some
good will come of it.

>
> Are you sure this is part of C99?

was shamefully baseing this off of the GNU manpage on a linux box
(hence '?')..  anyhoo, someone did the real homework here..

Like I mentioned, this is further down in my list of priorities at the
moment (e.g. not-started external vkernel shutdown & other ideas, etc
much more fun) ..

Let me know if I should file a bug or similar to track things,
otherwise I'll probably leave in the back of my mind to gather dust.

Thanks,

- Chris

Re: libm update plans?

[Chris Turner]

Simon corecode Schubert wrote:
> 
> yes, maybe you can post your analysis to bugs@ so that it gets recorded.
> 

okay.. all set.

I hereby wash my hands of this,
at least until it comes back to haunt me later :)

Re: libm update plans?

[Chris Turner]

Hasso Tepper wrote:

> Maybe you should pick up DragonFly C99 Standards Conformance Project? ;)
> 
> http://www.in-nomine.org/~asmodai/df/conformance/
> 
> 

gak!

wheres the trunc() ?

like I said.. wayyy down the list for me..

good to know this page exists though..

C-- (was: Re: Development)

[Chris Turner]

Simon corecode Schubert wrote:
> 
> However, there won't be any C++ in the list of things to do.  The only thing
> which uses C++ in DragonFly is groff, and I would be happier without this.
> 

was just thinking about this issue - any thoughts about:

  http://heirloom.sourceforge.net/doctools.html

Seems to be mostly under CDDL[1]_, which might be 'yet another'
license to throw into the mix, but based on a quick scan (IANAL),
doesn't seem much different than the GNU/Groff combination, plus
it's more like 'real' **ix  ..

Probably not anyone's top priority .. and from what I understand
groff is a bit 'smarter' about processing pipelines, so this might break
various pkgsrc items that expect a groff, but in any case, would remove
C++ code from base. Plus there's grap(1)!

- Chris

.. [1] http://www.opensolaris.org/os/licensing/opensolaris_license/

Re: C--

[Chris Turner]

Sascha Wildner wrote:
> 
> Hmm, looks interesting. I'll play with it...
> 
> Sascha
> 

p.s:
just discovered the 'utlities' portions are in pkgsrc-wip,
but the 'documentation tools' don't appear to be.

Re: umodem? driver

[Chris Turner]

Matthew Dillon wrote:
> 
> Try this:
> 
> cd /dev
> ./MAKEDEV ucom0
> 
> And then see if you can talk to it via ucom0.
> 

FWIW,

I had to do this personally for a GPRS device,
so I know it worked for me at one time..

(a.k.a. not sure how the 'default' devices are decided .. hint hint)

Re: cardbus/pccard patch

[Chris Turner]

Sepherosa Ziehau wrote:
> Hi all,
> 
> Following patch updates our cardbus/pccard support:
> http://leaf.dragonflybsd.org/~sephe/cbb_062307.diff
> 
> Developers and users that have laptops and run HEAD, please test it.
> I want to commit it before upcoming release.
> 

Just tried this out, and it seems to improve the situation with my
Sony/Ericson GC89 / Broadcom BCM4344 GPRS/EDGE Modem discussed earlier[1].

Thanks! much better! just got a day or two back I'm sure :)

As that's the only cardbus card I've got, not sure if thats a good
"test", but feel free to apply your updates from where I'm sitting :)

cbb-related portions of boot -v attached - this was a pci cbb adapter,
I've not tried my laptop as it's running -RELEASE.

As far as sio(4) goes:

A 'clean' version of the sio(4) patch for the card is attached -
It probed OK, and I was able to talk via cu(1), but there's some
strangeness where after entering things like e.g. 'AT', most of
the time I have to hit a second '' before I get the 'OK' back
from the card, but somtimes it returns immediately.

There's also the occasional:

   sio4: 1 more silo overflow (total 1)

message. As such my ppp isn't working yet.. From watch(8)ing the
chatscript, it looks like ppp is timing out waiting on the OK reponses.
Didn't try forcing '\n' in the chats though (seems incorrect..), and
unfortunately I'm not enough of a RS-232 / ppp / Hayes modem guru to
know if I'm configuring things incorrectly or if this is something
that needs fixing in the driver.

For now, I'll try the various device attach flags relating to the
on-chip queues and see if anything helps there with the basic 'cu' test.
But.. If anyone has any suggestions there, like from experience with
dialup ISP's *ahem* :) please let me know...

Thanks in advance,

- Chris



(btw, thanks Matt last time for the "pointers" ha ha on checking the
 card addresses. Didn't get around to trying things yet for this one,
 which might have been a good thing (TM) here ... but in the future,
 good to know...)

[1]http://leaf.dragonflybsd.org/mailarchive/kernel/2007-05/msg00063.html

Index: dev/serial/sio/sio.c
===
RCS file: /var/local/apps/dcvs/src/sys/dev/serial/sio/sio.c,v
retrieving revision 1.41
diff -u -p -r1.41 sio.c
--- dev/serial/sio/sio.c	9 May 2007 00:53:34 -	1.41
+++ dev/serial/sio/sio.c	23 Jun 2007 16:42:22 -
@@ -357,6 +357,7 @@ 	{ 0x95211415, "Oxford Semiconductor PCI
 	{ 0x7101135e, "SeaLevel Ultra 530.PCI Single Port Serial", 0x18 },
 	{ 0x151f, "SmartLink 5634PCV SurfRider", 0x10 },
 	{ 0x98459710, "Netmos Nm9845 PCI Bridge with Dual UART", 0x10 },
+	{ 0x434414e4, "Broadcom 4344 GPRS/EDGE Modem", 0x10 },
 	{ 0x, NULL, 0 }
 };
 
@@ -3052,6 +3053,7 @@ }
 
 DRIVER_MODULE(sio, isa, sio_isa_driver, sio_devclass, 0, 0);
 #if NPCI > 0
+DRIVER_MODULE(sio, cardbus, sio_pci_driver, sio_devclass, 0, 0);
 DRIVER_MODULE(sio, pci, sio_pci_driver, sio_devclass, 0, 0);
 #endif
 #if NPUC > 0
Index: bus/pci/pcidevs
===
RCS file: /var/local/apps/dcvs/src/sys/bus/pci/pcidevs,v
retrieving revision 1.27
diff -u -p -r1.27 pcidevs
--- bus/pci/pcidevs	3 May 2007 07:58:06 -	1.27
+++ bus/pci/pcidevs	23 Jun 2007 16:43:23 -
@@ -1141,6 +1141,7 @@ /* Broadcom Corporation products */
 product BROADCOM BCM4306_2	0x4321	BCM4306 802.11a Wireless Lan
 product BROADCOM BCM4309	0x4324	BCM4309 802.11a/b/g Wireless Lan
 product BROADCOM BCM4306_3	0x4325	BCM4306 802.11b/g Wireless Lan
+product BROADCOM BCM4344	0x4344	BCM4344 GPRS/EDGE Modem
 product BROADCOM BCM4401	0x4401	BCM4401 10/100 Ethernet
 product BROADCOM 5801		0x5801	5801 Security processor
 product BROADCOM 5802		0x5802	5802 Security processor
... 
found-> vendor=0x1180, dev=0x0475, revid=0x81
bus=0, slot=15, func=0
class=[BRIDGE]06-07-00, hdrtype=0x02, mfdev=0
subordinatebus=0secondarybus=0
intpin=a, irq=255
...
cbb0.pci0.pcib0.legacypci0.nexus0.root0
cbb0:  [tentative] at device 15.0 on pci0
cbb0: lazy allocation of 0x1000 bytes rid 0x10 type 3 at 0x8000
cardbus0.cbb0.pci0.pcib0.legacypci0.nexus0.root0
cardbus0:  [tentative] on cbb0
cardbus0:  [attached!] on cbb0
pccard0.cbb0.pci0.pcib0.legacypci0.nexus0.root0
pccard0: <16-bit PCCard bus> [tentative] on cbb0
pccard0: <16-bit PCCard bus> [attached!] on cbb0
pci_cfgintr_search: linked (62) to configured irq 11 at 0:14:0
pci_cfgintr: 0:15 INTA routed to irq 11
cbb0: PCI Configuration space:
  0x00: 0x04751180 0x0217 0x06070081 0x00022000 
  0x10: 0x8000 0x02dc 0x20030200 0xf000 
  0x20: 0x 0xf000 0x 0xfffc 
  0x30: 0x 0xfffc 0x 0x0700010b 
  0x40: 0x 0x0001 0x 0x 
  0x50: 0x 0x 0x 0x 
  0x60: 0x 0x 0x 0x 
  0x70: 0x 0x 0x 0x 
  0x80: 0x0001 0x 0x04630

bcm sio on cbb (Re: cardbus/pccard patch)

[Chris Turner]

Chris Turner wrote:
> A 'clean' version of the sio(4) patch for the card is attached -
> It probed OK, and I was able to talk via cu(1), but there's some
> strangeness where after entering things like e.g. 'AT', most of
> the time I have to hit a second '' before I get the 'OK' back
> from the card, but somtimes it returns immediately.
> 
> There's also the occasional:
> 
>sio4: 1 more silo overflow (total 1)
> 

disabling the other sio? at isa... lines seems to help the 'enter'
problem. Still working on ppp, but *gasp* had the brilliant idea to
check against other os'es :) So I'll do that first ..

I think I was so excited it probed that I forgot to think :)

- Chris

User-Land PPP vs. Free/Open

[Chris Turner]

While trying to investigate the SIO/PPP problems,
I came across this:

http://osdir.com/ml/os.openbsd.tech/2004-11/msg00085.html

Which seems to suggest that the DF implementation of user-PPP
is a bit behind w/r/t other OS'es.

That particular patch applied cleanly, appears to be in Free/Open,
and did absolutely nothing to fix my problem :)

I tried the kernel version, but it didn't work immediately and the
interactive mode of user-ppp seems a bit more 'debugging friendly'..

so anyway, I'll be setting up an OpenBSD box to troubleshoot /
diff the debug logs, etc.

But, it made me think -

if there is anything missing bug-wise in the ppp,
how would I go about submitting a patch ?

Thinking either:

  - diff with free and see what happens when I apply the patch (BAD!)
  - Manually merge changes against 2 known versions and submit the diff
( Better. Best? )

My other submits have been 'new code', so the 'synch procedure' hadn't
been an issue.. the 'Development' page on the wiki give kind of a
technical how-to, but not necessarily procedural..

Perhaps I should ask again, when I have a patch from another source,
but that's more than likely a matter of time .. :)

Also, anyone planning on a massive ppp synchup anytime soon ?

Thanks,
- Chris

Re: User-Land PPP vs. Free/Open

[Chris Turner]

walt wrote:
> 
> That patch is from 2004.  Have you been to the ppp maintainer's
> website to check for newer code?
> 
> http://www.awfulhak.org/ppp.html
> 

Thanks for the pointer -
started to.

Diffed freebsd-CURRENT and there were 4k lines of diff,
so rather than wade through it, I tried that particular patch,
which applied cleanly to the DF code.

I probably will check ppp-upstream later depending on testing
results if I can't fix the problem as the patch didn't work..

The portions related to the obsd diff here seems to have been applied to
FreeBSD post-fork however, and still needed, which brings up the more
general question of maintenance, etc than my ongoing ppp
'investigation', which is more what I was getting at when I posted..

which brings up a side question of - would it me more appropriate to
synch with awfulhack or fbsd?, etc, etc..

but thanks again, probably worth checking out, and should have mentioned
I'd looked there.

Thanks,
- Chris

Re: User-Land PPP vs. Free/Open

[Chris Turner]

Simon 'corecode' Schubert wrote:
> Probably you first should find out what's wrong and then fix it :)  If
> you suspect code (i.e. FreeBSD works, DragonFly doesn't), then do this:
...
> - if works, submit

Thanks for the suggestions , will see which approach seems more appropriate.

> 
> 4kloc are not so much, especially if it is a unified diff.
> 

This is true -
But still 4k more than I wanted to deal with at the time :)

thanks again,

- Chris

Re: Apache in pkgsrc has largefile support on?

[Chris Turner]

Granted, I don't think that many people were thinking in gigabytes
then, but:

http://minnie.tuhs.org/UnixTree/32VKern/usr/src/sys/h/param.h.html

and also:

  "To insure that it is possible to create files as large as 2^32
bytes with only two levels of indirection, the
minimum size of a file system block is 4096 bytes. The size of file system
blocks can be any power of two
greater than or equal to 4096." [1]

So I don't think Linux *invented* this UNIX(R)-ism, per-se -
although they might have copied it's behavior :)

HPUX11 requires that you specify 'mkfs -o largefiles / mount -o
largefiles' - and that was released in '98 fer some pretty beefy hardware..

but,:


Of course, our lab used Berkeley Unix, as do all right-thinking
folks. East Coast people were said to be biased towards AT&T Unix, but
then, they hadn't discovered hot tubs either.
[...]
"We're watching someone who's never used Berkeley Unix." He sucked
in his breath and whispered, "A heathen." [2]


As true as ever, 20 years on, imho :)

--

[1] Marshall Kirk McKusick, William N. Joy, Samuel J. Leffler and Robert
S. Fabry. A Fast File System for UNIX. Technical Report Computer Systems
Research Group, Computer Science Division, Department of Electrical
Engineering and Computer Science, University of California, Berkeley,
Berkeley, CA 94720.

[2] The Cuckoo's Egg: Tracking a Spy Through the Maze of Computer
Espionage, Clifford Stoll, 1989, ISBN 0-7434-1146-3

anyone using tw / xten / xtend for X10 devices?

[Chris Turner]

I stumbled across this the other day and it looks pretty nifty,
but would like to here positively from someone that the code is still
in good shape before buying some hardware...

Thanks,

- Chris

SMP vkernel speed on UP host

[Chris Turner]

I know UP hosts are so passe' these days,

But does anyone else think SMP vkernels are a bit slow
on UP hosts?

possibly due to the 'io processor' & context switching, methinks.

if correct, not sure if this is worth any code changes or
manual entry notes..

or perhaps I should just update my test network to the GHZ range .. :)

thoughts?

bridge(4) and pf tagging

[Chris Turner]

According to:

http://www.openbsd.org/faq/pf/tagging.html

it is possible to do ethernet-level pf tagging with OpenBSD's
bridge(4) device.

Being guilty of not having looked at the code myself,
but noting the lack of 'brconfig'

( http://www.openbsd.org/cgi-bin/man.cgi?query=brconfig&sektion=8 )

anyone have any idea of what it would take to get this ported over?

e.g. how far has the DF bridge & pf drifted from OpenBSD ..

Also, is anyone successfully using PF tagging in their rulesets on DF?

(not to imply it's broken, but just wondering..)

will probably investigate at some point,
just wanted to throw it out there..

Thanks,

- Chris

Re: Kerneltrap interview available

[Chris Turner]

Matthew Dillon wrote:
> Jeremy did another interview of me in Kernel Trap, here's the URL:
> 
> http://kerneltrap.org/node/14116
> 
>   -Matt

hey so what happened to the "Hammer" FS name ?

> JA: Does your new filesystem have a name?
> Matthew Dillon: Not yet. 'DFS' is taken unfortunately.

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

[Chris Turner]

[EMAIL PROTECTED] wrote:
> 
> Ive still got to solve one little problem which is, is there a way to run
> spamlogd on start up via rc.conf, such as spamlogd="YES" ?
> 
Not sure if this is the case for spamd,
but the apache httpd installed an rcng script into /usr/pkg/etc/rc.d/,
I just copied into /etc/rc.d/, configured rc.conf, things were OK.

You might pkg_info -L to see if spamd has a similar script..
if not, RCng scripting is pretty easy & I'm sure a pkgsrc patch
would be welcome..

- Chris

Re: Domain concept?

[Chris Turner]

Sdävtaker wrote:

> Can u please point me to some place with a nice explanation? I am
> googling it but only find some technical stuff about them and no the
> concept involved.
> Thanks for any help
> Sdäv

The main hints comes from resolver(5):

-8<-
On a normally configured system this file should not be necessary.  The
only name server to be queried will be on the local machine, the domain
name is determined from the host name, and the domain search path is
constructed from the domain name.
...
If no nameserver entries are present, the default is
to use the name server on the local machine.
...
Local domain name.  Most queries for names within this domain
can use short names relative to the local domain.  If no
domain entry is present, the domain is determined from the
local host name returned by gethostname(3); the domain part
is taken to be everything after the first `.'.
...
Search list for host-name lookup.  The search list is nor-
mally determined from the local domain name; by default, it
contains only the local domain name.  This may be changed by
listing the desired domain search path following the search
keyword with spaces or tabs separating the names.
->8-

Most of this documentation goes back to the original resolver
manual page (from 4.2BSD or 4.3BSD, I forget)- I was confused about this
issue myself & did the research :)

It looks like originally, the expectation was to run named
locally & have your hostname-implied search path match your
DNS topology - still probably a good idea today..

If you think of the machines at the time this page was probably written
('84 or so), your site would probably be running a timesharing mini
directly on the Internet rather than a lan workstation so this makes sense..

Unfortunately nowhere that I know of in the (manual page) documentation
does it directly say "Set your hostname to your FQDN and run named", so
you have to kind of scratch around to understand the assumption..

It also makes it mildly harder to stop SMTP relaying with some
mail servers if you don't set to FQDN.

Re: Screenshots for dbBSD desktop

[Chris Turner]

Justin Chan wrote:
> Hi folks,
> 
> Anyone has the screenshots for the dbBSD desktop.
> 
> For those who still couldn't get the X and WindowsManager up at least,
> they know it can be done.

http://themes.freshmeat.net/

Re: dfly mail archive not working

[Chris Turner]

Matthew Dillon wrote:
> 
> oops.  Ok, fixed.  I keep meaning to fix the rc script.  I'll just
> add a manual start in rc.local.
>  

If this is pkgsrc apache, there should be an RCng init script in
$PREFIX/etc for your convenience.. I had to hack it to add
'${apache_flags}' for a diferent wwwroot ..
Don't think patch has been committed yet, will get around to following
up oneday

Index: apache.sh
===
RCS file: /cvsroot/pkgsrc/www/apache/files/apache.sh,v
retrieving revision 1.27
diff -r1.27 apache.sh
39c39
<   ${ctl_command} ${action}
---
>   ${ctl_command} ${apache_flags} -k ${action}


Not knowing the leaf setup this may be extra info though..

Re: dfly mail archive not working

[Chris Turner]

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

Relates to apachectl argument processing.. With no flags "apachectl
[start|stop]" works fine but with flags it does not (for example
"apachectl -D FOO_IS_DEFINED start") fails. The -k works in either case
for start/stop.. but on closer inspection looks like it might break some
other apachectl features.. (e.g. configtest)

Probably I need to rework the patch to test for flags depending on the
argument, but thats getting a bit complex for my needs..

- Chris

Re: dfly mail archive not working

[Chris Turner]

Joerg Sonnenberger wrote:
> 
> Well, let me rephrase the question. Why can't you just use -k inside
> apache_flags?
> 

that requires the user of the flags to specify '-k'
as the last argument, which seemed incorrect

Re: libgtop2 and xorg break

[Chris Turner]

Pieter Dumon wrote:
> pkgsrc/sysutils/libgtop2 doesn't compile on my fresh 1.10 system
> (don't know if it did before):
> 
> proclist.c: In function 'glibtop_get_proclist_p':
> proclist.c:106: error: structure has no member 'kp_proc'

The error is probably due to DF struct proc diverging from freebsd
somewhere along the line. Probably just a few lines of diff would get
things working, but that may be more than you feel like doing..

I seem to recall this problem on either 1.6 or 1.8 for similar reasons,
but am not sure

Re: Disklabel Question

[Chris Turner]

Adrian Michael Nida wrote:
> 
> We're getting closer!
> 
> leviathan# disklabel -w ad4s0 auto

try '-r -w ad4s0 auto' here instead -
I think it needs to be 'read' before it's edited.

This could be wrong too -
I still get confused occasionally & have been using for some time.
it's a pain, but a flexible one.

Re: Disklabel Question

[Chris Turner]

Simon 'corecode' Schubert wrote:
>>
>> try '-r -w ad4s0 auto' here instead -
>> I think it needs to be 'read' before it's edited.
> 
> -r is "raw" :)

aaiight.. I promise to RTFM before posting.

since I'm spamming anyway - this perlish 'context dependant' -r thing
is way silly.

bah!

Re: NFS scalability

[Chris Turner]

steve wrote:
> On Mon, Sep 10, 2007 at 07:49:51AM -0700, Matthew Dillon wrote:
>> :Just out of interest (without the time/boxes to try it out myself), do 
>> :you guys think this is a "good" solution? I mean, does NFS handle a lot 
>> :concurrent accesses well? I never used NFS...
>> :

haven't played around with it,
but you can do interesting things with amd(8) maps to distribute the
load, handle failure, etc

I think this would be best suited for 'static' into (e.g. non-time
critical html, php scripts with only logic, etc) - as Matt points
out there are cache problems on NFS (have seen this occasionally first
hand when serving up NFS builds on a lan)

Re: Filesystem encryption

[Chris Turner]

Matthew Dillon wrote:
> 
> I'm not planning on it for 2.0, just getting HAMMER's base
> functionality implemented is already pretty challenging.  But there's
> nothing particularly difficult about the concept.
> 

lo.. been AWOL awhile but lurking.

anyway, what's the take on a simple hook for vn/vnconfig or similar.
seems like that would be a simple way to get the basic functionality
without too much VFS muckery..

2c

Re: Filesystem encryption

[Chris Turner]

Matthew Dillon wrote:

> 
> That would actually work, though it would be kinda hokey.
> 

yup. not as hokey as having your data stolen I suppose :)

my memory was rusty -  had looked into this previously &
looks like the open 'svnd' does as mentioned above:

http://www.openbsd.org/cgi-bin/man.cgi?query=svnd&sektion=4

goodness on the topic and various approaches here:

http://www.onlamp.com/pub/a/bsd/2005/12/21/netbsd_cgd.html?page=1

and here:

http://www.imrryr.org/~elric/cgd/cgd.pdf

if anyones interested.

- Chris

Re: Hardware compatibility

[Chris Turner]

Aggelos Economopoulos wrote:
> Hello all,
> I'll be buying parts for an AMD-based system tomorrow and I'd appreciate info 
> on what motherboards/chipsets you have booted DragonFly on without problems. 
> Problems you've encountered are helpful also. Hopefully your answers will be 
> useful to other buyers too (at least for the next few months).
> 
> TIA,
> Aggelos

Still honing in on these, my priorities have unfortunately been
elsewhere of late. in any case:

- Tyan Tomcat S3970-U w/1x Dual Core Opteron

  SMP OK with APIC_IO as well. HT1000 chipset, SATA seemed ok
  both in sata and in 'legacy' IDE mode, although I did have
  a strange filesystem corruption problem at one point.. not enough to
  go on of yet - machine was loaded via homebrew rconfig scripts
  using 'ftp |gunzip -f -| restore', which easily could have been the
  problem.. GigE ok via em(4). Didn't try non text graphics modes,
  but I suspect VESA stuffs should be ok. not much else on the board
  since it's a server, although I'm looking forward to investigating
  the new sensors port..

  still need more benchmarks / stress tests, planning on these shortly,
  haven't really run the machine for more than a few hours at a time
  yet, but it boots.

- Acer Ferrari 1000 w/ Turion 64 x2

  SMP OK with APIC_IO off.. ATI Radeon Xpress 1150 chipset. At one
  point, I had some problems with NATA (and therefore using the
  drive in SATA mode), but this was about 1/2 way through the 1.8->1.10
  release cycle, and I haven't had a chance to update this machine
  yet since I need to use it constantly. WMA, ACPI suspend, etc.
  was in a similar state (e.g. didn't work, but have been lots of
  positive improvements in tree since then). Overall, quite reliable
  for day to day user, and on-board bge works fine too. Haven't had
  a chance to try the new broadcom wireless driver. Will be
  updating this one soon as well - missing out on too much goodness :)

Hopefully I'll have a chance to focus more on testing/digging into
the various chips on these machines over the next few months as the
project I'm working on is winding down.. fingers crossed.

This brings up a question:

any thoughts on a dmesg@ or similar ala [1] ? such a database
might be useful..

I think I need to shut up and code more.. anyways - good luck..

- Chris

.. [1] http://www.openbsd.org/faq/faq4.html#SendDmesg

Re: Filesystem encryption

[Chris Turner]

Chris Turner wrote:
> Matthew Dillon wrote:
> 
>> That would actually work, though it would be kinda hokey.
>>
> 
> my memory was rusty -  had looked into this previously &
> looks like the open 'svnd' does as mentioned above:
> 
> http://www.openbsd.org/cgi-bin/man.cgi?query=svnd&sektion=4
> 

Missed the updated entry - hot off the presses !

http://www.openbsd.org/cgi-bin/man.cgi?query=mount_vnd&sektion=8&format=html

Using this to simplify the vn disk UI seems like a good idea although
the concept of 'mounting' an individual /dev/ entry makes me a little
queasy .. need to update my Open machines and check it out I suppose..

Possibly will peruse this over time.. gearing up for sio investigation
followed by a little vkernel.conf action first though.. need more hours
in the day!

Adding ad[4-6]s* to default devices

[Chris Turner]

Any thoughts?

More systems are coming out with 2 controllers - a PATA and a SATA,
and as a result there might be more users using ad4 and so on..

I know there's the MAKEDEV.local hook, just thought I'd mention it since
I keep having having to remember to MAKEDEV on some my systems after
installing them. If OK I'll probably go ahead and make the change.

Thanks,
- Chris

Re: Adding ad[4-6]s* to default devices

[Chris Turner]

Matthew Dillon wrote:
> 
> We should be creating them through ad7 already.  Check HEAD.  If it isn't
> doing it properly then we definitely want to adjust it.
> 

Okay - I think I asked my question incorrectly based on the (bad)
assumption that this was being done for the lower numbered 'ad' units..

I meant if we should add partition devices to more slices, e.g:

ad4s0[a-p] is present, s1[a-p] is present, but s2 - s4 only have the
'whole slice' device

current HEAD policy creates "{ad[1-7]s[01][a-p],ad[1-7]s[0-4]}"

moused + modular xorg 'sticking' ?

[Chris Turner]

Hello all -

fishing for similar experiences -

has anyone had any problems with ps2 mice getting 'stuck' under X?

just updated my 1.8 laptop to -HEAD before the Hammer work started
+ 2007Q3 Pkgsrc (modular-xorg-server-1.3.0) this seems to happen,
but only when using sysmouse or perhaps even sysmouse with radeon..

I tried several variations of acpi/noacpi/sysmouse/psm/radeon/vesa
and even reverted to the previous 1.24 psm.c, but it seems like
only using the direct /dev/psm0 will work 100% with this particular
combination since the upgrade.. or perhaps this is something in xorg
(was previously on monolithic 2007Q1 build)

In any case, since there's so many factors need to do more tests to zero
in on this for sure, but thought I'd ask..

Thanks,

- Chris

Re: moused + modular xorg 'sticking' ?

[Chris Turner]

Matthew Dillon wrote:
> 
> 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.
> 

Aaaight ..

I took a naive look at trying your idea, and my stack-discipline-foo
isn't up to snuff enough to even submit a buggy patch for the time
I have to work on it tonight..

in any case,

I suppose I'll save my notes, sulk away with my tail between my legs,
summarize everything said here into a coherent bug report,
and stick with userland patches for a while :)

"good" to hear others have the problem and there's something to try in
the future - wish I could do more to fix it right now.

Thanks,
- Chris

Re: moused + modular xorg 'sticking' ?

[Chris Turner]

Chris Turner wrote:
> Matthew Dillon wrote:
>> 
>> Should we start saving and restoring the FP context?  The ucontext
> 
> I suppose I'll save my notes, sulk away with my tail between my legs,
> summarize everything said here into a coherent bug report,
> and stick with userland patches for a while :)
> 

one other little detail before I do this -

does anyone know what posix has to say about fp in signal handlers?

perhaps this is a 'bug' w/r/t standards in xf86-input-mouse ...

Re: Should I jump into DFBSD?

[Chris Turner]

Jeremy C. Reed wrote:
> 
> I don't use it myself, but what about the video driver for NVIDIA cards 
> for the modular Xorg server in pkgsrc/x11/xf86-video-nv and available as a 
> DragonFly package.
> 

yes..

not using it here presently, but I'd think 'nv' or even 'vesa' would
work for regular 2d bitmapped graphics.. 3D + GLX is another story

Besides, if you're not running it,
how will you find a way to kludge the driver to work ?

:)

reeally hope the ATI/AMD specs open soon as theyre supposed to..

snd_hda

[Chris Turner]

Finally got around to testing the recent updates..

This works like a charm on a machine that was currently not playing
anything (intellegable at least)..

Thanks to all involved in fixing it !!

Re: Blacklisting (and blocking) remote sites - blt.tar.gz (0/1)

[Chris Turner]

> >
> >Like I said, there are many different ways and applications out there,
> >enough for everyone to find the one that fits. But I didn't see no
> >reason to install phython or whatever, just to have some big
> >application maintain my /etc/hosts.allow.
>

If anyone ever gets around to updating our PF as I would hope to do,
this can be done mostly automatically via tables..

see: http://johan.fredin.info/openbsd/block_ssh_bruteforce.html 

etc ..

Re: Mounting dragonfly ufs in freebsd

[Chris Turner]

Jonas Trollvik wrote:


I know this might be a little off topic for the DragonFly lists but
maybe someone could give me a pointer in the right direction to be
able to mount d-g.



kind of a PITA, but if you can read 'a', you could keep 2 labels in
'a', and relabel it as required..

(as long as the label doesn't extend into the partition &
  hose things up ..)

if the filesystem is working, it's probably just the label that has 
problems..


If you're just testing driver compatibility now & again,
I'd think that would might be good enough..

Re: rsync vs. cvsup benchmarks

[Chris Turner]

I have some benchmark test results comparing rsync to cvsup.  


...

okay.. so like:

you'd think with all of these repository copies flying around,
there'd be a lot less flaming and a lot more coding going on..

enough!

sheesh.. You people are making me want to write this email

IN EMACS

and,

WE ALL KNOW HOW HORRIBLE EMACS IS !!!

especially

WHEN COMPARED TO VI !!!

and also, just to finish this off (for now:)

People on each side are LIKE NAZIS

etc, etc, etc.

Re: Introduction

[Chris Turner]

Sascha Wildner wrote:


Hmm, am I the only one who is glad that we _don't_ have fancy codenames 
for releases?




mmm.. no

but this is >= 2^1024 times better than ubuntu release names

unless it is decided to add a silly adjective to the species

like: "Magical Medowhawk"

eek.. stomach turning over

so, no.

Re: Installation on Macbook Pro

[Chris Turner]

Christopher Rawnsley wrote:
first step is to partition the drive. I choose my hard drive which is 
detected as being 'ad4' (which struck me as odd. I thought they started 
at 0?). I have a 120GB drive of which I want to install DragonFly to a 
15GB slice. The installer tells me that it format is successful and then 
asks how I would like to set up my partitions (or sub-partitions). I 
stuck with the default and continue. Then get:


Execution of the command
/sbin/disklabel-r ad4s3
 >/tmp/install.disklabel.ad4s3
FAILED with a return code of 4.



Disclaimer: I don't have a macbook.. (or any intel mac)

ad4 : I've seen this as the first HDD on an Acer notebook, a shuttle 
xPC, and a tyan server board - In my case it had to do with both the 
presence of a legacy parallel ATA controller (ad0-3), and a SATA 
controller (ad4-..) . If e.g. the CD-ROM/DVD/etc. is on PATA (which I 
suspect to be the case), it is likely the reason that the first HDD is 
showing up as ad4.


In my case, it hasn't been any kind of problem.

For the disklabel problem:

typically I haven't labelled individual slices, but on occasion, it 
seems like disklabel -r gets confused if there is some residual data 
there.. you might try a manual install (see /README on the CD) and 
zeroing out that portion of the disk


e.g: dd if=/dev/zero of=/dev/ad4s3c count=32

(can't remember the exact size of the disklabel - I think 32 sectors 
should nuke it.. 1m will definately)


also:

I've unfortunately had some problems with native mode SATA which may or 
may not be related here (was HT1000 specific) - haven't had a chance to 
track down the bug. If it's possible to force 'legacy parallel' mode

in the apple machine, that might help things.

Sorry you ran into problems & good luck - and do let us know if you find 
something out

Re: pkgsrc DEVOSSAUDIO and DEVOSSSOUND (in oss.buildlink3.mk)

[Chris Turner]

Steve O'Hara-Smith wrote:



This led me to wonder if /dev/audio should have worked with the
current sound setup in DragonFly and what setting should be in
oss.buildlink3.mk.



not sure if/what/when this should be/was changed/ ... - but
depending on the outcome, could this be overridden in the platform 
makefile if necessary?


can't remember which includes what first..

might be worth checking the oss.buildlink3.mk history to see if this 
changed @ some point recently..


getting around to my personal bulk builds - I'm sure I'll encounter this 
and do more digging in the next few weeks unless you get it fixed it by 
then :)


- Chris

Re: DragonFly without net connection [Was: OpenBSD dhclient]

[Chris Turner]

Hasso Tepper wrote:


It's part of bigger set of issues called "DragonFly without net 
connection".




yeah..

This is problematic with RCNG - I tried putting a parameter in rc.conf
for something or the other to try and make this modular,

e.g. [ "$net_is_enabled" ] && foo_enable="yes" type of things..

but I kept forgetting to change this prior to reboots, etc and things
would get started inadvertently ..

on the one hand, you want foo_enable='yes' to be set so that 
/etc/rc.d/foo works, but on the other hand, you don't necessarily want

to start it at boot time ..

Anyone have any working 'multi configuration' setups for rcng?

I contemplated hooking in loader.conf somehow to make some menu choices, 
but didn't quite find the time ..


still miles ahead of monolithic rc.local or SysV style links, but still 
room for improvement


and please, no suggestions on reimplementing launchd ... *shudder*
and especially no suggestions on reimplementing solaris SMF ... *vomit*

Re: Building modular xorg

[Chris Turner]

Dave Hayes wrote:


This actually failed differently (I download the vulns file nightly)
than before, before I got this:

===> Building binary package for libXrender
...
ERROR: This package has set PKG_FAIL_REASON:
ERROR: Do not include x11.buildlink3.mk for X11_TYPE != "native".
*** Error code 1



Disclaimer: I haven't done my 1.12 builds yet, but..

on 1.10 / 2007Q3, my bulk build had the following:

meta-pkgs/modular-xorg-apps \
meta-pkgs/modular-xorg-fonts \
meta-pkgs/modular-xorg-drivers \
x11/modular-xorg-server \

Although I suspect one of these will pull in modular-xorg-libs..

I ran into similar problems building some packages here and there -
basically they need to be converted to use modular xorg -

e.g. based on LIBS in the Makefile, they should

.include "../../x11/x11libname/buildlink3.mk"

instead of the previous:

.include "../../mk/x11.buildlink3.mk"

tracing through the various includes might yield something fixable,
thought I'm surprised if noone else has run into this ...

Thanks,

- Chris

pkgbox down?

[Chris Turner]

Hello -

trying to build my 1.12 release image but 'make installer fetch'
is failing - and ping is failing from leaf..

so it looks like pkgbox is down ..

Guessing this is 'a Matt',
but if anyone else also has a way to get it back online ...

Cheers

- Chris

Re: Building modular xorg

[Chris Turner]

Chris Turner wrote:

might-have worked by just using 'meta-pkgs/xorg' ..


okay.. this doesn't even exist anymore.. staying quiet until I know for 
sure :)


cheers

- Chris

Re: Building modular xorg

[Chris Turner]

Chris Turner wrote:


Disclaimer: I haven't done my 1.12 builds yet, but..

on 1.10 / 2007Q3, my bulk build had the following:

meta-pkgs/modular-xorg-apps \
meta-pkgs/modular-xorg-fonts \
meta-pkgs/modular-xorg-drivers \
x11/modular-xorg-server \

Although I suspect one of these will pull in modular-xorg-libs..



Going through various mk.conf's lying around - a pkgsrc-on-linux build
might-have worked by just using 'meta-pkgs/xorg' ..

or it could have failed & reverted to the method above..

yeah. anyway - probably best for someone else to answer this :)

- Chris

Re: Building modular xorg

[Chris Turner]

Dave Hayes wrote:

I'm surprised too. I have no idea how the pbulk build Justin is doing
works at all. ;)


actually - glad you reminded me -

seems to be working for me here (using 'legacy' bulk)

will send mk.conf in a separate email for your perusal

lang/python24 build problems

[Chris Turner]

Not sure if this or the pkgsrc-users list is more appropriate ..

Anyone having trouble building this?

I noticed it wasn't available from pkgbox in the 1.12 area..

(previous minor versions built fine on on <= 1.10)

looks like some errors possibly relating changes in socket structures or 
similar?


anyhow.. will probably take a more detailed look later, wondering if
anyone else has done so already

also, on a related note - are the bulk build logs available from pkgbox 
for comparison / fixing purposes?


Thanks,

- Chris


--
Chris Turner 199Technologies, LLC.
[EMAIL PROTECTED]

Re: lang/python24 build problems

[Chris Turner]

Thomas E. Spanjaard wrote:


Fixed in pkgsrc -current, and possibly 2008Q2 as well? The problem was 
some bluetooth preprocessor directives had a conditional only for 
NetBSD, but we use NetBSD's bluetooth stack as well.


Cheers - just saw this as socketmodule.c was patched in Q4 by pkgsrc for 
the netbsd case


hacking in a change locally.. looks like 08Q1 isn't branched from HEAD 
yet, so this will probably go in then.


cheers

- Chris

Re: Ruby crash on DragonFly

[Chris Turner]

[EMAIL PROTECTED] wrote:

#0  0x2809a251 in rb_eval (self=697870460, n=0x29049808) at eval.c:2927
2927eval.c: No such file or directory.
in eval.c

Any idea why it can't find eval.c? And how to resolve this? Very urgent,
need to fix this to finish off a project for a client.



disclaimer: I know nothing of ruby, or rails, or the app in question..

but:

I'd think that eval.c:2927 is the line of code in the ruby interpreter
hitting a problem where it can't find a file..

as someone else suggested, cracking open the ruby sources and taking a 
look at what is happening there might make sense..


since the ruby source file is called 'eval', my theory is that perhaps 
the application is trying to load & evaluate some ruby code from a file 
that doesn't exist ..


perhaps in a compiled ruby module.. or similar.

see disclaimer..

- Chris

Re: Ruby crash on DragonFly

[Chris Turner]

Joerg Sonnenberger wrote:

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


anyone have any familarity of how the interpreter works?

wondering if it does wierd things with sigaltstack, etc to handle code 
generation that is interating with the FP/signals work a while back..


anyway.. I'm sure we'll soon see..

- Chris

Re: Cvsup kernel source only?

[Chris Turner]

Andre LeClaire wrote:
cvsup-ing the kernel source only? This was possible with FreeBSD 4*, but 
all the info I've found describes getting the entire source tree for 
Dragonfly, which wouldn't fit in the space available.


Thanks!

Andre


just do a regular CVS checkout.. e.g. 'co src/sys'

good luck.

Re: Announcing the pkgsrc-2008Q1 Branch

[Chris Turner]

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.


Even having regular builds doesn't save us from this kind of stuff :(.
  



Not knowing the intricacies of various m4 implementations, it seems like 
we have a GNU mode with '-g'.. is there any reason devel/m4 is really 
needed for DragonFly? (e.g. some kind of pkgsrc wrapper makefile could 
be used around the system m4 instead?)


I suppose some tools would need to be modified, but for those things 
just needing 'gnu' m4 for *builds* ..


just a thought..

- Chris

Re: Troubles building m4 from pkgsrc

[Chris Turner]

Matthew Dillon wrote:


It's a temporary solution until we can come up with a way to propogate
our own pkgsrc tree but also keep it automatically synchronized with
the master, so we can incorporate needed patches without creating a
mess for our users.



not sure I mentioned it before, but :

anyone have any experience using $LOCALPATCHES ?

(http://www.netbsd.org/docs/pkgsrc/components.html#components.patches.sources)

seems like we could use this for in-package code related fixes..

not sure what to do about actual makefile diffs, etc...

maybe the pkgsrc folks would not be opposed to adding some kind of 
makefile-check-warning into whatever mechanism handles $LOCALPATCHES


e.g.

if $LOCALPATCHES/mypackage/Makefile
  warn about presence of overriding makefile
  or potentially .include it
...

of course, brings up the whole which-df-branch to match 
which-pkgsrc-branch thing again..

Re: Troubles building m4 from pkgsrc

[Chris Turner]

Steve O'Hara-Smith wrote:



not sure what to do about actual makefile diffs, etc...


Tricky - but patching the diffs isn't necessary LOCALPATCHES can
contain patches that apply to the file as patched by the standard patches.



this was meant to read more like:

"not sure what to do about patches to the Makefiles themselves"

I definitely wasn't trying to advocate patching-patches

Re: How can I start a daemon under DFly?

[Chris Turner]

Justin C. Sherrill wrote:


- seems to imply that the file /var/lib/dbus/machine-id needs to be
created.  'touch /var/lib/dbus/machine-id' may do it.  I haven't run d-bus
on my own, so this is a guess on my part.



might be that touching an empty file will work,

I've been running

dbus-uuidgen > /var/lib/dbus/machine-id

when this crops up and that seems to fix things too..

not sure what the 'right' way to do it is, just started checking the
whatis database for 'dbus' when this cropped up, and it seemed to
make sense to try it..

probably the rc script should be updated to do "right thing" 
automagically ..


(since e.g. sshd creates keys, pgsql creates a database ?!?, etc)

Re: How can I start a daemon under DFly?

[Chris Turner]

thegraze wrote:

ImportError: No module named gtkmozembed


looks like this is missing:

http://www.pygtk.org/pygtkmozembed/index.html

I'm guessing either you need to get pygtk installed
or perhaps it is not compiled with this module..

if the latter, might just need a CONFIGURE_ARGS+=
something in the makefile.. but it could be more

probably will have to dig around a bit further if
that's the case..

good luck!

- Chris

Re: New site design

[Chris Turner]

Justin C. Sherrill wrote:

I'm happy with the layout and content; I'm looking for further suggestions
for tweaks.  If there's no show-stopping objections, I'd like to move the
dragonflybsd.org site to this.


- 1 on the cosmetic changes - looks less cohesive & less professional 
imho - that being said it probably could be tweaked and work just fine


navigation is simplified for sure

Re: multiple ENVironments

[Chris Turner]

rhymes against wrote:


I need the following environment variables set to compile the courier
mta and authentication components, and unset to compile the dragonfly
os.  Is there a convenient way to set this up so I don't need to
remember to do and undo this every 6-8 months?


is this within pkgsrc or a separate build?

if separate - pkgsrc might simplify things if it works for you..

- Chris

Re: OT: setrlimit equivalent to prevent unlink or truncate

[Chris Turner]

Johannes Hofmann wrote:

Hi,

I'm wondering whether there is a way to prevent a process to modify 
the file system. setrlimit(RLIMIT_FSIZE) to 0 almost does the trick,

but unfortunately it does not prevent unlink() or truncate().
Is there any reason why there is no limit to prevent unlink or
truncate?



if it works, mounting the FS readonly should work..

also, chflags might be helpful..

or is this a coding question about coding the program that calls 
setrlimit() ?

Re: ASUS Eee compatibility

[Chris Turner]

Sepherosa Ziehau wrote:

On Mon, Mar 23, 2009 at 3:35 AM, Justin C. Sherrill
 wrote:

On Sun, March 22, 2009 12:36 pm, Michael Neumann wrote:

Am Freitag, 20. März 2009 16:02:48 schrieb Justin C. Sherrill:

On Fri, March 20, 2009 6:22 am, Michael Neumann wrote:

I have this ASUS EEE PC 1000H lying next to me, and DragonFly works

...

What's the wireless chipset in those?

It's an RT2860.

...

was hoping to get my eee 901 machine (fully) working with some BSD OS -
ideally dragonfly, but one step at a time..

since my dragonfly time is unfortunately much less than it should be,
some (hopefully accurate) notes compiled 2/15 w/r/t status in the 
various BSD's is attached..


cheers,

- Chris



Bsd-on-eee901 status


OS   Ethernet Wifi Sleep
=  =
FreeBSD  yes  no   yes   
OpenBSD  no   yes  maybe
NetBSD   no   no   yes?
DragonflyBSD no   yes  maybe

Drivers are 'ale' for ethernet, and 'ral' for wifi.

Plan


  - get all srcs on dev01
  - setup freebsd on eee02
  - build freebsd on dev01
  - merge openbsd ral driver changes to freebsd
  - migrate eee01 to freebsd
  - reinstall eee02 with dragonfly
  - migrate ale to dragonfly
  - migrate netbsd sleep/suspend to dragonfly
  - profit!!

Links
=


hardware: 

  - keeps my head straight


http://wiki.freebsd.org/AsusEee#head-0ec25ebfd5cc91803a27ffefe989f34ee19d887e
 
  - notes on 700 setup

http://nighthack.org/wiki/EeeBSD

misc:

  - bootable freebsd images

http://bsdimp.blogspot.com/2007/10/building-bootable-freebsdi386-images.html

manpages / driver source:

  - openbsd:


http://www.openbsd.org/cgi-bin/man.cgi?query=ral&apropos=0&sektion=0&manpath=OpenBSD+Current&arch=i386&format=html
http://www.openbsd.org/cgi-bin/cvsweb/src/sys/dev/pci/if_ral_pci.c?rev=1.14

for another chip supported by the same driver, which is in net/free bsd, 
for comparison:


http://www.openbsd.org/cgi-bin/cvsweb/src/sys/dev/ic/rt2661.c?rev=1.47;content-type=text%2Fplain

  - netbsd:

http://netbsd.gw.com/cgi-bin/man-cgi?ral++NetBSD-current

http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/dev/pci/if_ral_pci.c?rev=1.9&content-type=text/x-cvsweb-markup&only_with_tag=MAIN
  

same other chip:


http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/dev/ic/rt2661.c?rev=1.25&content-type=text/x-cvsweb-markup

  - freebsd:


http://www.freebsd.org/cgi/man.cgi?query=ral&apropos=0&sektion=0&manpath=FreeBSD+8-current&format=html
http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/dev/ral/if_ral_pci.c?rev=1.8

same other chip:

http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/dev/ral/rt2661.c?rev=1.27

Re: Instant crash with Linux OpenOffice

[Chris Turner]

no idea w/r/t hammer -

IIRC I had to muck around with some version of fonts.conf in the past -
perhaps the linux-emulator installed version -
to make things work for some OOO release or another,

e.g. change the directories listed within, or similar..

and then possibly run the linux 'fc-cache' to update..

can't recall..

Francois Tigeot wrote:
 > Aniway, I tried your trick of null-mounting fonts/ from /boot. It 
works for

reading documents but fails when trying to save them:

Re: Dragonfly used in production

[Chris Turner]

Matthew Dillon wrote:


From my perspective, then, this means that kernel-supported software
RAID is the way to go.  e.g. vinum (if it could be made reliable enough),
or we could work up our own solution.  The mechanics of a software
RAID system are not all that complex in actual fact.  In fact,

>

Had my eye on openbsd's softraid for a while - not sure how it would
integrate with cam and all.. but looks clean and simple, compared to
vinum / raidframe..

http://www.openbsd.org/cgi-bin/man.cgi?query=softraid&sektion=4&format=html

And of course no raid 5 or 6 / ? support, but as a pluggable 
architecture, seems like that will happen one day / could be doable.


anyhow.. yeah. many ideas, no time to code :(

- Chris

Re: bsd broadcom netxtreme 57xx compatibility

[Chris Turner]

Niklas Rosencrantz wrote:
> and this computer, 2nd in a small network where ifconfig outputs
> bge0: flags= mtu 1500
> options=1b
> ether [ip6 address]
> media:Ethernet autoselect (100baseT )
> status:active
>
> After /sbin/dhclient bge0 it gets no link...sleeping and active turns
> to status: no carrier
>

I had some similar problems with a now-defunct laptop with a bge(4) -
(acer F1000) - IIRC, I had to ifconfig it to '0.0.0.0 up' before 
dhclient would work - something to do with the link detection not 
working properly if an address wasn't assigned..


e.g:

# ifconfig bge0 0.0.0.0 up
# dhclient bge0

would work..

you might try this, and if it still doesn't work, verify the
card is okay via a manually assigned address..

my apologies for not submitting a bug at the time :)

Thanks,

- Chris

Mozilla suite stability on 2.4/2009Q4

[Chris Turner]

Hello all

I recently (finally!) upgraded my main dev system to local builds of
both 2.4.1 and 2009Q4 pkgsrc packages -

Since doing this, I've been getting a ton of firefox / thunderbird
crashes. I've tried disabling my plugins, etc. but this isn't
helping too much - the crashes are confined to these applications,
so I'm thinking it's something XUL specific perhaps, or perhaps
some interaction between e.g. XUL, DBus and the like,
but I'm really at a loss for what is happening..

I'm in the process of getting wip/firefox to build, etc
to test other scenarios, and I'll be testing the official
packages later, but in the meantime I thought I'd ask if anyone
had any similar experiences?

(as I'd like my own builds to work anyway :))

Thanks in advance,

- Chris

Re: Mozilla suite stability on 2.4/2009Q4

[Chris Turner]

Dang.. I am not on top of my email fer sure.

These are locally built, Q4 in a spoofed 2.4 chroot on a 2.5 host.

(e.g. cd /usr/src/stable && make install DESTDIR=/foo
  && env UNAME_r chroot $DESTDIR)

Sooo, I wouldn't take this too seriously
as I certainly haven't investigated all the possibilities on my end!

Was just kind of putting out feelers if anyone had seen anything
anecdotally before I go chasing this sucker down.

Seems to be all my GTK-or-perhaps-gnome apps -
as evince has been crashing in a similar way too.

Do I need some new GNOME infrastructure stuff of late.. -
I have no idea what kind of daemons gnome is actually requiring
these days.. a quick ps reveals there is no 'bonobo-activation-server'.. 
but I still launch a 'gnome-settings-daemon' in my .Xinitrc, and dbus is

running just fine.. hmm.

(use a windowmaker + nautilus kludge-top)

though that being said I did muck around with fonts lately..

hmm.

Pointers welcome! and thanks much.

- Chris

Matthew Dillon wrote:

:Hello all
:
:I recently (finally!) upgraded my main dev system to local builds of
:both 2.4.1 and 2009Q4 pkgsrc packages -
:
:Since doing this, I've been getting a ton of firefox / thunderbird
:crashes. I've tried disabling my plugins, etc. but this isn't
:helping too much - the crashes are confined to these applications,
:so I'm thinking it's something XUL specific perhaps, or perhaps
:some interaction between e.g. XUL, DBus and the like,
:but I'm really at a loss for what is happening..
:
:I'm in the process of getting wip/firefox to build, etc
:to test other scenarios, and I'll be testing the official
:packages later, but in the meantime I thought I'd ask if anyone
:had any similar experiences?
:
:(as I'd like my own builds to work anyway :))
:
:Thanks in advance,
:
:- Chris

Are you using the binary packages from Avalon or did you build
firefox from /usr/pkgsrc from scratch?

I'm thinking we may have introduced some structural incompatibility
between 2.4 and 2.5 that is causing the Q4 binaries built on 2.5
to fail on 2.4.

-Matt
	Matthew Dillon 
	

Re: DragonFly 2.4.1 Released!

[Chris Turner]

Simon 'corecode' Schubert wrote:

Erik Wikström wrote:

Of course this is more or less what you are already doing, except for a
change of names, X.Y.0 -> X.Y.0 RC, X.Y.1 -> X.Y.0. The only difference
is that it will make casual users more aware of the fact that some
problems are to be expected.


I completely agree.



finally clearing out my email from *EEK* last october. this seems 
timely.. any conclusions?

Re: Printing with lpr in OpenOffice won't work

[Chris Turner]

Stephane Russell wrote:
Is there a way to send a printing to lpd with OpenOffice on DFBSD? There 
is no lpr or lp binary available in the SUSE package, so I have to send 
the print into a file and then do a lpr myself.


never tried this myself, but you might try either:

1) softlinking /compat/linux/usr/bin/lp to /usr/bin/lp
2) writing a /compat/linux/usr/bin/lp shell script that
   calls e.g. 'exec /usr/bin/lp $*'

this assumes openoffice uses lp to start with..
which I'd think it does, but anyway..

good luck.

- Chris

Re: prompt and time zone

[Chris Turner]

Pierre Abbat wrote:
name. There's a file /etc/login.conf which looks like it may be the right 
place to set this; 


login.conf for all users irrespective of shell,
use the shell-specific profile-alike for specific shells

(as /etc/profile mentioned before - but pointing out that bash is more a 
sh clone, than the other way around - though history is more complicated

than that - see sh(1)..)

also, adduer.conf has some relevance for adding new users..

Re: Monitoring CPU time

[Chris Turner]

Francois Tigeot wrote:


The FreeBSD plugin uses the sysctl kern.cp_time, which is not present on
DragonFly. Is there any way to easily get the different percentages of nice,
idle, user, system, etc... time from a shell ?

there might be a better way to do this - but you can snag that 
particular set of data from vmstat -c 1 ..

Re: Monitoring CPU time

[Chris Turner]

Francois Tigeot wrote:


Unfortunately, there's no data for nice or interrupt. Time to dig into top
sources...


interrupts are there - not by device, etc.. but anyhow..
yeah.. systat probably has code to get at everything you want..

--
Sent by An eMail

Re: First file to look at in dragonfly src tree

[Chris Turner]

Vivek Ayer wrote:


I just checked out the dragonfly src tree from avalon and was
wondering which source file you all would recommend to look at first
to get a feel for how the kernel and system works. I supposed wherever
'int main' is, right?



Agree with Sascha r.e. finding some place of interest.. I'd say:

I suppose this depends on what you are trying to learn first -

if you are wondering about generics of how the kernel runs processes, 
threads, handles system calls, etc, Personally, I'd take a look at the 
VKERNEL code -


/usr/src/sys/platform/vkernel/

As it's fairly simple to understand the bootstrap, without worrying 
about so much of the hardware-specific booting (such as bus device 
enumeration, etc)


If you're trying to write a new device driver, this might not be so 
helpful - in which case I'd check out the appropriate bus drivers first 
(e.g. pci, etc)


And do yourself a favor and fetch a copy of cscope. Just trust me on 
that one :)


Good luck!

Re: upgrade packs

[Chris Turner]

Gergo Szakal wrote:

Opinions?


+1 ASS KICKING

think this is great as it encourages the end user to 'build' their
system - or at least get used to using the source tree / make to
do updates..

just need to ensure that the make.conf is stock..

Re: Ideas and questions on pkgsrc

[Chris Turner]

Justin C. Sherrill wrote:

 - General ideas about the bulk builds and binary installs; I've been
staring at it so long I can't see the forest because there's all these
trees in the way.


Yeah.. Lots of these ideas crossing my mind today as well -
this DNS-in-base thing kind of spurs a lot of thinking about direction

(or maybe thats my life right now :)

aaanyhow..

some things come to mind:

1) I've got some 'pkgsrc profile' scripts lying around. They are
  basically lists of package names, which use 'rcorder' to create /
  install a heirarchy of packages.. e.g. just to pick a random example,
  your 'trac' profile might pull in your 'apache' profile,
  and also your 'python' and 'developer tools' profile.

  I'm in the process of dusting off code / revamping my website -
  I'll try to get that released somehow - maybe getting it
  into pkgsrc/pkgutils might make sense -

  and providing a few stock ones of these might help..

  which brings to mind:

2) nrelease vs pkgsrc vs contrib interfacing -

  If the plan is to modularize alot of contrib, it probably makes sense
  to get some kind of toolkit in place for this. I for one like having
  some things 'always built', and running nrelease every now and again
  has always seemed a bit like something that I do as an afterthought -
  and having this be a more 'usual' thing, might make the release
  go more smoothly / etc.

  e.g. if my 'make buildworld' would build dfly world, a pkgsrc
  bootstrap, and some core tools, like say BIND :) - that would be
  pretty hip - and if 'make installworld' would drop those into place
  without blowing up my base install, that would be hip too.

  or maybe having a separate target, like 'make buildpkgs' in the master
  makefile - in order to preserve modularity, I dunno..

  Side point:  Does the installer work in a vkernel, btw? I've never
  tried - but this would probably help with release engineering /
  end user documentation testing..

3) hack-a-thons - on line to start.

  I think these would be a great idea. I for one am always scratching my
  head to catch up with the list, etc.. Perhaps I should be on leaf /
  IRC more too. my fault there.. but basically, there's lots of stupid
  little things that I never get around to doing, coz there's always
  some other thing. But yeah.

anyhow.. I spend far too much time writing emails to the list and not 
writing code!


so, with that in mind - I'm off to catch up on git finally, so I can
commit my 1 liner from like a frigging month ago. w00t!

Re: HEADS UP: BIND Removal. Short instructions for migration to pkgsrc-BIND

[Chris Turner]

Jan Lentfer wrote:
As already announced I will push in my BIND removal patch-set to master 
in the next few days. So anyone running a base-BIND on their system and 
upgrading their world after I pushed that in will end up with no named 
binary around anymore.


Hello -

I tried to checkout your repository to take a look at this at:

http://gitweb.dragonflybsd.org/~lentferj/dragonfly.git

and was getting some errors. I'm pretty new to git,
so it's 100% possible I'm screwing something up..

problems were:

# git clone http://gitweb.dragonflybsd.org/~lentferj/dragonfly.git
Initialized empty Git repository in /usr/data/srcs/tmp2/dragonfly/.git/
warning: remote HEAD refers to nonexistent ref, unable to checkout.

I then tried to copy in my 'git add-remote' and 'git ls-remote' to find 
a branch

to 'pull' from, but this got errors too..

I totally see the repository in my web browser..

any ideas about what I'm screwing up - or is this repository
in flux?

cc'ing users@ for the hell of it..

cheers

- Chris

Re: HEADS UP: BIND Removal. Short instructions for migration to pkgsrc-BIND

[Chris Turner]

Chris Turner wrote:

http://gitweb.dragonflybsd.org/~lentferj/dragonfly.git


was pointed out that I was using http url.. DOH

Re: Ideas and questions on pkgsrc

[Chris Turner]

Aggelos Economopoulos wrote:


Well, w/o having seen the code, this sounds like a bit of a hack :) Also
I'm not sure what problem you're solving. Pkgsrc already has working
package dependencies. The serious issue is with handling upgrades.



yup. possibly so.

Problem solving: installing a box that does X, and specifying a box
  that does 'X' simply.

e.g.: my desktop is 80%

windowmaker
nautilus
firefox
aterm
emacs

and webdev is 80%

apache
postgres
p-languages

therefore 'workstation' ==

require desktop
require webdev

and I'm done - and I don't care if package 'foo' switched to 
latest-desktop-technology-blah v4.6d in the meantime, and I don't

have to track it in the makefiles.

also makes it easy to make tools that automate system builds, etc..



Eh, to do this properly you'd have to upgrade all packages depending on
your 'base' packages and all packages depending on those packages and so
on ("transitive closure").



I'd only really do this for 'core' things - like a DNS server, MTA, etc.

(those pesky types of things that keep getting moved out of base)

these tend to be highly contested leaf-nodes, which don't really
have dependancies - aka things that 'used to be' part of the OS.

I personally blow away the entire system at each release - and
don't really do much in the middle (effectively 'upgrading all 
packages') . Test box gets test packages - if I get the time to do it. 
Though I'm adressing my workflow as I get back up to speed from where I 
was ~2y ago..

Re: Ideas and questions on pkgsrc

[Chris Turner]

Aggelos Economopoulos wrote:


My by far most important gripe w/ pkgsrc is the inability to do mass
upgrades from binary packages in a straightforward manner. Not even sure
if it's anything the pkgsrc developers are concerned with.


Can't you do this if you have the right make targets (e.g. check binary 
packages first) and pkg_rolling-replace?


I think the problem you mean is more is 'mass upgrades from binary 
packages in a remote location' ..


I think yeah - a lot of pkgsrc folks have a site-local 
/usr/pkgsrc/packages hanging out in their automounter somewhere ..

Re: Ideas and questions on pkgsrc

[Chris Turner]

Aggelos Economopoulos wrote:

Heh. Those are all arbitrary choices really. Not terribly useful to a
general user audience.


s/above/gnome-and-3d-compositing-windowmanager/

and as for tracking, no - I just track the leaf nodes.
intermediate cruft stays buried in the makefile deps.


Yah, see, that's what Justin was talking about. Tell *that* to someone
who's been using apt-get/yum/urpmi/whatever and they'll just feel sorry
for you.


and I for them.

- C

Re: Ideas and questions on pkgsrc

[Chris Turner]

Justin C. Sherrill wrote:

that appears to be removable for the same reasons (overkill).  Sascha I
think was looking at a BSD-based replacement for groff, or maybe I'm
conflating it with the mandoc work.


I think someone was looking at "heirloom troff" at somepoint - but
I think that ran into incompatibilities with drift in the -man macros
since 4.4BSD, etc. anyhow - not really relavent, just for archive sake

Re: upgrading postgres

[Chris Turner]

Pierre Abbat wrote:

I meant *on DragonFly*. I don't think I have anything installed that depends 
on Postgres, but I am going to install the PHP module.


Same applies - but with the additional step of managing packages..

1) Backup & Dump databases w/old pg tools (e.g. separate copy of all db 
data, and an export - make sure you get the 'global' data - this

is in pg docs I believe)
2) Make sure you have old pg binary packages around - otherwise, 
generate some from live system w/'pkg_tarup', download, whatever

3) delete old pkgs w/pkg_delete
4) install new packages however (binary or build)
5) recreate db data dir (e.g. w/initdb)
6) import data using tools, global first, then db's (this depends on how 
you dumped them)


sorry this isn't exact steps.. but it should be good as a guide..

I hope

Re: HEADS UP: BIND Removal. Short instructions for migration to pkgsrc-BIND

[Chris Turner]

Sascha Wildner wrote:
The only question is: Do we want to _maintain_ a drill and libldns in 
our base from now on? Because that's what we would have to do if it's in 
base.


there needs to be some way to verify DNS out of the box. period. without 
pkgsrc. period. this is non optional, IMHO.


otherwise I might as well port System-III unix and start UUCPing my 
email with a 300bps acoustic coupler..


kernel!dragonfly!ucbvax anyone?

Re: HEADS UP: BIND Removal. Short instructions for migration to pkgsrc-BIND

[Chris Turner]

Freddie Cash wrote:
If you just need to verify that DNS works (resolve hostname to IP), then 
what's wrong with "ping some.host.com "?  ;)  
Kills two birds with one stone:  DNS lookups work, and Internet 
connection is live (if using a hostname on the Internet).


ok fair enough. was hopped up on coffee. blame it on the drugs. sorry 
bout that.


:)

Re: pkgsrc DragonFly 2.7/x86_64 2010-05-07 10:42

[Chris Turner]

Justin C. Sherrill wrote:

Built on pkgsrc-2010Q1 and this is 64-bit.

How much of these is too much?  Once the quarterly release is relativly
stable in terms of fixes/updates, the builds happen rapidly.



Saw similar discussion in

Re: [issue1761] pkg sysutils/bacula : build works with 5.0.2

on bugs@ - moved here so the conversation isn't in bug report..

Justin C. Sherrill wrote:
> As a tool for making that easier, the output from each bulk build is
> available here:
>
> http://avalon.dragonflybsd.org/reports/
>
> ... though it's a lot of text to swim through.

I'm thinking if there's a way to get some kind of httpd config
to make the 'meta/report.html' the first page in each
build, it might simplify browsing - as long as the build-log-links
worked..

not sure how to do that though - running a script to generate
 rewrite_rules in .htaccess files in the top level might do the trick..

then e.g. a simple announcment, kind of like agc makes e.g. ' builds for 
this branch are available' would do..


also - as for log mailing list - maybe some kind of rss setup makes 
sense? ditto on 'not sure how to do that' - though could probably

be manually generated in a similar fashion..

on another note - maybe it makes sense to have some kind of branch in 
the git repo for 'build environment tools' (e.g. 'meata-pbulk')- I know 
I've got my own little scripts to manage my own bulks - this could e.g. 
make ppls environments more consistant and help w/pkgsrc patch flow


I'm personally trying to get more of a 'continuous build server' kind of 
setup going so I'm not constantly trying to fix the same 10 packages

at every release.. which theoretically would free up my time to actually
get those 10 packages fixed in-tree :)

anyhow.. some idears. perhaps these should be streamlined via discusison
and moved to the wiki?

FF 3.6 CPU usage on 2.6?

[Chris Turner]

Anyone seeing any issues with Firefox running away on CPU on 2.6?

I had a problem with my 2.4 builds that was mitigated
buy a clean build - but now it's appeared again on my 2.6 builds -
trying to isolate this.

Thinking it might have something to do with my gnome settings, but
I dunno.

Cheers.

Thanks

Re: FF 3.6 CPU usage on 2.6?

[Chris Turner]

Chris Turner wrote:

Thinking it might have something to do with my gnome settings, but
I dunno.


This is 3.6 from the 2010Q1 branch, btw. Strangely, bug seems to be 
acutely triggered by visiting slashdot, but not say, the DF site. I 
guess /. doesn't just waste head-cycles :). Hmm. ok. shutting up now.

Re: FF 3.6 CPU usage on 2.6?

[Chris Turner]

Chris Turner wrote:

Thinking it might have something to do with my gnome settings, but
I dunno.


sorry for chatter - was a few bugfixes behind on 3.6 or something -
a fresh installworld seems to have resolved it.. or at least a quick
few tabs in a few sites seems to suggest that.

Re: starting Apache

[Chris Turner]

Justin C. Sherrill wrote:

It seems that maybe we should ship with /etc/rc.d set by default.


-1 for me - usual 'system is system and pkgsrc is pkgsrc' reasoning.

(and just had my tweaked apache start script botched by pkgsrc)

pkgsrc / rcng are super flexible - isn't there some way to spec the
rcorder scan dirs? (seem to recall.. maybe freebsd does this?) -
perhaps setting the rc default to scan pkgsrc as well might be the way 
to go if a change is required?


then we could throw in some quasi-rh-like 'service foo start' command
which would locate the appropriate rc script (according to dir scan 
config), and run - this would let the newer-folks have an easy setup:


- tweak rc.conf
- service foo start

to prevent the 'hey whered my script go', and
and people with custom scripting wouldn't be clobbered..

though I'd like something linguistically tweaked a smidge like 'rc run' 
- 'service' is WAAAY too windowsy and muddles things

Re: starting Apache

[Chris Turner]

Thomas Nikolajsen wrote:

rc variable namespace still isn't separated, e.g. no reserved name space
for base rc scripts, any ideas for this?


this could probably be tweaked somehow in the scan / startup sequence 
with rcorder(1) statements.. e.g. pass 1 for 'basic init', pass 2 for 
'secondary init' -


but then it's getting complicated..

I like the simplicity of rcng, which lets the user kind of 'make their 
own' policy, if desired..


for example, I just tweaked my own rc.conf to have a
little 'ops mode' toggle.

e.g: I set box to ops_mode='basic' for heavy maintinance,
do maintinance
tweak ops_mode='full'
and reboot

this keeps me from thinking 'oh right - that needs this, so don't enable 
that' when doing maintinance.

Re: WINE - DFBSD

[Chris Turner]

Siju George wrote:

Virtualbox installation and qemu installation from pkgsrc failed for
me :-(


Just built qemu - will submit patches / fixes in a bit - but if you're 
itching to try.. and assuming our pkgsrc dependencies are similar..


Note: I haven't actually *used* the build yet..
  so it might just crash instantly, etc.

Against 2010Q1 branch of pkgsrc, on 2.6:
(other branches / versions might work)

  - build until you get the SDL failure, then tweak the
sdl 'main.c' file to comment out the block containing 'mlockall',
resume w/'bmake package' - might need to cleanup the .o file
from the error first..
  - continue build until you get the QEMU failure about OSSVERSION -
comment it out, and set 'version = 0x03',
resume w/'bmake package'

incedentally, this makes pulseaudio build, if anyone has any broken
deps on that. (again - no idea if it works :)

2x things to fix :

1) we need mlockall()

   http://www.opengroup.org/onlinepubs/009695399/functions/mlockall.html
   for any ambitious VMem patch-making-ppl.

2) we probably should? define OSS_VERSION - or generally be more
   in step with oss.

I'll submit patches for pkgsrc / crosscheck pkgsrc current / file OS 
bugs in a bit.. time for a nap :)


cheers

- Chris

Re: WINE - DFBSD

[Chris Turner]

Antonio Huete Jimenez wrote:

We already return ENOSYS for mlockall() / munlockall() in both master
and 2.6 so pulseaudio should build fine since then.


Hmm.. I was getting undefined link-time symbols -
I wondered how it was able to even compile..

This explains the problem I suppose.

It wasn't in the 'mlock' manpage, and I didn't dig deeper.

still failed my build in any case - again, I didn't crosscheck anywhere.
I'm definately running 2.6..

FYI my build works so far w/VNC qemu display & tap networking -
setting up my router dev 'box' now..

haven't tested anything else.

cheers.

Re: running i386 vkernel on amd64

[Chris Turner]

YONETANI Tomokazu wrote:

You can install which ever applications you want, but it doesn't
mean they all work as you expected :)  But seriously, since the X server
and the clients don't need to be running on the same node, you should be
able to run any X clients (firefox, for instance) inside a vkernel by
setting DISPLAY environment variable to point to the vkernel host which
is also running X server and accepts X protocol from the vkernel.


Or if anyone wants to step up and:

1) get Xvfb working so Xvnc+Xvfb could work as a VNC display
2) Just get a vncserver building in pkgsrc

that could work too..

(thought I'd mention it)

Re: BSDStats: Status Report

[Chris Turner]

Aggelos Economopoulos wrote:
 > It's a lower bound (assumming no cheating which you can't detect 
anyway)
on the number of users you have, which might not be /completely/ 
useless. Using those numbers for userbase comparisons is silly of course :)


Hey - gotta start somewhere:

http://www.columbia.edu/acis/history/census-tabulator.html

http://2010.census.gov/espanol/:

"Para Progresar, Necesitamos Sus Respuestas"

indeed.

Re: which pkgsrc version do I get via git?

[Chris Turner]

Justin C. Sherrill wrote:

Seconded; gstreamer is one of the worst "offenders" right now in terms of
dependent package breakage.  It's the same on 2.6,2.7, i386, and x86_64.


I've really gotta get in synch w/my patches (I think I tweaked a few of 
these), but anyhow - using 2010Q1 on 2.6/i386 I've been able to so far 
get the following:


# pkg_info |grep gst |cut -d ' ' -f 1
gstreamer0.10-0.10.28
gst-plugins0.10-base-0.10.28
gst-plugins0.10-theora-0.10.28
gst-plugins0.10-ogg-0.10.28
gst-plugins0.10-vorbis-0.10.28
gst-plugins0.10-good-0.10.21
gst-plugins0.10-bad-0.10.18
gst-plugins0.10-flac-0.10.21
gst-plugins0.10-gio-0.10.28
gst-plugins0.10-soup-0.10.21
gst-plugins0.10-oss-0.10.21
gst-plugins0.10-x11-0.10.28
py26-gstreamer0.10-0.10.18

If I recall correctly, py-gstreamer needed the following:

PTHREAD_AUTO_VARS=  yes
PTHREAD_OPTS+=  require
.include "../../mk/pthread.buildlink3.mk"
.include "../../mk/bsd.pkg.mk"

will submit PR eventually..