gh. How do people feel about this sort of thing?
--
Matthew Garrett | [EMAIL PROTECTED]
#x27;ll take you a while to compile
everything :)
--
Matthew Garrett | [EMAIL PROTECTED]
d then the files don't get copied into packages
correctly.
> 2) GTK
Mm. Never tried that one...
> 3) the utmp bug (blocks PAM and anything that uses it)
I've had working PAM before - what's the issue? Just the different utmpx
semantics?
--
Matthew Garrett | [EMAIL PROTECTED]
d be handy
would be a set of packages that fail to build - what would be even more
useful would be some means of working out which ones are the most
important blockers.
--
Matthew Garrett | [EMAIL PROTECTED]
in the current version of tar in sid - it doesn't honour the
--no-recursive option. Bdale's uploaded a new version that fixes it, I
guess it'll be installed tonight.
--
Matthew Garrett | [EMAIL PROTECTED]
than they've been in
> the past, though.
As far as I know, NetBSD weren't pushing their stuff upstream terribly
hard. I think that's mostly resolved now (gcc<3.2ish defaulted to a.out
on NetBSD-i386 - gcc 3.3 at least defaults to ELF sanity)
--
Matthew Garrett | [EMAIL PROTECTED]
On Mon, Jan 19, 2004 at 12:36:05PM -0600, John Goerzen wrote:
> On Mon, Jan 19, 2004 at 05:16:42PM +0000, Matthew Garrett wrote:
> > The second effort was started up by Robert Millan and uses glibc. I'm
> > not clear on how this interacts with stuff that's more tightly
x27;t - arguably, the improved
NetBSD support for Sun4c (ie, it doesn't suck) would be handy as well. I
believe the subsets of real SGI hardware supported by Linux and NetBSD
are intersecting but don't entirely overlap, so that might be worthwhile
too.
--
Matthew Garrett | [EMAIL PROTECTED]
ity. Doing it right is more important, even if that
means it takes longer.
--
Matthew Garrett | [EMAIL PROTECTED]
ncourage a
decent working relationship.
> And that doesn't apply to the other port? Well, I think we should postpone the
> discussion untill the port is ready for production use, in a pair of weeks.
Nothing is ready for production use with that little testing.
--
Matthew Garrett | [EMAIL PROTECTED]
patching required to make our sysvinit work under
*BSD is pretty trivial.
--
Matthew Garrett | [EMAIL PROTECTED]
wouldn't be
Linux-centric - however, it is, and the fact that the Linux ports are
the only ones in an even vaguely releasable state gives them a
significant degree of priority.
--
Matthew Garrett | [EMAIL PROTECTED]
arc deb just has
000ae288 51703 R_SPARC_3200090644 __gxx_personality_v0 + 0
which is presumably roughly what my copy should look like. I'm guessing
that my toolchain is at least semi-broken - what's the best way of
finding out what's generating these UA32 entries and fixing it?
--
Matthew Garrett | [EMAIL PROTECTED]
ECTED]:/autobuilder/build/alpha/OBJ/autobuilder/build/src/sys/arch/alpha/compile/GENERIC
alpha
bash-2.05$ apt-get moo
(__)
(oo)
/--\/
/ |||
* /\---/\
~~ ~~
"Have you mooed today?"...
bash-2.05$
--
Matthew Garrett | [EMAIL PROTECTED]
(__)
(oo)
/--\/
/ |||
* /\---/\
~~ ~~
"Have you mooed today?"...
bash-2.05$
dpkg and apt for NetBSD-Sparc32 exist. I'll get the rest of a base
system thrown together as soon as possible and put up a chroot tarball
for people to play with.
--
Matthew G
Debian system to work on it -
simply working on it under NetBSD is sufficient.
--
Matthew Garrett | [EMAIL PROTECTED]
so on). sysctl -w netbsd.vendor=Debian, for example.
--
Matthew Garrett | [EMAIL PROTECTED]
-nc
-d. Stuff should build. Hack debian/rules to deal with the absence of
doc. You should now have some debs. Create the files dpkg bitches about,
and install them. You're now able to set about building more debs.
--
Matthew Garrett | [EMAIL PROTECTED]
x, but if someone wants
to take a look at that it'd be a sensible thing to package.
--
Matthew Garrett | [EMAIL PROTECTED]
compile
on non-i386. They're mostly based heavily on the NetBSD source, so can in
most cases probably be fixed just by copying stuff out of the NetBSD
source tree.
--
Matthew Garrett | [EMAIL PROTECTED]
that i'd expect a lot of it
> not to apply to debian/netbsd. only those parts in sys/, lib/libc/ and
> sundry programs ... which is probably 90% of the list anyway.
A quick grep of libc reveals that it's mostly either The NetBSD foundation
or UCB, but there's random others in th
> seem to be much activity on this channel. But maybe there is there another
> channel elsewhere i could knock at too when i am looking for help?
#debian-bsd is probably your best bet, but it does tend to be pretty
quiet.
--
Matthew Garrett | [EMAIL PROTECTED]
p this
> project (apart from using Debian/NetBSD on a regular basis?)
Yes. Read the list, try to rebuild packages, suggest ways of making things
that don't work work :)
--
Matthew Garrett | [EMAIL PROTECTED]
On Sun, Sep 08, 2002 at 08:40:49AM +1000, matthew green wrote:
>
>The only packages where this caused any great trouble were gcc and
>binutils, and that was fairly easily rectified.
>
> GDB?
Oh yes, that too.
--
Matthew Garrett | [EMAIL PROTECTED]
r case. Having a new config string means we can fix this without
breaking things elsewhere, which means there's a better chance of upstream
accepting them.
The only packages where this caused any great trouble were gcc and
binutils, and that was fairly easily rectified.
--
Matthew Garrett | [EMAIL PROTECTED]
x27;ve been meaning to get around to this for ages. I'm about to go
away for a few days (taking my laptop with me), so I'll try to write some
notes up by the time I get back on Monday.
--
Matthew Garrett | [EMAIL PROTECTED]
moving at the moment, so haven't had much time to
work on things - things should start moving again in early September, with
a bit of luck.
--
Matthew Garrett | [EMAIL PROTECTED]
On Wed, Jul 24, 2002 at 12:52:23AM +1000, matthew green wrote:
> can't we have both?
There's certainly no problem shipping both - what I was thinking about
more is which do we link everything against?
--
Matthew Garrett | [EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PR
g BSD libc - does anyone have any comments?
--
Matthew Garrett | [EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
at look like the NetBSD one instead?
--
Matthew Garrett | [EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
On Sat, Jun 15, 2002 at 05:01:10PM +0100, Matthew Garrett wrote:
> Debconf doesn't seem entirely happy at the moment
This can be rectified by setting TERM
> Large quantities of includes are currently missing, so building anything
> is likely to be a pain. Should be fixed shortly.
Joel Baker for hacking on the NetBSD stuff
in all sorts of useful ways and VMWare for providing a wonderfully useful
program for testing installation (Plex86 doesn't seem to support swapping
floppies yet, sadly).
--
Matthew Garrett | [EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
the moment. I've dealt with the
incompatible install issue by modifying stuff in /usr/share/mk. This seems
to do the job for the moment.
--
Matthew Garrett | [EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
fic patch against autotools-dev (at least, I
> think that's where config.{sub,guess} are supposed to come from, right?)
Keeping it as a Debian specific patch is easy enough, but it would be nice
if we could get it included upstream. I can't see how it could break
anything else.
nel which seems happy
to build with gcc 3.04 :) )
--
Matthew Garrett | [EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
ges seem to build happily
(I've turned off objc and java support due to not having the right garbage
collection libraries, but everything else seems fine). I haven't run the
test suite, though, so I'm not sure how useful this is.
--
Matthew Garrett | [EMAIL PROTECTED]
--
To UNSUBS
Net or
Open. If anyone has any idea how this is done, implementing it would
certainly be useful.
--
Matthew Garrett | [EMAIL PROTECTED]
he process) or leave it as is?
--
Matthew Garrett | [EMAIL PROTECTED]
I've managed to get this to approximate functionality. What are we still
missing from a base system at this point?
--
Matthew Garrett | [EMAIL PROTECTED]
On Tue, Feb 26, 2002 at 02:03:29PM +1100, matthew green wrote:
> hmm. i see this as far as real syscalls:
(snip)
Ok, in that case I think we're fine (the only ones of those wrapped by
fakeroot are the stat ones).
--
Matthew Garrett | [EMAIL PROTECTED]
t, stat, chown, chown, fchown, lchown and
possibly a couple of others that I've missed are affected. It ought to be
pretty trivial to fix them, though.
--
Matthew Garrett | [EMAIL PROTECTED]
On Sun, Feb 24, 2002 at 08:43:11PM +, Matthew Garrett wrote:
> Fakeroot on NetBSD is dying inside libfakeroot. The mess of wrap* has left
> me sufficiently confused that I'm not really sure what's going on, and
> I've certainly got no idea why it dies. Does anyo
Fakeroot on NetBSD is dying inside libfakeroot. The mess of wrap* has left
me sufficiently confused that I'm not really sure what's going on, and
I've certainly got no idea why it dies. Does anyone who understands these
things better than me want to take a look at it?
--
Matthew G
it ought to work.
--
Matthew Garrett | [EMAIL PROTECTED]
diff -urN Linux-PAM-0.72/Makefile Linux-PAM-0.72.new/Makefile
--- Linux-PAM-0.72/Makefile Sun Feb 24 15:05:03 2002
+++ Linux-PAM-0.72.new/Makefile Sun Feb 24 14:14:49 2002
@@ -126,7 +126,7 @@
# build composite defines
#
-LOADLIBES = $(PA
/run/log - symlinking the two work, and you can pass an option to
msyslog to make it produce /var/run/log instead. What's the preferable way
to do this?
--
Matthew Garrett | [EMAIL PROTECTED]
hough I'm not sure why), so that can probably be assumed to just be me
being stupid. I don't see any reason not to use pmake instead (other than
it being a bit out of date).
--
Matthew Garrett | [EMAIL PROTECTED]
Patches to shadow enclosed. These are for the build tree rather than the
package, and break compatibility with Linux. Make sure that you compile
without PAM enabled.
--
Matthew Garrett | [EMAIL PROTECTED]
diff -u -N /tmp/shadow-2902/build-tree/shadow-2902/lib/Makefile.in
./Makefile.in
On Wed, Feb 06, 2002 at 12:14:30PM +0100, smoerk wrote:
> On Sun, 3 Feb 2002 23:02:22 +0000, Matthew Garrett wrote:
>
> >I now have a self-hosting Debian NetBSD system.
>
> Great! :) What is easiest way to install Debian BSD/NetGNU to an empty
> harddisk? Do I have to
On Tue, Feb 05, 2002 at 04:32:08PM -0500, [EMAIL PROTECTED] wrote:
> I'll have to try it out, and see how it works. Do you have adduser working
> yet?
Yup. Now that groups work again, adduser runs without any trouble. I'll
try to generate a new diff and send it on.
--
Matthew
really is a lot of extra work or stuff is tied to the
kernel (ifconfig, mount and so on), so I'm inclined to modify the debian
one rather than packaging the BSD ones.
--
Matthew Garrett | [EMAIL PROTECTED]
t can probably be added in later)
--
Matthew Garrett | [EMAIL PROTECTED]
On Sun, Feb 03, 2002 at 06:53:28PM -0500, [EMAIL PROTECTED] wrote:
> Excellent. Mind sending me a diff against my patch, so I can merge it
> together,
> and send upstream?
Modified now that Matthew's pointed out that RB_POWERDOWN exists.
--
Matthew Garrett | [EMAIL PROTECT
ELOAD=$LIB sh
-c "$*"
make: *** [4] Error 1
make: Leaving directory `/tmp/fakeroot-0.4.5/test'
and running it gives me a shell but claims that I'm still me. I'll look
into this.
Although it's less useful without fakeroot running, I can provide access
to this system now for anybody who's interested in developing on it.
--
Matthew Garrett | [EMAIL PROTECTED]
ill
come up with sysvinit and give me a shell. I'm very impressed.
--
Matthew Garrett | [EMAIL PROTECTED]
eed adding/updating?
--
Matthew Garrett | [EMAIL PROTECTED]
ight just be that the libstdc++ with 3.0 is better. I don't really know,
> but it helped a lot.
I've altered the patch so that it doesn't do any messing around with
_POSIX_C_SOURCE - new patch attached. Still needs environment.mak hacked,
I'll do that properly later on.
--
M
On Sun, Feb 03, 2002 at 04:01:32AM +, Matthew Garrett wrote:
> Oh, no, got it. Unless a certain set of magic words appears in the host
> architecture, it builds with static libraries. buildlib/environment.mak.in
> probably needs patching to be a touch saner.
Yup, that does it
On Sun, Feb 03, 2002 at 03:55:32AM +, Matthew Garrett wrote:
> The COMPILING doc actually implies that the shared library problem and the
> "Unable to determine a suitable system type" may be linked. Again, this is
> probably a post-sleep problem.
Oh, no, got it. Unle
On Sun, Feb 03, 2002 at 03:53:29AM +, Matthew Garrett wrote:
> apt now builds - diff attached. The package still won't build due to lack
> of doc generation (missing debian-sgml), and for some reason that I
> haven't been able to work out the shared libraries for libapt-pkg
le to determine a suitable
system type". I'll have a look at these problems once I've had some sleep.
--
Matthew Garrett | [EMAIL PROTECTED]
diff -u -r ftparchive/apt-ftparchive.cc
../apt-0.5.4.work/ftparchive/apt-ftparchive.cc
--- ftparchive/apt-ftparchive.ccTue Jun 26 02:5
ies on "GNU libc or libiberty".
Would it make more sense to include it in the libc package? That way we
end up with a single package that approximates the functionality of glibc,
rather than requiring several of them.
--
Matthew Garrett | [EMAIL PROTECTED]
in unstable, rather than the CVS. I haven't
looked at the CVS version.
> should we try merging our patches before feeding them to
> upstream?
Sure, I don't see why not.
--
Matthew Garrett | [EMAIL PROTECTED]
e tarball...
Patch attached.
--
Matthew Garrett | [EMAIL PROTECTED]
diff -u -N -r ./archtable /source/stage2/done/dpkg-1.9.18.work/archtable
--- ./archtable Sun May 13 21:50:59 2001
+++ /source/stage2/done/dpkg-1.9.18.work/archtable Mon Jan 7 16:16:54 2002
@@ -51,3 +51,5 @@
s390-ibm
On Wed, Jan 30, 2002 at 04:06:03PM -0500, Jimmy Kaplowitz wrote:
> Well, I didn't really want to take control of that away from Matthew
> Garrett.
> Of course, he could always email me a snapshot and I would put it on the
> site,
> if he wants. If that is OK with him, I will g
s/FreeBSD/NetBSD/g on it seems to work wonderfully. Installing the package
updated my master.passwd and then gave a "postinst failed" error.
Reinstalling it worked. Thanks!
--
Matthew Garrett | [EMAIL PROTECTED]
One line addition to the OpenSSL Configure which seems to get it to build:
"debian-netbsd-i386","gcc:-DL_ENDIAN -DTERMIOS -O3 -fomit-frame-pointer
-Wall::-D_REENTRANT::BN_LLONG ${x86_gcc_des}
${x86_gcc_opts}::dlfcn:bsd-gcc-shared:-fPIC",
--
Matthew Garrett | [EMAIL PROTECTED]
lable at
http://www-jcsu.jesus.cam.ac.uk/~mjg59/debian-netbsd/source - I'll add the
source packages to that, since it sounds like people would find them
useful.
--
Matthew Garrett | [EMAIL PROTECTED]
On Tue, Jan 22, 2002 at 08:37:43AM -0600, Colin Watson wrote:
> I'm working on an experimental upload of groff which should make life a
> bit easier there; it should be ready in the next couple of days for
> somebody to try to build. Does NetBSD have ?
Yes.
--
Matthew Ga
if so, but I would
> sort of like to know...)
If it's stuff like libc and libutil, that'll be because I neglected to put
any shlibdeps stuff in the libc package. This probably needs to be sorted
at some point (and it needs to be split into a -dev too, for that matter).
--
Matthew Garrett | [EMAIL PROTECTED]
On Mon, Jan 21, 2002 at 01:50:19AM +, Matthew Garrett wrote:
> Should now be appearing at
> http://www.srcf.ucam.org/~mjg59/debian-netbsd/source
Make that http://www-jcsu.jesus.cam.ac.uk/~mjg59/debian-netbsd/source - my
quota appears to be smaller than I'd remembered.
--
Matt
Should now be appearing at
http://www.srcf.ucam.org/~mjg59/debian-netbsd/source
--
Matthew Garrett | [EMAIL PROTECTED]
I'd recommend uploading
> them if you can even if you can also produce a proper source archive.
I'll upload the tarred build trees to
http://www.srcf.ucam.org/~mjg59/debian-netbsd/source once they've finished
compressing. It's likely to take a while for them all to app
uild the deb of that might be worth it.
--
Matthew Garrett | [EMAIL PROTECTED]
blocking factor on this has been bandwidth. I'll try go get them
tarred up tonight and then set them uploading before I go to bed in the
hope that my housemates won't complain about their inability to do
anything with the network :)
--
Matthew Garrett | [EMAIL PROTECTED]
Difficult to say :) Way below 1, but some bits are solid and some bits are
completely missing.
> - how can i help?
There's a todo list on the website - I'd say that those are the most
pressing tasks. Testing is also useful.
--
Matthew Garrett | [EMAIL PROTECTED]
non-US main contrib
non-free
I haven't tested with binaries - apt-get source foo seems to work here.
--
Matthew Garrett | [EMAIL PROTECTED]
On Thu, Jan 17, 2002 at 04:27:33AM +, Matthew Garrett wrote:
> Ok, it turns out that this doesn't look too bad. In shadow-foo/lib we have
> a stack of code that performs the actual manipulation of the files, with
> all the utilities using that. There's code there for ma
en playing with PAM today without complete success - I'll try again
with 0.75 once it appears. It's certainly not vital for the moment.
Thanks,
--
Matthew Garrett | [EMAIL PROTECTED]
o idea - I'm unable to reproduce
it here.
--
Matthew Garrett | [EMAIL PROTECTED]
h is fairly short. The main problem is that passwd itself has defined
pw_lock, with entertaining consequences :( Fixing it shouldn't be too
difficult, and once pwio is modified/added to things should just work
(tm) :)
--
Matthew Garrett | [EMAIL PROTECTED]
ion of GNU code are we replacing with BSD code and
does that reduce the amount of GNU code present to the point that it
should be considered a major constituent rather than the major
constituent?
(ignore the kernel here, obviously)
--
Matthew Garrett | [EMAIL PROTECTED]
7;d be too
difficult.
> That wouldn't be compatible with NetBSD passwd databases, though. But then,
> rewriting update-passwd didn't seem like fun to me. I couldn't even get it to
> compile on FreeBSD.
In a way, it ought to be easier - all the shadow functions can just be
hack
that are offered over traditional password files, but we keep the ability
to use PAM. Combining both sets of features into one set of tools looks
like a little more work :)
--
Matthew Garrett | [EMAIL PROTECTED]
le ifconfig than to try to make some replacement that's compatible.
> Besides, Linux already has both ifconfig/route and "ip", with totally
> different
> semantics.
Ah, that's true. Yup, the init scripts all use ifup now, so just having
our own version of that ought to deal with the problem nicely.
Thanks,
--
Matthew Garrett | [EMAIL PROTECTED]
such useful advice so far. Comments?
--
Matthew Garrett | [EMAIL PROTECTED]
ke-bsd; DESTDIR=$(currdir)/debian/tmp make-bsd install seems to
do the job adequately.
--
Matthew Garrett | [EMAIL PROTECTED]
On Wed, Jan 16, 2002 at 08:48:55PM +, Matthew Garrett wrote:
> On Wed, Jan 16, 2002 at 09:05:43PM +0100, Fabbione wrote:
> > I get exactly the same error with your source code.
> > and looking at gcc I noticed:
>
> Yeah, you're right - it's not building her
debian with whatever directory your debian system is
in). Check the manpage for pwd_mkdb for further information :)
--
Matthew Garrett | [EMAIL PROTECTED]
t
wasn't one of the files I patched. I've stuck my source at
http://www.srcf.ucam.org/~mjg59/debian-netbsd/apt/ if you want to give
that a go. I'll try rebuilding it myself now.
> PS last question: does dpkg-buildpackage works
Yup :)
--
Matthew Garrett | [EMAIL PROTECTED]
d try to work out what it's doing when it
segfaults?
Thanks,
--
Matthew Garrett | [EMAIL PROTECTED]
to either copy the webpage across or give me access so that I can
upload it myself?
(oh, and we're in DWN tomorrow. Smile :) )
--
Matthew Garrett | [EMAIL PROTECTED]
ing utilities.
ii slang1 1.3.9-1The S-Lang programming library - runtime ver
ii slang1-dev 1.3.9-1The S-Lang programming library, development
ii tar1.13.17-2 GNU tar
ii tcpd 7.6-4 Wietse Venema's TCP wrapper utilities
ii
x27;ll disable that for the moment and come back to it later on)
--
Matthew Garrett | [EMAIL PROTECTED]
Is the 2.95 in the ports tree suitable for all architectures?
--
Matthew Garrett | [EMAIL PROTECTED]
ironment for further building. The machine the files were previously
available on has been stuck behind a firewall I have no control over, so
once I have more stuff built I'll move it to an accessible system.
--
Matthew Garrett | [EMAIL PROTECTED]
any of these has gcc
> or at least can give me a hint how to compile a 2.95.2 gcc on NetBSD,
> I'd be very thankful.
You ought to be able to get 2.95 to compile using the patches that come
with the NetBSD source.
--
Matthew Garrett | [EMAIL PROTECTED]
st of the init system can work in pretty much any way
we want.
--
Matthew Garrett | [EMAIL PROTECTED]
stent base package before sticking
another root tarball and package repository up (and I'd like the
opportunity to tidy up some of the packages I've got here :) ), but I'm
happy to stick them somewhere if people want to work on them.
--
Matthew Garrett | [EMAIL PROTECTED]
on both "sides" who see the idea as some sort of personal insult
designed to weaken either Debian or NetBSD. However, I also imagine that
there will be people who are more concerned about the quality of the code
than petty squabbling. I expect we'll see how it balances out before too
long :)
> However, the benefits might be very large for all concerned. It might
> be worth discussing.
Agreed.
--
Matthew Garrett | [EMAIL PROTECTED]
On Sat, Jul 21, 2001 at 12:51:17AM +0100, Matthew Garrett wrote:
> Great. I'll prod this at some point over the weekend and try to get NetBSD
> ones built. Have you done anything about differences between the FreeBSD
> and Debian file layouts?
I've built binutils packages,
1 - 100 of 115 matches
Mail list logo
