Hello,
Here is a log of a discussion held with vorlon and Zomb about this
issue, just so that this knowledge doesn't dissapears again as my
original followup apparently did.
10:58 < vorlon> svenl: why exactly did you mark 242068 as "grave"?
11:08 < svenl> vorlon: let me check.
11:10 < svenl> vorlon: mmm, maybe i would not have.
11:10 < svenl> vorlon: it makes the package unusable on powerpc.
11:11 < svenl> vorlon: since the new airport is unsupported (being a broadcom
chip), this mean a major chunk of wifi chips is
not supported.
11:12 < svenl> vorlon: the bug has been open 286 days, is reported to be
functional on powerpc (in the 2.6.6 days). Not sure
what the situation is on other arches, but the maintainer don't
seem to do his job seriously about this one.
11:12 < svenl> vorlon: so i would either remove the package or fix this.
11:13 < svenl> vorlon: this could be a primary installation method for d-i in
addition.
11:14 < svenl> vorlon: notice that #286305 is 27 days old and also grave, and a
solution to it would fix the other one too.
And there is #290047 too.
11:14 < svenl> vorlon: what is the MIA status of Bradley Bell <[EMAIL
PROTECTED]> ?
12:51 < svenl> vorlon: do you think 242068 should be downgraded ?
12:52 < vorlon> svenl: yes, I don't see why having a lack of binaries on an
architecture is RC. It's common enough for
module packages to only provide i386 binaries.
12:53 < svenl> vorlon: when there is only x86 binaries.
12:53 < svenl> vorlon: and i tried to build the stuff for myself, and failed.
12:53 < svenl> vorlon: so there is more than just that broken.
12:53 < svenl> vorlon: and there is also the bug about x86 2.4.27 stuff.
12:53 < svenl> vorlon: and there is a report of them working on powerpc.
12:54 < vorlon> ok, well you didn't *say* any of that when you raised the bug
severity. You raised the severity on a bug
report asking for *powerpc modules in the archive*.
12:54 < svenl> vorlon: yep, but i told it to you a bit above.
12:54 < svenl> you didn't seem to react on it.
12:55 < svenl> vorlon: that package seem utterly abandoned by its maintainer
anyway. 3 RC bugs it has right now.
12:55 < svenl> vorlon: i would just quick it out of sarge for now, more secure
like that.
12:56 < vorlon> the other bugs I was already planning to NMU for, because I
understand why they're RC. You didn't mention
anything about the package being unusable on powerpc when you
try to rebuild it above, and a lack of binaries
on powerpc isn't RC.
12:56 < jvw> svenl: mail [EMAIL PROTECTED] with suspecions
12:57 < svenl> vorlon: i am not entirely sure about this.
12:57 < svenl> vorlon: and i raised the RC level to catch attention, which it
did.
12:57 < vorlon> 286305 may have been open for 27 days, but it wasn't actually
fixable until 2.4.27-2 was available.
12:58 < svenl> vorlon: i don't understand why there is no more info, i though i
replied to this bug report, i wonder if it
has dissapeared or i did a mismanipulation.
12:58 < vorlon> Er... that's a crappy reason to raise the severity of a bug to
RC when it's not an RC issue.
12:58 < svenl> vorlon: the one in question is open since 248 days or something
such.
12:58 < svenl> vorlon: if i had not done it, we would not be discussing this.
12:59 < svenl> vorlon: and i disagree with you about this. Raising RC level on
certain bugs we want fixed is a good thing to
do, expecially as we will be multiplicating bug-fix-parties and
such.
12:59 < vorlon> what failed when you tried to build the powerpc binaries for
yourself?
12:59 < svenl> vorlon: but i know not everyone agrees with me because of that.
13:00 < svenl> vorlon: i don't know even how i should do it, it seem to me that
a -source package was missing or something, i
was fully unable to understand how even the x86 modules where
built.
13:00 < svenl> vorlon: i mailed doko, but he also didn't remember how he
managed to build them.
13:00 < vorlon> The release managers *and* the BTS admins disagree with you
about this. You should not be raising the
severity just because it's a pet bug of yours -- the RC
severities are clearly defined.
13:01 < svenl> vorlon: sure, so you can downgrade them once you notice.
13:02 < svenl> vorlon: the package is fully unusable for me though. I mean, i
do package a kernel module and am in the kernel
maintainer team, and was not able to find the stuff to build it
from, and the maintainer seems MIA, or at
st not caring about the package, do we really want to ship a package like that ?
13:02 < vorlon> What the hell good did it do to have me notice it? If it
wasn't RC (and you should have enough experience
with the BTS to figure this out for yourself), all I was going
to do with it when I saw it was downgrade it.
If the package couldn't build from source, someone NMUing it
would have noticed this anyway without you
raising the severity of an unrelated bug.
13:02 < svenl> vorlon: what do you want to do about this ? just add a followup
which will be mostly ignored ?
13:03 < svenl> vorlon: i said it, i thought i added more mail, i don't know
where it went. it was some weeks ago, before
christmas even i think.
13:03 < svenl> vorlon: and if it is x86 only, please mark it such in the
control file, but i got RC bugs against packages for
exactly that, so ...
13:06 < svenl> vorlon: and this driver is essential for wifi support on recent
apple notebooks, and if we don't fix it, we
will see a flurry of bogus "howto compile wifi drivers on my
ibook G4" documents out there, which are more
harm than anything else.
13:06 < svenl> vorlon: i have just apt-get sourced the package, and the control
file has :
13:07 < svenl> Package: linux-wlan-ng
13:07 < svenl> Architecture: i386 powerpc arm alpha hppa
13:07 < svenl> Description: utilities for wireless prism2 cards
13:07 < svenl> Package: linux-wlan-ng-doc
13:07 < svenl> Where is the source of the linux-wlan-ng-0.2.1-modules package ?
13:07 < svenl> where is the -source package ?
13:07 < Zomb> svenl: it was a proof-of-concept package for apt-src, now joeyh
dropped it
13:08 < svenl> Zomb: so what do i build the modules from ?
13:08 < vorlon> well, I'm looking at the package now and agree that we
shouldn't have modules binary packages in the archive
from this, because there's no way to build them from source.
13:08 < svenl> vorlon: see, this was the reason why i raised the RC level, i
wonder where my explanation did go.
13:08 < Zomb> svenl: you get the source package and extract it manually. I did
not look how to fix the download method for
module-assistant yet.
13:09 < vorlon> I wonder that, too. :)
13:09 < vorlon> svenl: as for wifi support for apple notebooks, the *recent*
ones all have Airport Extremes, which is not
supported under Linux at all unless something has changed in
the two weeks since I last looked.
13:10 < svenl> it was just for christmas.
13:10 < vorlon> Hrm? Someone has the AE working under Linux now?
13:10 < Zomb> vorlon: what is the key to get support? Nothing like ndiswrapper
for OS-X drivers?
13:11 < svenl> Zomb, vorlon: in any case, such a package is not fit to be
distributed by debian, at least i would have
13:11 < svenl> Zomb, vorlon: in any case, such a package is not fit to be
distributed by debian, at least i would have
nothing to do with such a thing, and would prefer it dropped
than shipped in the current state.
13:11 < svenl> vorlon: not that i know.
13:11 < vorlon> Zomb: hah, you think apple is going to provide a standard
driver interface for OSX that we could wrap around?
They couldn't even be bothered to use a standard *hardware*
interface for the damn thing.
13:11 < vorlon> Zomb: anyway, from what you said there is no source package
corresponding to the modules packages in the
archive, correct?
13:11 < svenl> vorlon: so the next best thing is a usb-wifi dongle, and the
wlang stuff is the best next thing.
13:12 < svenl> vorlon: i am investigatting a project to build airport-extreme
compatible boards from the rallink chip though.
13:12 < Zomb> vorlon: unless something has changed in the last monts since I
looked at it, there was no -source deb for linux
wlan ng
13:12 < svenl> So, it should be kicked out of the archive for breaking the DFSG.
13:12 < Zomb> svenl: no, the stuff can be built from source
13:12 < liw> er, no support for the builtin wifi on apple *books? that would be
inconvenient
13:12 * liw hugs his laptop, again :)
13:13 < Zomb> svenl: you just need to construct the debian/rules call to make
it create module packages
13:13 < svenl> Zomb: but the source is not in debian.
13:13 < vorlon> liw: the Apple AirPort Extreme is beyond proprietary. The
original AirPort is well supported, but is already
deprecated.
13:13 < svenl> liw: airport-extreme is broadcom, which will never be supported.
13:13 < Zomb> svenl: apt-cache showsrc linux-wlan-ng | grep gz
13:14 < svenl> liw: which is why i got a second hand ibook G3.
13:14 < svenl> Zomb: that is no excuse. stuff in main need to be buildable from
main, and even autobuildable for the most
part.
13:14 < vorlon> Zomb: it cannot be built from *this* source. It's not
supportable in its current state for security updates.
13:14 < svenl> Zomb: or i am going to insist about having ppc64 kernels in
sarge.
13:14 < Zomb> oh, I see
13:15 < svenl> Zomb: and remember, the prefered form of modification include
the debian/rules and other stuff.
13:16 < svenl> vorlon: so, you want me to fill a new bug, add some info, or can
you handle.
13:17 < svenl> vorlon: i am doing parted stuff right now, which will add RAID
and LVM support for apple boxes.
13:18 < vorlon> svenl: I think I have enough information without another bug
report at this point.
13:18 < svenl> vorlon: yeah, if you intent to fix it, if not all this part of
the log should be added to the bug report or
something ?
13:19 < vorlon> svenl: but I'm not NMUing it for this, I'm kicking it out of
testing; it's too much of a mess. If you want
this package fixed for sarge, please look into it yourself, or
find someone else with an interest in this
package.
13:20 < svenl> vorlon: do you keep a log and can add this discussion to the bug
report ?
13:20 < vorlon> svenl: a follow-up to 286305 seems more appropriate than a new
bug report, if you want to send more info.
13:20 < svenl> as i would rather not rewrite another time all this.
13:20 < svenl> Zomb, vorlon: it is ok to include your part of it ?
13:20 < vorlon> yes, you can quote me.
13:21 < Zomb> svenl: sure
13:21 < Zomb> svenl: wait a moment
13:21 < svenl> Mmm, does one of you keep logs, i don't.
13:21 < vorlon> no.
13:22 < svenl> Ok, will do some copy pasting .
13:22 < Zomb> svenl:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=184899&archive=yes
13:22 < Zomb> svenl: I have submitted a patch there to make it work as all
other module packages do
13:23 < Zomb> for what I have been flamed to death by joeyh on d-d :(
Friendly,
Sven Luther
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
