Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Peter S Galbraith]

Junichi Uekawa <[EMAIL PROTECTED]> wrote:

> > > * New upstream release \1
> > >   * fixed "BTS summary line of #1" Closes: #1
> > >   * fixed "BTS summary line of #2" Closes: #2
> > >   * fixed "BTS summary line of #3" Closes: #3
> > > 
> > > in changelogs would probably go a lot further to correcting this
> > > very minor issue than reopening dozens of bug reports that belong
> > > closed, annoying users with BTS garbage, and repeating the same
> > > thread on debian-devel over and over.
> > 
> > Yes, that sounds pretty intesresting; an addition to debian-changelog-mode
> > would be interesting.
> > 
> > It should be possible to get bugs.debian.org/ and parse the resulting 
> > HTML for the title and the original submitter, and placing it.
> > 
> > Some addition to debian-changelog-close-bug, possibly optional, 
> > that utilizes debian-bug-* (there's no debian-bug-to-buffer... hmm?)
> > to parse the bugreport ?
> 
> 
> Here is what the output looks like:
> 
> ecasound2.2 (2.3.0-1) UNRELEASED; urgency=low
> 
>   * New upstream release
>   * Run autoconf-automae-libtool in build.
>   * jackd requires /proc/cpuinfo information not available on zaurus
> From: Junichi Uekawa  (closes: #207435)
> 
>  -- Junichi Uekawa <[EMAIL PROTECTED]>  Sat, 30 Aug 2003 13:16:07 +0900
> 
> and here is the patch against debian-bug.el and debian-changelog-mode.el

Thanks for CC'ing me.  I catch up on -devel only sporadically.

I'm not sure what the goal is?
Why do you want the bug submitter named in the Closes entry?
If its the title you want, then that is already available after the bug
list has been fetched.  I've been wanting to use the title as initial
input to the close command for a wgile, but wasn't convinced it was good
enough.  I guess it's a good first stab.  That's what you guys want?

Peter

Re: debian archive disk space requirements.

[Martijn van Oosterhout]

On Sat, Aug 30, 2003 at 11:40:34AM +1000, jason andrade wrote:
> 
> Hi,
> 
> I've noticed the growth in the debian archive with some concern
> over the last few weeks/months.  We're now hitting the limit of
> the partition that debian is on and it will be quite difficult
> technically for us to deal with this in the short term.

[snip]

> Is there any way to reduce the size of the archive over the next
> 4-6 weeks ?

If you have info on what people dowload, you could delete things that people
aren't fetching. apt falls-back to other archives if the first one fails,
assuming the same package is listed in one of the pther archives in the
sources.list. I'm not sure how difficult this would be. If you also
regenerate the Package file then they will simply become invisible to users.
This could get confusing.

Hope this helps,
-- 
Martijn van Oosterhout  http://svana.org/kleptog/
> "All that is needed for the forces of evil to triumph is for enough good
> men to do nothing." - Edmond Burke
> "The penalty good people pay for not being interested in politics is to be
> governed by people worse than themselves." - Plato


pgppEwun2nGvL.pgp
Description: PGP signature

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Peter S Galbraith]

Junichi Uekawa <[EMAIL PROTECTED]> wrote:

> > I'm not sure what the goal is?
> > Why do you want the bug submitter named in the Closes entry?
> > If its the title you want, then that is already available after the bug
> > list has been fetched.  I've been wanting to use the title as initial
> > input to the close command for a wgile, but wasn't convinced it was good
> > enough.  I guess it's a good first stab.  That's what you guys want?
> 
> Erm.. although having the submitter name is probably not the point,
> I wanted to point out to you that I want the bug subject,

Right.  I understood both points.  I was wondering about having the bug
submitter there.  Maybe change the phrasing?

  * Bug fix "debian-changelog mode to support fetching of bug to fill
   in changelog", Thanks to Junichi Uekawa <[EMAIL PROTECTED]>
   (Closes: #207852).

But I've always thought listing the thitle didn't really say _what_ was
fixed and _how_.  Most times, the title mentions a symptom but not the
actual problem.  That why I never added that before.

I could make it optional...

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Junichi Uekawa]

> I'm not sure what the goal is?
> Why do you want the bug submitter named in the Closes entry?
> If its the title you want, then that is already available after the bug
> list has been fetched.  I've been wanting to use the title as initial
> input to the close command for a wgile, but wasn't convinced it was good
> enough.  I guess it's a good first stab.  That's what you guys want?

Erm.. although having the submitter name is probably not the point,
I wanted to point out to you that I want the bug subject,
and the patch I have submitted you does that (and add the submitter name,
for some kind of credibility).



regards,
junichi

Re: MEI Whitelist Autoresponse

[Adam McKenna]

On Fri, Aug 29, 2003 at 09:20:53AM +1000, Russell Coker wrote:
> The comparison to mailing list software makes no sense.

Maybe not in the context of viruses, but for the "Joe Job" problem it does.

Viruses can and should be filtered out before they reach the C-R system.

--Adam
-- 
Adam McKenna  <[EMAIL PROTECTED]>  <[EMAIL PROTECTED]>

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Brian May]

On Thu, Aug 28, 2003 at 10:05:21AM -0500, Adam Heath wrote:
> A proper entry is as follows:
> 
> * New upstream release.
>   * no longer does foo when bar happens. Closes: #12345
>   * wrapper script rewritten to not use $$ in tempfile names.  Closes: #12345
> 
> Please, everyone remember, a changelog documents *changes*.  It's not a tool
> to close bugs automatically.

Maybe a tool that integrates debian BTS/changelog behaviour could help.

eg. You run program and give name of package.

You see a list of bugs associated with the source packge.

You select one of the bugs.

It comes up with a screen with two windows:

1. One where the can review the bug history.

2. You type in extra information with what has changed,
and an entry is automatically added to the changlog.

You then repeat the process for the next bug report.

This is still very rough, but I think it gives the general idea.

Such a program could also make it easier to do common BTS tasks,
eg. merge duplicate bug reports, etc.

If we could make it easier and quicker to write correct changelogs,
I think they are more likely to occur.


(Assuming such a tool doesn't already exist...)
-- 
Brian May <[EMAIL PROTECTED]>

Re: .iso conflict, discussion of resolution

[Chris Cheney]

Stephan,

Would it be possible to get the two desktop filesi mentioned below merged
into kdelibs so that arson and k3b are easily installable at the same time? 
I can do the commit myself if you approve.

Thanks,

Chris Cheney
Debian KDE Maintainer

PS - Christian/Jean-Michel there is a new debian-qt-kde developer mailing
 list for Qt/KDE debian maintainers, debian-kde is still available
 for users.


On Fri, Aug 29, 2003 at 10:34:39PM -0700, Mike Markley wrote:
> All,
> 
> A quick summary of this bug:
> Arson, a KDE CD burning application, includes two .desktop files to
> associate certain files with it:
> /usr/share/mimelnk/application/x-iso.desktop
> /usr/share/mimelnk/application/x-cue.desktop
> 
> For more info on what's been suggested and what's been discussed, see
> 195214, 195218, and 203954.


pgp2OTQn6ZCJZ.pgp
Description: PGP signature

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Andreas Metzler]

Peter S Galbraith <[EMAIL PROTECTED]> wrote:
[...]
> Right.  I understood both points.  I was wondering about having the bug
> submitter there.  Maybe change the phrasing?

I usually don't list him/her, my changelogs are too long already. I do
list submitters who send the report by private mail instead of via
BTS, because othervise they wouldn't have any public record of their
help.

>  * Bug fix "debian-changelog mode to support fetching of bug to fill
>   in changelog", Thanks to Junichi Uekawa <[EMAIL PROTECTED]>
>   (Closes: #207852).

> But I've always thought listing the thitle didn't really say _what_ was
> fixed and _how_.  Most times, the title mentions a symptom but not the
> actual problem.  That why I never added that before.

The symptom often is much more interesting than the cause for the
audience, i.e. "blah does not crash immediately on startup on PowerPC"
instead of "Properly initialize char *foo in src/bar.c".
cu andreas

Re: configure web proxy via DHCP server

[Russell Coker]

On Sat, 30 Aug 2003 12:57, H. S. Teoh wrote:
> > It would be really "cool"(tm) if I didn't have to reconfigure every
> > program on my laptop to use a different proxy server every time I plug
> > it into a different network.
>
> Run a local proxy that forwards connections to the (external) proxy of
> your choice, and point all applications at it?

Then of course there's the issue of how to configure the local proxy.

For the local proxy on my laptop I use m4 to change it's configuration from a 
script which knows about the networks I regularly connect to.

The ideal solution however would be an addition to the DHCP standard for a 
field to specify this information.  My solution only works AFTER I have 
figured out the correct data for each network by some other means and 
customised my script.

-- 
http://www.coker.com.au/selinux/   My NSA Security Enhanced Linux packages
http://www.coker.com.au/bonnie++/  Bonnie++ hard drive benchmark
http://www.coker.com.au/postal/Postal SMTP/POP benchmark
http://www.coker.com.au/~russell/  My home page

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Joe Drew]

On Fri, 2003-08-29 at 18:12, Glenn Maynard wrote:
> I'm confused.  We have three cases:
> 
> 1. Close bug #12345 directly (12345-done), noting the version that fixed it.
> 2. Note in the changelog that bug #12345 is fixed; the bug receives a
> notification of the version that fixed it.
> 3. Note in the changelog that bug #12345, "ls --crash crashes", is
> fixed; the bug receives a notification of the version that fixed it.
> 
> #3 is obviously ideal, if the maintainer has time; no debate there.
> 
> However, you're saying that #1 is preferable to #2.  Why?  It seems to
> have no disadvantage to #1, with the added advantages that I can check
> which version fixed bug #12345 without hitting the network (since it's
> documented in the changelog), and saves developer time.  What am I missing?

Start with Herbert Xu's premise...
> We've gone through this many times already.  Upstream changes should
> not be documented in the Debian changelog, even if they fix bugs in
> the Debian BTS.

... and follow the bouncing ball.

-- 
Joe Drew <[EMAIL PROTECTED]> <[EMAIL PROTECTED]>

My weblog doesn't detail my personal life: http://me.woot.net

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[benfoley]

On Saturday 30 August 2003 03:47, Branden Robinson wrote:
> On Fri, Aug 29, 2003 at 06:00:51PM -0400, Glenn Maynard wrote:
> > A script to convert eg.
> >
> > * New upstream release .* (Closes: #1, #2, #3)
> >
> > to
> >
> > * New upstream release \1
> >   * fixed "BTS summary line of #1" Closes: #1
> >   * fixed "BTS summary line of #2" Closes: #2
> >   * fixed "BTS summary line of #3" Closes: #3
> >
> > in changelogs would probably go a lot further to correcting this very
> > minor issue than reopening dozens of bug reports that belong closed,
> > annoying users with BTS garbage, and repeating the same thread on
> > debian-devel over and over.
>
> One big problem with this approach is that the same maintainers who are
> too lazy to write proper entries for bug-closers in their changelog
> entries are going to be too lazy to ensure that a bug report has a
> meaningful summary in the first place.

maybe that should be a rule in policy. the issue of proper notification is 
valid, or not?  there should be a means to at least reduce redundant bug 
reports, other than some nasty response indicating that a bug has already 
been reported or attended. us plain old users play a valid part in the 
project, but it's beginning to feel as though the gap between users and 
maintainers requires a whole new dialogue. it's not unreasonable to expect 
that user bug reports deserve a coherent response, including some degree of 
clarification of the means of bug resolution.

ben

Re: Bug#207300: tmda: Challenge-response is fundamentally broken

[Russell Coker]

On Sat, 30 Aug 2003 10:42, Brian May wrote:
> On Fri, Aug 29, 2003 at 03:48:13PM +1000, Craig Sanders wrote:
> > the point that you keep on missing is that TMDA and similar programs send
> > "confirmation" emails to innocent third-parties who did *NOT* send an
> > email.
> >
> > TMDA and all C-R systems are broken-by-design, just as many stupid
> > end-user "autoresponders" and AV-scanners that send notifications back to
> > the forged sender address are broken-by-design.
>
> You saying that any SMTP MTA that sends bounces to unauthenticated
> E-Mail addresses is also broken?

Yes.

> That is the idea behind autorespoonders after all, to tell the sender
> that his mail didn't get through because it didn't meet some required
> criteria.

A SMTP 550 code can convey all the information that is needed for bounces.

> E-Mail that looks suspicious can be valid mail at times, for instance
> somebody I knew tried to send a ZIP file that happened to be executable
> via E-Mail.

If the mail server it was sent to responded with:
550 Don't want ZIP files of .exe content

Then the bounce message would have been clear and there would be no chance of 
it going to the wrong person.

If the C-R systems we are discussing would send their challenge in the 550 
SMTP code then I doubt that anyone would have any problem with them.

> The problem is that I see no easy way to fix this problem to the large
> scale required on the Internet while keeping store-and-forward "feature"
> of SMTP.

The old-style store and forward is dead.

Backup MX servers serve no useful purpose in the modern Internet, this is why 
big sites such as microsoft.com and hotmail.com don't have them.

If you have a backup MX then it should know all the acceptable email addresses 
in your domain and enforce all rules regarding acceptable content.  Then it 
can block content through SMTP 550 and 450 codes.

-- 
http://www.coker.com.au/selinux/   My NSA Security Enhanced Linux packages
http://www.coker.com.au/bonnie++/  Bonnie++ hard drive benchmark
http://www.coker.com.au/postal/Postal SMTP/POP benchmark
http://www.coker.com.au/~russell/  My home page

.iso conflict, discussion of resolution

[Mike Markley]

All,

A quick summary of this bug:
Arson, a KDE CD burning application, includes two .desktop files to
associate certain files with it:
/usr/share/mimelnk/application/x-iso.desktop
/usr/share/mimelnk/application/x-cue.desktop

For more info on what's been suggested and what's been discussed, see
195214, 195218, and 203954.

Since I've seen no action in three months on possible help from other
maintainers with fixing this SOMEHOW, I'd like some thoughts on the
two solutions that I *can* implement, and the one that I can at least
ram down others' throats:

1. Conflicts: k3b (very ugly)
2. Depends: k3b (also very ugly)
3. Depends: kde-burner-common, which contains the .iso files (which
would have to Conflicts: k3b anyway, unless Jean-Michel is willing to
cooperate, and which is basically archive bloat)

I will do #3 no later than one week from today, unless this discussion
convinces me that it's a bad idea (or the release manager/ftp-masters
object). If that happens, I will implement #1. Why? As a recently
liberated former dialup user, I appreciate how utterly ludicrous it is
to Depend: on a 2.2mb package for 1107 bytes of freaking text.

Please Cc: me on replies, as I'm not subscribed these days.

-- 
Mike Markley <[EMAIL PROTECTED]>
GPG: 0x3B047084 7FC7 0DC0 EF31 DF83 7313  FE2B 77A8 F36A 3B04 7084

Re: debian archive disk space requirements.

[Goswin von Brederlow]

jason andrade <[EMAIL PROTECTED]> writes:

> Hi,
> 
> I've noticed the growth in the debian archive with some concern
> over the last few weeks/months.  We're now hitting the limit of
> the partition that debian is on and it will be quite difficult
> technically for us to deal with this in the short term.
> 
> We have some new storage which is being comissioned but it won't
> be in place for about a month and the debian archive has already
> filled up a 100G partition we've dedicated to it here (that doesn't
> include the debian-cd or debian-non-US or other debian related
> archives.. just the main debian one).
> 
> I'm stuck between a rock and a hard place here as we have a large
> dependency tree before we can move equipment over and get access
> to the new disk..
> 
> Is there any way to reduce the size of the archive over the next
> 4-6 weeks ?

What architectures are you mirroring? You could drop some of the more
uncommon arches (see access logs if any are completly unused) till you
have space for them again. Preferably just drop the Packages file for
the arch and delete as little as possible of the actual debs.

MfG
Goswin

PS: What are you using to mirror?

Re: MEI Whitelist Autoresponse

[Marc Wilson]

On Thu, Aug 28, 2003 at 07:12:47PM -0700, Joshua Kwan wrote:
> Hmm, how about giving tmda its own special header so we can auto-filter
> out messages from people who use C-R systems?

It adds itself to X-Delivery-Agent, so it's not hard to filter out.  I've
started capturing C-R signatures where I can find them and adding them to
procmail /dev/null recipies.  Haven't got many yet, but I'm working on it.

-- 
 Marc Wilson | Stop searching.  Happiness is right next to you.
 [EMAIL PROTECTED] | Now, if they'd only take a bath ...


pgpaBsXpKihMD.pgp
Description: PGP signature

Re: configure web proxy via DHCP server

[Goswin von Brederlow]

Russell Coker <[EMAIL PROTECTED]> writes:

> On Sat, 30 Aug 2003 12:57, H. S. Teoh wrote:
> > > It would be really "cool"(tm) if I didn't have to reconfigure every
> > > program on my laptop to use a different proxy server every time I plug
> > > it into a different network.
> >
> > Run a local proxy that forwards connections to the (external) proxy of
> > your choice, and point all applications at it?
> 
> Then of course there's the issue of how to configure the local proxy.
> 
> For the local proxy on my laptop I use m4 to change it's configuration from a 
> script which knows about the networks I regularly connect to.
> 
> The ideal solution however would be an addition to the DHCP standard for a 
> field to specify this information.  My solution only works AFTER I have 
> figured out the correct data for each network by some other means and 
> customised my script.

Many browsers allow the proxy to be autoconfigured from some info
file. You can easily generate that file depending on your network
setup. google around a bit for the file format.

MfG
Goswin

Re: Bug#207300: tmda: Challenge-response is fundamentally broken

[Karsten M. Self]

on Sat, Aug 30, 2003 at 10:42:17AM +1000, Brian May ([EMAIL PROTECTED]) wrote:
> On Fri, Aug 29, 2003 at 03:48:13PM +1000, Craig Sanders wrote:
> > the point that you keep on missing is that TMDA and similar programs send
> > "confirmation" emails to innocent third-parties who did *NOT* send an email.
> > 
> > TMDA and all C-R systems are broken-by-design, just as many stupid end-user
> > "autoresponders" and AV-scanners that send notifications back to the forged
> > sender address are broken-by-design.
> 
> You saying that any SMTP MTA that sends bounces to unauthenticated
> E-Mail addresses is also broken?

At the very least, this is a small subset of the incoming mail.  There
are probably bad practices, which should be fixed.

The aim is also one which is presumably useful:  if the sender is valid,
then advising them that a message was not delivered is arguably useful
(note that I regard most delivery failure messages as junk).

Most importantly:  the MTA isn't sending mail out willy-nilly to offload
a cost (filtering, content assessment) to a third party.  It's taking an
action on a (hopefully) limited number of mails which cannot be
delivered.

SMTP Envelope reply address should be given precedence, and an SMTP
error precedence over any bounce.

> That is the idea behind autorespoonders after all, to tell the sender
> that his mail didn't get through because it didn't meet some required
> criteria.

"The message can't be delivered because of addressing errors" is a
different class of error than "I can't be bothered to see if this mail
is worth reading, despite its being properly addressed to me".

> Even encryption does not help here, or at least I have not seen any
> proposals for any system that could scale to the Internet. GPG for
> instance only verifies the sender to the receiver, it could not be used
> to verify every sender to the MTAs involved.

A publicly available key, with an email address (or addresses),
validated against contents, is useable.  It doesn't validate the sender,
but it provides a level of indication that someone went through the
trouble of getting a key, posting it publicly, and signing (and/or
encrypting) content with it.

That's more elbow grease than your garden-variety spammer.

Peace.

-- 
Karsten M. Self http://kmself.home.netcom.com/
 What Part of "Gestalt" don't you understand?
   Hollings:  bought, paid for, but couldn't deliver the CBDTPA:
 http://www.politechbot.com/docs/cbdtpa/hollings.s2048.032102.html


pgpea857fP6eC.pgp
Description: PGP signature

Re: configure web proxy via DHCP server

[Javier Fernández-Sanguino Peña]

On Sat, Aug 30, 2003 at 12:01:42PM +1000, Brian May wrote:
> Is this possible?
> 
> It would be really "cool"(tm) if I didn't have to reconfigure every
> program on my laptop to use a different proxy server every time I plug
> it into a different network.
> 

The answer is close to you:
$ apt-cache search laptop configure
(manually removing useless entries)
ifplugd - A configuration daemon for ethernet devices
laptop-net - Automatically adapt laptop ethernet
laptop-net-doc - Automatically adapt laptop ethernet - documentation
laptop-netconf - network detection and configuration program for laptops
whereami - Automatically reconfigure your (laptop) system for a new 
location

All of these, IIRC, provide means to run user-defined scripts for
autoconfiguration. Setting up the proxy server might be just a manner of
setting up the http_proxy environment variable depending on your location
(although some browsers might not use it). 

Regards

Javi


pgpHqNRoNwnBE.pgp
Description: PGP signature

Re: debian archive disk space requirements.

[Josip Rodin]

On Sat, Aug 30, 2003 at 09:29:16AM +0200, Goswin von Brederlow wrote:
> What architectures are you mirroring? You could drop some of the more
> uncommon arches

No, he cannot, it's ftp.au.debian.org. All official mirrors have been, are
or will be in a similar situation.

-- 
 2. That which causes joy or happiness.

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Junichi Uekawa]

>   * Bug fix "debian-changelog mode to support fetching of bug to fill
>in changelog", Thanks to Junichi Uekawa <[EMAIL PROTECTED]>
>(Closes: #207852).
> 
> But I've always thought listing the thitle didn't really say _what_ was
> fixed and _how_.  Most times, the title mentions a symptom but not the
> actual problem.  That why I never added that before.

Yes, that's one point, so the title usually needs editing.
That's why it needs to be done at the level of debian-changelog-mode, 
and not dpkg.
It's not something that can be completely automated, but at least
some template is there to help the maintainer :)


regards,
junichi

Re: configure web proxy via DHCP server

[Jamie Wilkinson]

This one time, at band camp, Russell Coker wrote:
>The ideal solution however would be an addition to the DHCP standard for a 
>field to specify this information.  My solution only works AFTER I have 
>figured out the correct data for each network by some other means and 
>customised my script.

SRV records?

-- 
[EMAIL PROTECTED]   http://people.debian.org/~jaq

Re: debian archive disk space requirements.

[Josip Rodin]

On Sat, Aug 30, 2003 at 11:40:34AM +1000, jason andrade wrote:
> I've noticed the growth in the debian archive with some concern
> over the last few weeks/months.  We're now hitting the limit of
> the partition that debian is on and it will be quite difficult
> technically for us to deal with this in the short term.

I hear you.

/dev/sda3 97G   97G  470M 100.0 [] /bla

> Is there any way to reduce the size of the archive over the next
> 4-6 weeks ?

We are still waiting for Joey to officially announce the obsolescence of
potato on -announce so that it can be moved to archive.debian.org. That will
buy us six, seven gigabytes at most. Other than that, I've no idea. Pray to
god that testing and unstable stop diverging so much? :)

-- 
 2. That which causes joy or happiness.

Re: configure web proxy via DHCP server

[Russell Coker]

On Sat, 30 Aug 2003 18:33, Jamie Wilkinson wrote:
> This one time, at band camp, Russell Coker wrote:
> >The ideal solution however would be an addition to the DHCP standard for a
> >field to specify this information.  My solution only works AFTER I have
> >figured out the correct data for each network by some other means and
> >customised my script.
>
> SRV records?

Almost no-one uses them.  Unless something can get accepted by >50% of the 
servers out there then it's of little use for such things.  If something 
works <50% of the time I may as well do it manually.

-- 
http://www.coker.com.au/selinux/   My NSA Security Enhanced Linux packages
http://www.coker.com.au/bonnie++/  Bonnie++ hard drive benchmark
http://www.coker.com.au/postal/Postal SMTP/POP benchmark
http://www.coker.com.au/~russell/  My home page

Re: MEI Whitelist Autoresponse

[Richard Braakman]

On Thu, Aug 28, 2003 at 02:55:35PM -0700, Adam McKenna wrote:
> How many were challenges from mailing list software?  Yes, another class of
> software that automatically issues challenges (specifically, to new
> subscriptions and to non-list members if the list is closed).  So I guess you
> should also file bugs against majordomo, mailman, ezmlm-src, and any other
> mailing list managers that do this.

I think virus scanners are in a different class, though.  Mailing list
software isn't designed to recognize viruses, while virus scanners are.
It's disgustingly incompetent to recognize a mail as Sobig.F, which is
known to fake the sender, and then reply to it anyway.  (And yes, I
get a lot of "notifications" that mention Sobig.F by name.)

Richard Braakman

Bug#195214: Info received (was .iso conflict, discussion of resolution)

[Debian Bug Tracking System]

Thank you for the additional information you have supplied regarding
this problem report.  It has been forwarded to the package maintainer(s)
and to other interested parties to accompany the original report.

Your message has been sent to the package maintainer(s):
 Mike Markley <[EMAIL PROTECTED]>

If you wish to continue to submit further information on your problem,
please send it to [EMAIL PROTECTED], as before.

Please do not reply to the address at the top of this message,
unless you wish to report a problem with the Bug-tracking system.

Debian bug tracking system administrator
(administrator, Debian Bugs database)

Re: Bug#207300: tmda: Challenge-response is fundamentally broken

[John Hasler]

Brian May writes:
> You saying that any SMTP MTA that sends bounces to unauthenticated
> E-Mail addresses is also broken?

Karsten M. Self writes:
> At the very least, this is a small subset of the incoming mail.

This is about a quarter of my incoming mail.
-- 
John Hasler
[EMAIL PROTECTED]
Dancing Horse Hill
Elmwood, Wisconsin

Re: .iso conflict, discussion of resolution

[Stephan Kulow]

Am Saturday 30 August 2003 08:06 schrieb Chris Cheney:
> Stephan,
>
> Would it be possible to get the two desktop filesi mentioned below merged
> into kdelibs so that arson and k3b are easily installable at the same time?
> I can do the commit myself if you approve.
>
Approved.

Greetings, Stephan

-- 
"SCO doesn't have enough understanding of the case to be able
to lie convincingly. Indeed they have so little understanding
that it's difficult to accuse them of lying." - Greg Lehey

Re: debian archive disk space requirements.

[jason andrade]

On Sat, 30 Aug 2003, Josip Rodin wrote:

> /dev/sda3 97G   97G  470M 100.0 [] /bla

/dev/sdf1100798036  98652428   2145608  98% /raid/lun1p1

and that is by moving a chunk of debian archives into another disk and
symlinking back in..

> > Is there any way to reduce the size of the archive over the next
> > 4-6 weeks ?
>
> We are still waiting for Joey to officially announce the obsolescence of
> potato on -announce so that it can be moved to archive.debian.org. That will
> buy us six, seven gigabytes at most. Other than that, I've no idea. Pray to
> god that testing and unstable stop diverging so much? :)

hmm ok. 6-7G would probably help for 4 weeks.  once we move to the new
disk i don't mind too much about debian growing again :-)

/dev/sdb11220623064 32828 1208380596   1% /data1

i just need to get the time to be able to migrate stuff across..

and i really don't want to break ftp.au.debian.org for people..

regards,

-jason

Re: configure web proxy via DHCP server

[Santiago Garcia Mantinan]

> It would be really "cool"(tm) if I didn't have to reconfigure every
> program on my laptop to use a different proxy server every time I plug
> it into a different network.

The thing is that MS has an extension for DHCP to do this, the code is 252,
it works only with IE in windows, but I've read it also works with konqueror
in kde, I haven't tried this in Debian, though, anybody has tried this?

I think it would be good if we could get this working.

Both dhcp (version 2) and dhcp3-server (version 3) support this extension if
you declare it, here are a couple of examples:

for dhcp3-server:
option auto-proxy-config code 252 = string;

subnet 192.168.1.0 netmask 255.255.255.0 {
  range 192.168.1.101 192.168.1.199;
  option auto-proxy-config "http://192.168.1.1/wpad.pac";;
  option routers 192.168.1.1;
}

for dhcp (version 2) in a normal subnet declaration you add:
  option option-252 "http://192.168.1.1/wpad.pac";

The pac files are the ones that are used to configure netscape, mozilla,
opera or IE, it is a javascript that could be in its simpler form something
like:

function FindProxyForURL(url, host)
{
   return "PROXY 192.168.1.1:3128" ;
}

Here are some links to doc on the topic where you can read more and find
more examples about all this:

http://www.dit.gov.bt/~takeshi/material/win2000srv/internet-gw/wpad/
http://www.microsoft.com/technet/treeview/default.asp?url=/technet/prodtechnol/ie/reskit/ie6/part6/c26ie6rk.asp
http://www.davidpashley.com/tutorials/automatic-proxy.html
http://wp.netscape.com/eng/mozilla/2.0/relnotes/demo/proxy-live.html
http://www.squid-cache.org/Doc/FAQ/FAQ-5.html

Hope this helps.

Regards...
-- 
Manty/BestiaTester -> http://manty.net

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Ross Burton]

On Fri, 2003-08-29 at 20:17, Colin Watson wrote:
> On Fri, Aug 29, 2003 at 08:48:13PM +0200, Mathieu Roy wrote:
> > There's at least one other solution: what if, when a bug tagged
> > "upstream" was closed, the mail sent would include the upstream
> > ChangeLog (hopefully named ChangeLog in the top directory of the
> > package)?
> > Can someone familiar with the BTS code tell whether this change is
> > trivial or not?
> 
> It's not trivial in the slightest, sorry. The BTS doesn't remotely have
> this information available to it, and it's not even easy to arrange for
> it to be available.

I'd also be quite annoyed if I was mailed the upstream changelog when a
bug was closed, as these can often get rather large.

Ross
-- 
Ross Burton mail: [EMAIL PROTECTED]
  jabber: [EMAIL PROTECTED]
 www: http://www.burtonini.com./
 PGP Fingerprint: 1A21 F5B0 D8D0 CFE3 81D4 E25A 2D09 E447 D0B4 33DF




signature.asc
Description: This is a digitally signed message part

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Junichi Uekawa]

> One big problem with this approach is that the same maintainers who are
> too lazy to write proper entries for bug-closers in their changelog
> entries are going to be too lazy to ensure that a bug report has a
> meaningful summary in the first place.

Maintainers who are lazy cannot be fixed, but some may improve, and 
the load on those who do write proper changelogs may be lightened.

I'd not expecting perfect results from here... just some improvements.
Less excuse for making sloppy changelogs :P


regards,
junichi

Re: configure web proxy via DHCP server

[Steve Langasek]

On Sat, Aug 30, 2003 at 02:15:28PM +0200, Santiago Garcia Mantinan wrote:
> > It would be really "cool"(tm) if I didn't have to reconfigure every
> > program on my laptop to use a different proxy server every time I plug
> > it into a different network.

> The thing is that MS has an extension for DHCP to do this, the code is 252,
> it works only with IE in windows, but I've read it also works with konqueror
> in kde, I haven't tried this in Debian, though, anybody has tried this?

> I think it would be good if we could get this working.

> Both dhcp (version 2) and dhcp3-server (version 3) support this extension if
> you declare it, here are a couple of examples:

On the client side, however, at least dhcp3-client requires recompiling
for each option you want to export to the client hook scripts.  This was
recently done for the netbios options that Samba uses.  I know there are
many more unexported options that are potentially useful to other
packages, including ntp timeserver information; it might be nice to
gather a list of them all at once for submission to the dhcp3
maintainers, so they're not constantly recompiling to special-case each
attribute.

-- 
Steve Langasek
postmodern programmer


pgpwdu9mPLuUF.pgp
Description: PGP signature

Re: GDM in sid does not read /etc/environment anymore

[Daniel Ruoso]

But this doesn't solve the problem that the login screen appears without
locales... I already have pam_env.so in /etc/pam.d/gdm, but this only
take effect (maybe I'm wrong) after the user logs in. The greeter will
still appear without locales. (and that's the bug reported in bts).

Em Sáb, 2003-08-30 às 14:20, Gustavo Noronha Silva escreveu:
> Em Fri, 29 Aug 2003 16:38:16 -0300, Daniel Ruoso <[EMAIL PROTECTED]> escreveu:
> 
> > I've actually sent him an email but got no answer. I've posted in
> > debian-devel few days ago and nobody complained that GDM could source
> > /etc/environment in the init script. That's an one-line patch (already
> > tagged as patch in bts for more than a year)... 
> > 
> > I think that if the maintainer doesn't take that, a NMU will be
> > neecessary.
> 
> I don't think that's right. The Right Thing is adding this:
> 
> auth   required   pam_env.so
> 
> to /etc/pam.d/gdm
> 
> My /etc/pam.d/gdm is like this after my own fix:
> 
> [EMAIL PROTECTED] $ cat /etc/pam.d/gdm
> #%PAM-1.0
>  
> # This module parses /etc/environment (the standard for setting
> # environ vars) and also allows you to use an extended config
> # file /etc/security/pam_env.conf.
> # (Replaces the `ENVIRON_FILE' setting from login.defs)
> auth   required   pam_env.so
>  
> @include common-auth
> @include common-account
> @include common-session
-- 
Atenciosamente,

Daniel Ruoso
Desenvolvimento de Sistemas
[EMAIL PROTECTED]
Oktiva Telecomunicaes e Informtica

Re: Bug#207300: tmda: Challenge-response is fundamentally broken

[Osamu Aoki]

I think challenge response needs extra care.  

Anyway, current e-mail worm/virus incident is pretty bad.

On Sat, Aug 30, 2003 at 07:44:56AM -0500, John Hasler wrote:
> Brian May writes:
> > You saying that any SMTP MTA that sends bounces to unauthenticated
> > E-Mail addresses is also broken?
> 
> Karsten M. Self writes:
> > At the very least, this is a small subset of the incoming mail.
> 
> This is about a quarter of my incoming mail.

I filter e-mail worm/virus mail bounces by reading the attached original
mail header.  Most bounces keep the good amount of original header
information.

## Worm e-mails by the header
:0
* ^X-Mailer: Microsoft
* ^X-MailScanner: Found to be clean
Xworm/

## Worm bounces by the header&body
:0 BH
* ^FROM_MAILER
* ^X-Mailer: Microsoft
* ^X-MailScanner: Found to be clean
Xworm-bounce/

I guess our e-mail server can do the similar checks.

Osamu

Re: debian archive disk space requirements.

[Richard Atterer]

On Sat, Aug 30, 2003 at 12:18:50PM +0200, Josip Rodin wrote:
> > Is there any way to reduce the size of the archive over the next 4-6
> > weeks ?
> 
> We are still waiting for Joey to officially announce the obsolescence of
> potato on -announce so that it can be moved to archive.debian.org. That
> will buy us six, seven gigabytes at most. Other than that, I've no idea.
> Pray to god that testing and unstable stop diverging so much? :)

Hm, another possibility is to just make a new stable release, and then to
delete all those woody packages. ;-)

Cheers,

  Richard

-- 
  __   _
  |_) /|  Richard Atterer |  GnuPG key:
  | \/¯|  http://atterer.net  |  0x888354F7
  ¯ '` ¯

Re: debian archive disk space requirements.

[Heikki Vatiainen]

[I am not on debian-devel but reading the list through Usenet gateway]

jason andrade <[EMAIL PROTECTED]> writes:

> /dev/sdf1100798036  98652428   2145608  98% /raid/lun1p1
> 
> and that is by moving a chunk of debian archives into another disk and
> symlinking back in..

You may also want to see mount(8) and look for bind from the manual
page. Since there is not very much about it, I will just quote it
below. From mount(8):

   Since Linux 2.4.0 it is possible to remount part of the file
   hierarchy somewhere else. The call is
  mount --bind olddir newdir

ftp.fi.debian.org has the whole debian/ on the same partition but a
couple of directories, such as debian-cd/, that were previously on the
same partition with the mirror root are now mounted from a new
location with the --bind option.  /etc/fstab also works with bind
option with something like:

/newloc/debian-cd /oldloc/debian-cd none bind,noatime,noexec,nodev

Using bind should keep all mirroring programs happy since they do not
have to get confused when they find a symlink instead of
directory.

-- 
Heikki Vatiainen  * [EMAIL PROTECTED]
Tampere University of Technology  * Tampere, Finland

Re: debian archive disk space requirements.

[Marcelo E. Magallon]

On Sat, Aug 30, 2003 at 12:18:50PM +0200, Josip Rodin wrote:

 > Pray to god that testing and unstable stop diverging so much? :)

 FYI, this is the size of all the binaries belonging to the given
 architecture, in the specifies suites:

 architecture | any| unstable| u+t| u+t+s  | u+t+s+o
--++-+++
 all  | 6907585148 | 3400497424  | 4086646374 | 5709009750 | 686078
 alpha| 8181555068 | 3407524450  | 4906783652 | 6562026256 | 7611929650
 arm  | 6594851918 | 2914482000  | 4192165600 | 5566376046 | 6233666136
 hppa | 6447723766 | 3185688176  | 4554324284 | 5977389114 | 5977389114
 hurd-i386|  554627760 |  554464016  |  554464016 |  554464016 |  554464016
 i386 | 9027667988 | 3693410444  | 5352013018 | 7153730830 | 8206129322
 ia64 | 7683523572 | 3714676258  | 5380423374 | 7121518964 | 7121518964
 m68k | 6396308550 | 2809338886  | 4069205302 | 5393380388 | 6114386710
 mips | 5416612358 | 2851405680  | 3920631098 | 5215452326 | 5215452326
 mipsel   | 5550145962 | 2819073082  | 3995118416 | 5262682024 | 5262682024
 powerpc  | 7680418788 | 3336303070  | 4780596742 | 6309202432 | 7112407388
 s390 | 6147469798 | 3053291988  | 4427782760 | 5820464436 | 5820464436
 sparc| 7331766068 | 3071979288  | 921552 | 5917654266 | 6790963196

 "any" means any of:

 experimental
 oldstable
 old-proposed-updates
 proposed-updates
 unstable
 testing
 testing-proposed-updates
 stable

 stable+proposed-updates:

 architecture |size
--+
 all  | 2195553250
 alpha| 2438601088
 arm  | 1951980644
 hppa | 260648
 i386 | 2753622878
 ia64 | 2531882184
 m68k | 1888659560
 mips | 1801959356
 mipsel   | 1763094946
 powerpc  | 2196413794
 s390 | 1947506370
 sparc| 2065059050

 testing+testing-proposed-updates:

 architecture |size
--+
 all  | 3113135590
 alpha| 3050200486
 arm  | 2667813072
 hppa | 2850080658
 i386 | 3378532922
 ia64 | 3394287226
 m68k | 2571810596
 mips | 2569942598
 mipsel   | 2531507858
 powerpc  | 2996108486
 s390 | 2713124396
 sparc| 2740423638

 From the numbers quoted above you can make a reasonable estimation of
 how much unstable and testing overlap.  You can assume
 testing-proposed-updates is empty (you'll be off only by a few MB).  My
 napkin says 50%, which IMO is hideously bad, but that's me.

 Drop a factor of 2-3 somewhere if you want to consider source.

 This data does not come from the filesystem, but from the database used
 to wrangle the whole mess.  Account for 10% wasted space or so for a
 guess of how much space you actually need for a mirror (adapt to your
 filesystem of choice -- information which I don't really care about).

-- 
Marcelo

Re: Work-needing packages report for Aug 29, 2003

[Marcelo E. Magallon]

On Sat, Aug 30, 2003 at 06:57:38PM +0200, Simon Richter wrote:

 > > [NEW] 3dwm (#206870), orphaned 5 days ago
 > >  Description: libzorn development files
 > >  Reverse Depends: 3dwm-pickclient 3dwm-texclient 3dwm-csgclient
 > >  libcelsius-dev libpolhem-dev libgarbo-dev libnobel-dev
 > >  3dwm-vncclient libsolid-dev 3dwm-clock 3dwm-server libzorn-dev
 > >  3dwm-geoclient
 > 
 > Hrm, I don't think that description matches the package, i.e. there
 > might be something wrong with the scripts. Could anyone familiar with
 > them please check?

 It is, as strange as it might seem, correct.  Source packages don't
 have descriptions and Martin obviously hacked up something to write the
 bug reports automatically (which I find irritating but it's better than
 the option).  The libzorn-dev description was picked up as "the"
 description for the source package.

-- 
Marcelo

Re: debian archive disk space requirements.

[Josip Rodin]

On Sat, Aug 30, 2003 at 08:25:36PM +0300, Heikki Vatiainen wrote:
> ftp.fi.debian.org has the whole debian/ on the same partition but a
> couple of directories, such as debian-cd/, that were previously on the
> same partition with the mirror root are now mounted from a new
> location with the --bind option.

It's harder within debian/ because you have to hand-pick parts of pool/ to
bind-mount... messy.

-- 
 2. That which causes joy or happiness.

Re: Work-needing packages report for Aug 29, 2003

[Simon Richter]

Hi,

> [NEW] 3dwm (#206870), orphaned 5 days ago
>  Description: libzorn development files
>  Reverse Depends: 3dwm-pickclient 3dwm-texclient 3dwm-csgclient
>  libcelsius-dev libpolhem-dev libgarbo-dev libnobel-dev
>  3dwm-vncclient libsolid-dev 3dwm-clock 3dwm-server libzorn-dev
>  3dwm-geoclient

Hrm, I don't think that description matches the package, i.e. there
might be something wrong with the scripts. Could anyone familiar with
them please check?

   Simon

-- 
GPG Fingerprint: 040E B5F7 84F1 4FBC CEAD  ADC6 18A0 CC8D 5706 A4B4


pgpt2KwqFl1oc.pgp
Description: PGP signature

Re: GDM in sid does not read /etc/environment anymore

[Gustavo Noronha Silva]

Em Fri, 29 Aug 2003 16:38:16 -0300, Daniel Ruoso <[EMAIL PROTECTED]> escreveu:

> I've actually sent him an email but got no answer. I've posted in
> debian-devel few days ago and nobody complained that GDM could source
> /etc/environment in the init script. That's an one-line patch (already
> tagged as patch in bts for more than a year)... 
> 
> I think that if the maintainer doesn't take that, a NMU will be
> neecessary.

I don't think that's right. The Right Thing is adding this:

auth   required   pam_env.so

to /etc/pam.d/gdm

My /etc/pam.d/gdm is like this after my own fix:

[EMAIL PROTECTED] $ cat /etc/pam.d/gdm
#%PAM-1.0
 
# This module parses /etc/environment (the standard for setting
# environ vars) and also allows you to use an extended config
# file /etc/security/pam_env.conf.
# (Replaces the `ENVIRON_FILE' setting from login.defs)
auth   required   pam_env.so
 
@include common-auth
@include common-account
@include common-session


-- 
[EMAIL PROTECTED]: Gustavo Noronha 
Debian:   *  
Dúvidas sobre o Debian? Visite o Rau-Tu: http://rautu.cipsga.org.br

Re: debian archive disk space requirements.

[Josip Rodin]

On Sat, Aug 30, 2003 at 12:18:50PM +0200, Josip Rodin wrote:
> We are still waiting for Joey to officially announce the obsolescence of
> potato on -announce so that it can be moved to archive.debian.org.

I've found out today that Joey doesn't feel there should be any more
announcements. I've sent an e-mail so I guess ftpmasters just need to find
the time to pull out the magic wand, move leftover source packages to pool
update anything else that is necessary, and move old potato stuff out.

-- 
 2. That which causes joy or happiness.

[no subject]

[Dr. Rasoul Shahpari]

To whom it may concern 

Pls be informed that we are 
an international company active in the commercial fields of the importation of 
the 
1- 
ironeries
2- 
aluminum , 
copper 
3- 
sugar 

4- 
rice
5- 
food 
grains
6- 
petrochemical 
items
7- 
selling and 
buying of the lands
We are interested to work 
with the investors or the financers or partners or the money lenders in all over 
the world that they want to cooperate in the partnership with 
us,
in case u are interested in 
these businesses with us , u are kindly requested to contact with us 
,
pls be advised that the 
minimum profit that will be guaranteed by our company will be 20% per cent  of the total amount of the investment in 
the year  , we will guarantee the 
whole capital of your self in this profitable business 

With thanks 

saeed 
bradaran
The managing director of 
the beh azin shargh
http://www.iranazin.net
e-mail: [EMAIL PROTECTED]
00989132583522
meanwhile , pls be informed 
that , I will never , ever pay any thing before or upfront for the different 
fees or expenses that will be occurred by the lawyer or the bank transfer , 
iwill not be responsible for the fees and will never pay anything to u or the 
lawyer  or 
etc………..,
frankly in case u are 
writing this letter to arrange for the payment of some u.s.d. and with writing 
this letter that u have this amount and try to get some small money like for 
example 8000 , u.s.d. or more and less , I will not pay even one cent upfront in 
any form to anybody , incase the business is okay or real and genuine for the 
investment , of course will not need these tricks and playing , If  you will accept my condition . no 
problem , the a/c no of mine is as follows :
CREDIT SUISSE ( HEAD OFFICE 
) – ZURICH
SWIFT CODE : CRESCHZZ80A 

A/C NO. 
903137-44
FOR FURTHER CREDIT TO 
EDB
IN FAVOUR OF ……… RASOUL 
SHAHPARI
A/C NO……….. 
010092971
SENDER TO RECEIVER 
INFORMATION  

PLS TRANSFER UNDER MT 
100/103 MSG
U ARE KINDLY REQUESTED TO 
ARRANGE FOR THE TRANSFER AND THEN COME TO 
IRAN TO MAKE THE INVESTMENTS AND GIVE ME 
UR A/C TO PAY MONTHLY TO UR A/C 
BE ASSURED OF ALL OUR 
SERVICES WITH THE HONESTY 
(HONESTY IS THE BEST POLICY 
)
THANKS 

[EMAIL PROTECTED]

[no subject]

[Dr. Rasoul Shahpari]

To whom it may concern 

Pls be informed that we are 
an international company active in the commercial fields of the importation of 
the 
1- 
ironeries
2- 
aluminum , 
copper 
3- 
sugar 

4- 
rice
5- 
food 
grains
6- 
petrochemical 
items
7- 
selling and 
buying of the lands
We are interested to work 
with the investors or the financers or partners or the money lenders in all over 
the world that they want to cooperate in the partnership with 
us,
in case u are interested in 
these businesses with us , u are kindly requested to contact with us 
,
pls be advised that the 
minimum profit that will be guaranteed by our company will be 20% per cent  of the total amount of the investment in 
the year  , we will guarantee the 
whole capital of your self in this profitable business 

With thanks 

saeed 
bradaran
The managing director of 
the beh azin shargh
http://www.iranazin.net
e-mail: [EMAIL PROTECTED]
00989132583522
meanwhile , pls be informed 
that , I will never , ever pay any thing before or upfront for the different 
fees or expenses that will be occurred by the lawyer or the bank transfer , 
iwill not be responsible for the fees and will never pay anything to u or the 
lawyer  or 
etc………..,
frankly in case u are 
writing this letter to arrange for the payment of some u.s.d. and with writing 
this letter that u have this amount and try to get some small money like for 
example 8000 , u.s.d. or more and less , I will not pay even one cent upfront in 
any form to anybody , incase the business is okay or real and genuine for the 
investment , of course will not need these tricks and playing , If  you will accept my condition . no 
problem , the a/c no of mine is as follows :
CREDIT SUISSE ( HEAD OFFICE 
) – ZURICH
SWIFT CODE : CRESCHZZ80A 

A/C NO. 
903137-44
FOR FURTHER CREDIT TO 
EDB
IN FAVOUR OF ……… RASOUL 
SHAHPARI
A/C NO……….. 
010092971
SENDER TO RECEIVER 
INFORMATION  

PLS TRANSFER UNDER MT 
100/103 MSG
U ARE KINDLY REQUESTED TO 
ARRANGE FOR THE TRANSFER AND THEN COME TO 
IRAN TO MAKE THE INVESTMENTS AND GIVE ME 
UR A/C TO PAY MONTHLY TO UR A/C 
BE ASSURED OF ALL OUR 
SERVICES WITH THE HONESTY 
(HONESTY IS THE BEST POLICY 
)
THANKS 

[EMAIL PROTECTED]

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Matt Zimmerman]

On Fri, Aug 29, 2003 at 04:34:58PM -0500, Adam Heath wrote:

> On Fri, 29 Aug 2003, Glenn Maynard wrote:
> 
> > If I report "segmentation fault in ls", I--as a user of ls, not a
> > developer--couldn't care less about why it was segfaulting or how the
> > bug was fixed; I only care that it's been fixed.  If a developer wants
> > to spend their limited time researching how the bug was fixed and
> > summarizing it in a changelog, great, but it's certainly not something I'd
> > expect everyone to do.
> 
> It's not about summarizing how the bug was fixed.  It's about summarizing the
> bug *itself* in the changelog.

I certainly prefer it if the changelog tells how the bug was fixed.  This
documents the difference between:

 * New upstream release
   - Removed the entire subsystem which contained this bug (Closes: #xxx)

 * New upstream release
   - Made the "foo" option create its file with sane permissions (Closes: #xxx)

-- 
 - mdz

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Matt Zimmerman]

On Sat, Aug 30, 2003 at 08:36:16AM +0200, Andreas Metzler wrote:

> Peter S Galbraith <[EMAIL PROTECTED]> wrote: [...]
> > Right.  I understood both points.  I was wondering about having the bug
> > submitter there.  Maybe change the phrasing?
> 
> I usually don't list him/her, my changelogs are too long already. I do
> list submitters who send the report by private mail instead of via BTS,
> because othervise they wouldn't have any public record of their help.

I list the submitter when they have provided a patch, so as to provide for
attribution, and therefore credit or blame, as appropriate.

-- 
 - mdz

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Matt Zimmerman]

On Sat, Aug 30, 2003 at 02:39:01PM -0400, Matt Zimmerman wrote:

> On Sat, Aug 30, 2003 at 08:36:16AM +0200, Andreas Metzler wrote:
> 
> > Peter S Galbraith <[EMAIL PROTECTED]> wrote: [...]
> > > Right.  I understood both points.  I was wondering about having the bug
> > > submitter there.  Maybe change the phrasing?
> > 
> > I usually don't list him/her, my changelogs are too long already. I do
> > list submitters who send the report by private mail instead of via BTS,
> > because othervise they wouldn't have any public record of their help.
> 
> I list the submitter when they have provided a patch, so as to provide for
> attribution, and therefore credit or blame, as appropriate.

And also, I suppose, if the submitter did a good job of tracking down the
bug and saved me time, and deserved credit for that.

-- 
 - mdz

Re: .iso conflict, discussion of resolution

[Simon Richter]

Hi,

> A quick summary of this bug:
> Arson, a KDE CD burning application, includes two .desktop files to
> associate certain files with it:
> /usr/share/mimelnk/application/x-iso.desktop
> /usr/share/mimelnk/application/x-cue.desktop

This sounds like you may want to use an alternative here, so that the
sysadmin can choose which program to associate .iso files with by
default. Of course, users must be able to override this, but I think
that is given (is it?).

Alternatives will have to be supported by any package that brings such a
file, but need no support from unrelated packages. Putting the file in
question into a base package is bad because it may associate the file
with a program that is not installed.

   Simon

-- 
GPG Fingerprint: 040E B5F7 84F1 4FBC CEAD  ADC6 18A0 CC8D 5706 A4B4


pgpvonHOLMhYo.pgp
Description: PGP signature

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Branden Robinson]

On Sat, Aug 30, 2003 at 02:39:38PM -0400, Matt Zimmerman wrote:
> On Sat, Aug 30, 2003 at 02:39:01PM -0400, Matt Zimmerman wrote:
> > I list the submitter when they have provided a patch, so as to provide for
> > attribution, and therefore credit or blame, as appropriate.
> 
> And also, I suppose, if the submitter did a good job of tracking down the
> bug and saved me time, and deserved credit for that.

Yup.  Both of those scenarios are exactly how I have been using my
"thanks" messages in the XFree86 changelogs for years.

-- 
G. Branden Robinson|Religion is regarded by the common
Debian GNU/Linux   |people as true, by the wise as
[EMAIL PROTECTED] |false, and by the rulers as useful.
http://people.debian.org/~branden/ |-- Lucius Annaeus Seneca


pgpMvEDC1NHUp.pgp
Description: PGP signature

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Branden Robinson]

On Sat, Aug 30, 2003 at 11:06:20PM +0900, Junichi Uekawa wrote:
> > One big problem with this approach is that the same maintainers who are
> > too lazy to write proper entries for bug-closers in their changelog
> > entries are going to be too lazy to ensure that a bug report has a
> > meaningful summary in the first place.
> 
> Maintainers who are lazy cannot be fixed, but some may improve, and 
> the load on those who do write proper changelogs may be lightened.
> 
> I'd not expecting perfect results from here... just some improvements.
> Less excuse for making sloppy changelogs :P

Fair enough.  Go for it, then!  :)

-- 
G. Branden Robinson|
Debian GNU/Linux   |   If ignorance is bliss,
[EMAIL PROTECTED] |   is omniscience hell?
http://people.debian.org/~branden/ |


pgpWfmWh3DSws.pgp
Description: PGP signature

overwriting files from modules packages

[martin f krafft]

I am the (new) maintainer of bcm5700-source, a modules package for
the broadcom gigabit adapter. The final package,
bcm5700-module-${KVERS}, includes a manpage,
/usr/share/man/man4/bcm5700.4.gz. I just now ran into the problem
that while installing the 2.4.22 image and modules, the
bcm5700-modules-2.4.22 package attempted to install that manpage,
even though bcm5700-modules-2.4.21 already put it there.

This is a special situation, and I am not sure how to handle it.
I don't want to create the manpage in the postinst because then it
would not be under control of the package management system. I could
install a manpage including the version number and simply symlink,
but then it may be that i have symlinked to 2.4.21 even though
2.4.22 includes an updated version, which is the one that's relevant
if I booted 2.4.22.

What to do? I appreciate any advice.

-- 
Please do not CC me when replying to lists; I read them!
 
 .''`. martin f. krafft <[EMAIL PROTECTED]>
: :'  :proud Debian developer, admin, and user
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!


pgp93jwMlpSzr.pgp
Description: PGP signature

Re: debian archive disk space requirements.

[Matt Zimmerman]

On Sat, Aug 30, 2003 at 11:40:34AM +1000, jason andrade wrote:

> Is there any way to reduce the size of the archive over the next
> 4-6 weeks ?

Drop potato?

-- 
 - mdz

Bug#181429: retitle 181429 ITP: grubconf -- Gnoretitle 181429 ITP: grubconf -- Gnome2 based GRUB configuration editor

[Luis R. Rodriguez]

retitle 181429 ITP: grubconf -- Gnome2 based GRUB configuration editor
thanks

* Package name : grubconf
Version : 0.5
Upstream Author :   Joseph Monti <[EMAIL PROTECTED]>,
Ryan Scotka <[EMAIL PROTECTED]>
* URL : http://grubconf.sourceforge.net/
* License : GPL


ABOUT

Grubconf is a Gnome2 based GRUB configuration editor. It provides
an easy to use interface allowing effortless modification of OS's
and the flexibility to configure the most obscure options.
Designed to require minimal user interaction while providing tools
for the most adventurous user.

Grubconf must be run as ROOT. This is to gain access to the
grub configuration file.

Notes: I need a sponsor. The package is ready. 

Re: .iso conflict, discussion of resolution

[Matt Zimmerman]

On Fri, Aug 29, 2003 at 10:34:39PM -0700, Mike Markley wrote:

> A quick summary of this bug:
> Arson, a KDE CD burning application, includes two .desktop files to
> associate certain files with it:
> /usr/share/mimelnk/application/x-iso.desktop
> /usr/share/mimelnk/application/x-cue.desktop

And presumably another application also wants to include these files?  Is
KDE's file association system really so broken that two programs cannot both
declare themselves able to handle a certain type of file?

-- 
 - mdz

Re: overwriting files from modules packages

[Matt Zimmerman]

On Sat, Aug 30, 2003 at 11:05:40PM +0200, martin f krafft wrote:

> I am the (new) maintainer of bcm5700-source, a modules package for the
> broadcom gigabit adapter. The final package, bcm5700-module-${KVERS},
> includes a manpage, /usr/share/man/man4/bcm5700.4.gz. I just now ran into
> the problem that while installing the 2.4.22 image and modules, the
> bcm5700-modules-2.4.22 package attempted to install that manpage, even
> though bcm5700-modules-2.4.21 already put it there.
> 
> This is a special situation, and I am not sure how to handle it.  I don't
> want to create the manpage in the postinst because then it would not be
> under control of the package management system. I could install a manpage
> including the version number and simply symlink, but then it may be that i
> have symlinked to 2.4.21 even though 2.4.22 includes an updated version,
> which is the one that's relevant if I booted 2.4.22.
> 
> What to do? I appreciate any advice.

Kernel module packages should be treated with the same care as shared
library packages, since different versions need to coexist.  Anything which
could conflict needs to either be versioned in the filesystem, or moved into
a separate package.

If you want two different versions of documentation, they need to be named
differently.  Alternatively, you can just split the documentation into a
separate package and allow only one version to be installed at once.

-- 
 - mdz

Re: Bug#195214: .iso conflict, discussion of resolution

[Mike Markley]

On Sat, Aug 30, 2003 at 08:17:05PM +0200, Simon Richter <[EMAIL PROTECTED]> 
wrote:
> This sounds like you may want to use an alternative here, so that the
> sysadmin can choose which program to associate .iso files with by
> default. Of course, users must be able to override this, but I think
> that is given (is it?).
> 
> Alternatives will have to be supported by any package that brings such a
> file, but need no support from unrelated packages. Putting the file in
> question into a base package is bad because it may associate the file
> with a program that is not installed.

I misspoke about the purpose; these files don't actually associate
file types, just describe them. The only reasona arson and k3b both
provide them is that nothing else currently does.

-- 
Mike Markley <[EMAIL PROTECTED]>
GPG: 0x3B047084 7FC7 0DC0 EF31 DF83 7313  FE2B 77A8 F36A 3B04 7084

Re: overwriting files from modules packages

[martin f krafft]

also sprach Matt Zimmerman <[EMAIL PROTECTED]> [2003.08.31.0013 +0200]:
> If you want two different versions of documentation, they need to
> be named differently.  Alternatively, you can just split the
> documentation into a separate package and allow only one version
> to be installed at once.

I don't think a single manpage warrants another binary package. So
the only real solution is to put the manpage into
/usr/share/doc//docs ...

-- 
Please do not CC me when replying to lists; I read them!
 
 .''`. martin f. krafft <[EMAIL PROTECTED]>
: :'  :proud Debian developer, admin, and user
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!


pgpbYIHLfTSCA.pgp
Description: PGP signature

Re: Bug#195214: .iso conflict, discussion of resolution

[Mike Markley]

On Sat, Aug 30, 2003 at 02:13:10PM +0200, Stephan Kulow <[EMAIL PROTECTED]> 
wrote:
> Am Saturday 30 August 2003 08:06 schrieb Chris Cheney:
> > Would it be possible to get the two desktop filesi mentioned below merged
> > into kdelibs so that arson and k3b are easily installable at the same time?
> > I can do the commit myself if you approve.
> >
> Approved.
> 
> Greetings, Stephan

Thank you, guys. I will remove the files from arson. I'm glad we were
able to get a resolution here. :)

-- 
Mike Markley <[EMAIL PROTECTED]>
GPG: 0x3B047084 7FC7 0DC0 EF31 DF83 7313  FE2B 77A8 F36A 3B04 7084

Re: debian archive disk space requirements.

[jason andrade]

On Sun, 30 Aug 2003, Heikki Vatiainen wrote:

> [I am not on debian-devel but reading the list through Usenet gateway]

i didn't realise debian-devel is gatewayed to a newsgroup.. interesting.

> You may also want to see mount(8) and look for bind from the manual
> page. Since there is not very much about it, I will just quote it
> below. From mount(8):

[...]

> ftp.fi.debian.org has the whole debian/ on the same partition but a
> couple of directories, such as debian-cd/, that were previously on the
> same partition with the mirror root are now mounted from a new
> location with the --bind option.  /etc/fstab also works with bind
> option with something like:
>
> /newloc/debian-cd /oldloc/debian-cd none bind,noatime,noexec,nodev
>
> Using bind should keep all mirroring programs happy since they do not
> have to get confused when they find a symlink instead of
> directory.

i'm not entirely sure how this helps unless you mean we can use the
above to start moving parts of the debian tree to different partitions
and then remounting them inside the tree..  this would be a nice
temporary solution except that we have a more complicated architecture
as we have a back end fileserver where everything is mirrored and
front end servers that actually provide the ftp/http/rsync services
which read only NFS mount from the back end..

once you start doing loops and binds on the back end it tends to
fall over with NFS structures :-/

thanks for the tip though.

regards,

-jason

Re: configure web proxy via DHCP server

[Brian May]

On Sat, Aug 30, 2003 at 10:30:31AM +0200, Javier Fern?ndez-Sanguino Pe?a wrote:
> All of these, IIRC, provide means to run user-defined scripts for
> autoconfiguration. Setting up the proxy server might be just a manner of
> setting up the http_proxy environment variable depending on your location
> (although some browsers might not use it). 

No, setting http_proxy wont work unless youo log out and back in again,
or at least close all your terminal sessions.

It won't work for programs like Mozilla.
-- 
Brian May <[EMAIL PROTECTED]>

Re: overwriting files from modules packages

[Matt Zimmerman]

On Sun, Aug 31, 2003 at 12:18:26AM +0200, martin f krafft wrote:

> also sprach Matt Zimmerman <[EMAIL PROTECTED]> [2003.08.31.0013 +0200]:
> > If you want two different versions of documentation, they need to
> > be named differently.  Alternatively, you can just split the
> > documentation into a separate package and allow only one version
> > to be installed at once.
> 
> I don't think a single manpage warrants another binary package. So
> the only real solution is to put the manpage into
> /usr/share/doc//docs ...

Or name it -..

-- 
 - mdz

Re: configure web proxy via DHCP server

[Brian May]

On Sat, Aug 30, 2003 at 09:56:38AM -0500, Steve Langasek wrote:
> On the client side, however, at least dhcp3-client requires recompiling
> for each option you want to export to the client hook scripts.  This was

You do? I was kindof hoping that with /etc/dhcp3/dhclient*-hooks.d/,
it would be just a matter of adding an extra script that runs...

> recently done for the netbios options that Samba uses.  I know there are
> many more unexported options that are potentially useful to other
> packages, including ntp timeserver information; it might be nice to
> gather a list of them all at once for submission to the dhcp3
> maintainers, so they're not constantly recompiling to special-case each
> attribute.

Sounds like a good idea...

So we can put the proxy server on the DHCP server, in the form
of a reference to a PAC file.

We can (with modifications) retrieve the reference to the PAC
file on the client.

Then what?

I think that there are two cases:

1. browsers that don't support PAC files, eg. lynx, links, and
w3m probably don't (correct me if I am wrong).

2. browsers that do support PAC files, but need to be configured
to use it for each user on the system.
-- 
Brian May <[EMAIL PROTECTED]>

Re: configure web proxy via DHCP server

[Steve Langasek]

On Sat, Aug 30, 2003 at 11:18:25PM +, Brian May wrote:
> On Sat, Aug 30, 2003 at 09:56:38AM -0500, Steve Langasek wrote:
> > On the client side, however, at least dhcp3-client requires recompiling
> > for each option you want to export to the client hook scripts.  This was

> You do? I was kindof hoping that with /etc/dhcp3/dhclient*-hooks.d/,
> it would be just a matter of adding an extra script that runs...

Ah, I'm sorry, I misremembered -- the actual problem is not the lack of
option exporting, it's that recent dhcp clients and servers will
negotiate with one another to decide what options will actually be sent
(presumably to reduce the payload size); so unless the dhcp3-client is
configured to explicitly request such options, they won't be sent by the
server.  So strictly speaking, this doesn't have to be done with a
recompile, but it does require the intervention of the dhcp3 maintainers
if you want this support by default (i.e., without an admin having to
edit dhclient.conf).

> > recently done for the netbios options that Samba uses.  I know there are
> > many more unexported options that are potentially useful to other
> > packages, including ntp timeserver information; it might be nice to
> > gather a list of them all at once for submission to the dhcp3
> > maintainers, so they're not constantly recompiling to special-case each
> > attribute.

> Sounds like a good idea...

> So we can put the proxy server on the DHCP server, in the form
> of a reference to a PAC file.

> We can (with modifications) retrieve the reference to the PAC
> file on the client.

> Then what?

> I think that there are two cases:

> 1. browsers that don't support PAC files, eg. lynx, links, and
> w3m probably don't (correct me if I am wrong).

> 2. browsers that do support PAC files, but need to be configured
> to use it for each user on the system.

For 1. above, I imagine most (if not all) browsers that don't support
javascript /do/ support global config files.  At minimum, therefore,
each such browser package could add its own hook script to
/etc/dhcp3/dhclient-enter-hooks.d/ that would update the config file on
disk.  This still leaves the question of how to refresh the
configuration of any running browsers, but that's probably a question to
be sorted out individually with the respective upstream developers
rather than here.

-- 
Steve Langasek
postmodern programmer


pgp5SZVbioNwN.pgp
Description: PGP signature

Re: configure web proxy via DHCP server

[Marco d'Itri]

On Aug 30, Russell Coker <[EMAIL PROTECTED]> wrote:

 >The ideal solution however would be an addition to the DHCP standard for a 
 >field to specify this information.  My solution only works AFTER I have 
http://www.wpad.com/

-- 
ciao, |
Marco | [1564 scuoJF5IqmYsk]

Re: debian archive disk space requirements.

[Marco d'Itri]

On Aug 30, Josip Rodin <[EMAIL PROTECTED]> wrote:

 >It's harder within debian/ because you have to hand-pick parts of pool/ to
 >bind-mount... messy.
BTDT. It's even harder because the archive contain hard links between
random directories.

-- 
ciao, |
Marco | [1565 id6qO1u0SBERA]

Re: debian archive disk space requirements.

[Diego Calleja García]

El 30 Aug 2003 20:25:36 +0300 Heikki Vatiainen <[EMAIL PROTECTED]> escribió:

>Since Linux 2.4.0 it is possible to remount part of the file
>hierarchy somewhere else. The call is
>   mount --bind olddir newdir

with 2.6 we also got mount --move olddir newdir 8)

Re: Bug#207300: tmda: Challenge-response is fundamentally broken

[Brian May]

On Sat, Aug 30, 2003 at 04:01:19PM +1000, Russell Coker wrote:
> > That is the idea behind autorespoonders after all, to tell the sender
> > that his mail didn't get through because it didn't meet some required
> > criteria.
> 
> A SMTP 550 code can convey all the information that is needed for bounces.

There are two problems with this.

1. The modular design of SMTP agents like postfix do not allow 
scanning of messages before the message has been accepted by the
MTA at the SMTP session. I think you would have to add hooks
into smtpd, but that is going to complicate the code.

2. All checks have to be automatic, and there is no chance of manual
review to ensure that the messages where geniune before bouncing it.

The list of known solutions follows:

(nil)


;-)

I have considered to possibility of using something like
zorp to act as a proxy SMTP server between the client and the
real server, but that would not work for encrypted SMTP
communications.

...and Yes, now that I have enabled SSL on my postfix,
at least some spammers do use encrypted SMTP sessions.
-- 
Brian May <[EMAIL PROTECTED]>

Re: overwriting files from modules packages

[martin f krafft]

also sprach Matt Zimmerman <[EMAIL PROTECTED]> [2003.08.31.0130 +0200]:
> > I don't think a single manpage warrants another binary package. So
> > the only real solution is to put the manpage into
> > /usr/share/doc//docs ...
> 
> Or name it -..

Which is what I will do, probably. It's just annoying... but
a statement in README.Debian will do...

-- 
Please do not CC me when replying to lists; I read them!
 
 .''`. martin f. krafft <[EMAIL PROTECTED]>
: :'  :proud Debian developer, admin, and user
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!


pgpuOrEbjBUo7.pgp
Description: PGP signature

Re: Bug#207300: tmda: Challenge-response is fundamentally broken

[Steve Lamb]

On Sat, 30 Aug 2003 23:49:40 +
Brian May <[EMAIL PROTECTED]> wrote:
> 1. The modular design of SMTP agents like postfix do not allow 
> scanning of messages before the message has been accepted by the
> MTA at the SMTP session. I think you would have to add hooks
> into smtpd, but that is going to complicate the code.

Well, that's postfix's problem.  After years of hearing how modular is the
superior method it is kind of ironic that a simple solution is complicated.
 
> 2. All checks have to be automatic, and there is no chance of manual
> review to ensure that the messages where geniune before bouncing it.

Trust me, if SA scores it high enough do you really want to worry
about it?  Running SA-Exim here and with sensible defaults I've 550'd most
spam and, after the Bayesian filtering caught up, most of the recent viruses.
I am perfectly capable of of reviewing the messages as I have the option of
saving them for review.  I define the levels at which things get accepted,
rejected or even teergrubed. So far my false positives have been exactly 0 for
rejects as I allow a small window for grey area messages to get through for
the user to manage.  In that window exactly 2 messages were false-positives
and they were/barely/ high enough to be marked as spam.  This has been so
effective I've been working on getting clamav hacked into Spamassassin so that
future virus attacks can be avoided without the painful Bayesian ramp-up.

-- 
 Steve C. Lamb | I'm your priest, I'm your shrink, I'm your
   PGP Key: 8B6E99C5   | main connection to the switchboard of souls.
---+-


pgpV7r5ETQv8e.pgp
Description: PGP signature

Re: Bug#207300: tmda: Challenge-response is fundamentally broken

[Steve Langasek]

On Sat, Aug 30, 2003 at 05:26:59PM -0700, Steve Lamb wrote:
> On Sat, 30 Aug 2003 23:49:40 +
> Brian May <[EMAIL PROTECTED]> wrote:
> > 1. The modular design of SMTP agents like postfix do not allow 
> > scanning of messages before the message has been accepted by the
> > MTA at the SMTP session. I think you would have to add hooks
> > into smtpd, but that is going to complicate the code.

> Well, that's postfix's problem.  After years of hearing how modular is the
> superior method it is kind of ironic that a simple solution is complicated.

I don't see any reason why postfix couldn't be readily adapted to
support hooks to external program checks, if it really doesn't support
them already.  You can do just about any other kind of check you want to
at any point in the receiving process, external program checks shouldn't
be all that hard to add.

-- 
Steve Langasek
postmodern programmer


pgpShfmHfA2tw.pgp
Description: PGP signature

Re: .iso conflict, discussion of resolution

[Chris Cheney]

On Sat, Aug 30, 2003 at 06:08:17PM -0400, Matt Zimmerman wrote:
> On Fri, Aug 29, 2003 at 10:34:39PM -0700, Mike Markley wrote:
> 
> > A quick summary of this bug:
> > Arson, a KDE CD burning application, includes two .desktop files to
> > associate certain files with it:
> > /usr/share/mimelnk/application/x-iso.desktop
> > /usr/share/mimelnk/application/x-cue.desktop
> 
> And presumably another application also wants to include these files?  Is
> KDE's file association system really so broken that two programs cannot both
> declare themselves able to handle a certain type of file?

Nope, mime types are completely independent of them declaring if they
support it. That is why the mimetype really belongs in kdelibs itself
and not the 3rd party package.

Chris


pgpbRneFe2HLg.pgp
Description: PGP signature

Re: Bug#207300: tmda: Challenge-response is fundamentally broken

[Karsten M. Self]

on Sat, Aug 30, 2003 at 07:44:56AM -0500, John Hasler ([EMAIL PROTECTED]) wrote:
> Brian May writes:
> > You saying that any SMTP MTA that sends bounces to unauthenticated
> > E-Mail addresses is also broken?
> 
> Karsten M. Self writes:
> > At the very least, this is a small subset of the incoming mail.
> 
> This is about a quarter of my incoming mail.

Which?  Bounces to spoofed senders, or improperly addressed mail?

What prevents you from 550ing this at SMTP connect?

Peace.

-- 
Karsten M. Self http://kmself.home.netcom.com/
 What Part of "Gestalt" don't you understand?
   KQED FM:  The bright spot on the dial:  http://www.kqed.org/fm/


pgpIh22S5cRqH.pgp
Description: PGP signature

Re: GDM in sid does not read /etc/environment anymore

[Gustavo Noronha Silva]

Em Sat, 30 Aug 2003 14:31:32 -0300, Daniel Ruoso <[EMAIL PROTECTED]> escreveu:

> But this doesn't solve the problem that the login screen appears without
> locales... I already have pam_env.so in /etc/pam.d/gdm, but this only
> take effect (maybe I'm wrong) after the user logs in. The greeter will
> still appear without locales. (and that's the bug reported in bts).

Indeed, but the problem is gdm is not reading /etc/environment anymore
when the user logs in. So that also needs to be fixed. I agree with
trying to find LANG in /etc/environment.

[]s!

-- 
[EMAIL PROTECTED]: Gustavo Noronha 
Debian:   *  
Dúvidas sobre o Debian? Visite o Rau-Tu: http://rautu.cipsga.org.br

Re: overwriting files from modules packages

[Junichi Uekawa]

> Which is what I will do, probably. It's just annoying... but
> a statement in README.Debian will do...

update-alternatives is usually used in managing multiple 
manpages.

I don't know if it's a overkill, but it's somewhat friendlier than documenting 
in README.Debian.



regards,
junichi.

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Gunnar Wolf]

Glenn Maynard dijo [Fri, Aug 29, 2003 at 04:03:16PM -0400]:
> If I report "segmentation fault in ls", I--as a user of ls, not a
> developer--couldn't care less about why it was segfaulting or how the
> bug was fixed; I only care that it's been fixed.  If a developer wants
> to spend their limited time researching how the bug was fixed and
> summarizing it in a changelog, great, but it's certainly not something I'd
> expect everyone to do.
> 
> (As a user, I'd certainly be rather annoyed at receiving duplicate close
> reports because someone reopened the bug for frivelous reasons, however.
> I get enough junk mail already.)

Ok, you could not care much about why was it broken and how was it fixed
- but then again, we have a lot of different users somehow different
from you, and I don't think you would be bothered by receiving this
extra information. Some users might find it even useful. Some fellow
developers might be interested in knowing how did you do it.

Greetings,

-- 
Gunnar Wolf - [EMAIL PROTECTED] - (+52-55)5630-9700 ext. 1366
PGP key 1024D/8BB527AF 2001-10-23
Fingerprint: 0C79 D2D1 2C4E 9CE4 5973  F800 D80E F35A 8BB5 27AF

Re: LWN subscription for Debian developers

[Dan Jacobson]

Bdale> ... I announced a group subscription to lwn.net for Debian
Bdale> developers, sponsored by HP...

Debian may be seen as supporting non-disclosure conditions /
protected proprietary information / trade secrets / etc. whatever.

Bdale> If you are a Debian developer and want full LWN access, go to
Bdale> lwn.net and create an account for yourself (no money is
Bdale> required...

But the Debian developer cannot disclose the information seen with non
Debian developers.

Re: Bug#195214: .iso conflict, discussion of resolution

[Matt Zimmerman]

On Sat, Aug 30, 2003 at 08:06:18PM -0500, Chris Cheney wrote:

> On Sat, Aug 30, 2003 at 06:08:17PM -0400, Matt Zimmerman wrote:
> > On Fri, Aug 29, 2003 at 10:34:39PM -0700, Mike Markley wrote:
> > 
> > > A quick summary of this bug:
> > > Arson, a KDE CD burning application, includes two .desktop files to
> > > associate certain files with it:
> > > /usr/share/mimelnk/application/x-iso.desktop
> > > /usr/share/mimelnk/application/x-cue.desktop
> > 
> > And presumably another application also wants to include these files?  Is
> > KDE's file association system really so broken that two programs cannot both
> > declare themselves able to handle a certain type of file?
> 
> Nope, mime types are completely independent of them declaring if they
> support it. That is why the mimetype really belongs in kdelibs itself
> and not the 3rd party package.

Ah, so these files are to define a MIME type, and not to associate files (as
they were described above).  That makes more sense.

-- 
 - mdz

Re: Bug#207300: tmda: Challenge-response is fundamentally broken

[John Hasler]

I wrote:
> This is about a quarter of my incoming mail.

Karsten writes:
> Which?  Bounces to spoofed senders, or improperly addressed mail?

Bounces.

> What prevents you from 550ing this at SMTP connect?

The absence of any such connections.  I'm on a dialup.
-- 
John Hasler
[EMAIL PROTECTED] (John Hasler)
Dancing Horse Hill
Elmwood, WI

Re: LWN subscription for Debian developers

[David B Harris]

On Sun, 31 Aug 2003 05:42:06 +0800
Dan Jacobson <[EMAIL PROTECTED]> wrote:
> Bdale> ... I announced a group subscription to lwn.net for Debian
> Bdale> developers, sponsored by HP...
> 
> Debian may be seen as supporting non-disclosure conditions /
> protected proprietary information / trade secrets / etc. whatever.
> 
> Bdale> If you are a Debian developer and want full LWN access, go to
> Bdale> lwn.net and create an account for yourself (no money is
> Bdale> required...
> 
> But the Debian developer cannot disclose the information seen with non
> Debian developers.

Jealous? :)


pgpKuaboT6Ya9.pgp
Description: PGP signature

Re: LWN subscription for Debian developers

[Lukas Geyer]

Dan Jacobson <[EMAIL PROTECTED]> writes:

> Bdale> ... I announced a group subscription to lwn.net for Debian
> Bdale> developers, sponsored by HP...
> 
> Debian may be seen as supporting non-disclosure conditions /
> protected proprietary information / trade secrets / etc. whatever.

What are you smoking? lwn.net makes articles available after one week,
they are only trying to get some money from subscriptions. I don't see
any relations to trade secrets and non-disclosure agreements, except
if you see all subscription-based services that way. But that would
also apply to people subscribing to newspapers, using the local
library (certain services only available for residents in public
libraries or university affiliates for academic libraries...) and tons
of other stuff.

> Bdale> If you are a Debian developer and want full LWN access, go to
> Bdale> lwn.net and create an account for yourself (no money is
> Bdale> required...
> 
> But the Debian developer cannot disclose the information seen with non
> Debian developers.

I don't think that is true. We cannot set up mirror sites, sure, but
if I share this information with some friends, I don't see how this
would violate anything.

Lukas

P.S.: I still would prefer lwn.net to be free. In fact, I would prefer
to get my NY Times without paying for it, but unfortunately my wishes
do not always come true...

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Junichi Uekawa]

> I certainly prefer it if the changelog tells how the bug was fixed.  This
> documents the difference between:
> 
>  * New upstream release
>- Removed the entire subsystem which contained this bug (Closes: #xxx)
> 
>  * New upstream release
>- Made the "foo" option create its file with sane permissions (Closes: 
> #xxx)

I think there are two different kinds of bug-closing scenarios; and IMO,
having the bug title and the submitter information would be benefitial,
to get it edited to something useful.

1. A patch submitted through BTS which is directly applied :

 * Updated translation for ja.po
 From: Junichi Uekawa (closes: #)
 * Fix behavior of SIGSEGV handling
 From: Junichi Uekawa (closes: #)

2. A user report that did not contain a patch or anything that helped 
track down the problem, but at least the problem got fixed:

 * debian/rules: Removed unnecessary checks for environmental friendliness
 fixes "Package does not build from source on my dual-opteron machine"
 From: Junichi Uekawa (closes: #)
 * New upstream release
 implements "ACPI methods for controlling teapots"
 From: Junichi Uekawa (closes: #)



For the second case, the maintainer needs to perform some manual editing
to obtain a resonable changelog entry, as you suggest.


Just a random thought.


regards,
junichi

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Herbert Xu]

Matt Zimmerman <[EMAIL PROTECTED]> wrote:
> 
> I list the submitter when they have provided a patch, so as to provide for
> attribution, and therefore credit or blame, as appropriate.

Well the credit should definitely be directed at the submitter.  The
blame however is squarely at the feet of the maintainer.

Cheers,
-- 
Debian GNU/Linux 3.0 is out! ( http://www.debian.org/ )
Email:  Herbert Xu ~{PmV>HI~} <[EMAIL PROTECTED]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Herbert Xu]

Gunnar Wolf <[EMAIL PROTECTED]> wrote:
> 
> Ok, you could not care much about why was it broken and how was it fixed
> - but then again, we have a lot of different users somehow different
> from you, and I don't think you would be bothered by receiving this
> extra information. Some users might find it even useful. Some fellow
> developers might be interested in knowing how did you do it.

For Debian changes, sure.

For upstream changes, definitely not.  This is what the upstream changelog
and failing that, the upstream source is for.

Listing random upstream changes in debian/changelog just because they
happen to fix bugs in the Debian BTS makes no sense.

Cheers,
-- 
Debian GNU/Linux 3.0 is out! ( http://www.debian.org/ )
Email:  Herbert Xu ~{PmV>HI~} <[EMAIL PROTECTED]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: Accepted kaffe 1:1.1.1-1 (i386 source)

[Matt Zimmerman]

On Sun, Aug 31, 2003 at 02:23:46PM +1000, Herbert Xu wrote:

> Matt Zimmerman <[EMAIL PROTECTED]> wrote:
> > 
> > I list the submitter when they have provided a patch, so as to provide for
> > attribution, and therefore credit or blame, as appropriate.
> 
> Well the credit should definitely be directed at the submitter.  The
> blame however is squarely at the feet of the maintainer.

This is not at all the case.  A large percentage of patches that I apply to
my packages are done more or less blindly.  Specifically, the message
translations.

-- 
 - mdz