Re: [gentoo-dev] Bugzilla 4 migration

2011-03-08 Thread Fabian Groffen
On 07-03-2011 17:25:02 -0500, Mike Frysinger wrote:
  As outsider, I don't like to accept another certificate thing, just to
  view a bugtracker.
 
 if we're only forcing *login*, then this isnt an issue

+1


-- 
Fabian Groffen
Gentoo on a different level



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-08 Thread Donnie Berkholz
On 07:50 Tue 08 Mar , Hans de Graaff wrote:
 On Mon, 2011-03-07 at 08:13 -0600, Donnie Berkholz wrote:
 
  Thanks! One thing I've been very interested about in 3.x and 4.x is API 
  access that's better than screen-scraping. I tried using the 
  python-bugzilla client that accesses Bugzilla via XML-RPC but it didn't 
  seem to work. Do we have anything available?
 
 I've tried an ipad application that uses xmlrpc and that seemed to work
 fine.

Confirmed with my iphone one. Guess the Python one's broken with BZ 4. 
Fiddling around manually with xmlrpclib works alright, too.

-- 
Thanks,
Donnie

Donnie Berkholz
Sr. Developer, Gentoo Linux
Blog: http://dberkholz.com


pgpN0Hq8SDQ9s.pgp
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-08 Thread Michał Górny
On Mon, 07 Mar 2011 15:06:25 -0500
Olivier Crête tes...@gentoo.org wrote:

 On Mon, 2011-03-07 at 20:47 +0100, Michał Górny wrote:
  Why does everyone assume it needs to be enforced? If user is
  interested in protecting his/her data, he/she can simply use
  https://. If he/she is not, there is no real reason to enforce
  slower (and not always supported) SSL.
 
 Maybe it's not to protect the user, but to protect the Gentoo
 infrastructure.. And really, SSL has been supported by every browser
 for the last 15 years. And it is not in any way slow or slower than
 non-SSL.

If you really think you need to force all users to use SSL, thus
assuming they're unable to make their own decisions, why don't you
restrict bugzie access completely?

-- 
Best regards,
Michał Górny


signature.asc
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-08 Thread Antoni Grzymała

On Tue, 8 Mar 2011 15:26:34 +0100, Michał Górny wrote:

On Mon, 07 Mar 2011 15:06:25 -0500
Olivier Crête tes...@gentoo.org wrote:


On Mon, 2011-03-07 at 20:47 +0100, Michał Górny wrote:
 Why does everyone assume it needs to be enforced? If user is
 interested in protecting his/her data, he/she can simply use
 https://. If he/she is not, there is no real reason to enforce
 slower (and not always supported) SSL.

Maybe it's not to protect the user, but to protect the Gentoo
infrastructure.. And really, SSL has been supported by every browser
for the last 15 years. And it is not in any way slow or slower than
non-SSL.


If you really think you need to force all users to use SSL, thus
assuming they're unable to make their own decisions, why don't you
restrict bugzie access completely?


Michał,

You don't seem to (or pretend not to) understand that using SSL 
protects not *the user* (in which case, yes, a user is free to leave the 
door to *his own* house wide open), but the Gentoo infrastructure that 
is far from his own and that all of us are using. Besides, complaining 
about SSL being slow is absurd considering how mildly interactive and 
how low-traffic a typical bugzilla session is. You could do just fine 
over a 9600 bps modem.


Regards,

Antoni



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-08 Thread Michał Górny
On Tue, 08 Mar 2011 16:41:08 +0200
Antoni Grzymała awa...@chopin.edu.pl wrote:

  On Tue, 8 Mar 2011 15:26:34 +0100, Michał Górny wrote:
  On Mon, 07 Mar 2011 15:06:25 -0500
  Olivier Crête tes...@gentoo.org wrote:
 
  On Mon, 2011-03-07 at 20:47 +0100, Michał Górny wrote:
   Why does everyone assume it needs to be enforced? If user is
   interested in protecting his/her data, he/she can simply use
   https://. If he/she is not, there is no real reason to enforce
   slower (and not always supported) SSL.
 
  Maybe it's not to protect the user, but to protect the Gentoo
  infrastructure.. And really, SSL has been supported by every
  browser for the last 15 years. And it is not in any way slow or
  slower than non-SSL.
 
  If you really think you need to force all users to use SSL, thus
  assuming they're unable to make their own decisions, why don't you
  restrict bugzie access completely?
 
  You don't seem to (or pretend not to) understand that using SSL 
  protects not *the user* (in which case, yes, a user is free to leave
 the door to *his own* house wide open), but the Gentoo infrastructure
 that is far from his own and that all of us are using.

Please explain to me how not using SSL for a particular bugzie user is
going to hurt Gentoo infra. Even if we're talking about a dev,
and we're really assuming a dev is completely unaware of security
issues he/she's dealing with, I'd say power outage could cause more
damage.

 Besides, complaining about SSL being slow is absurd considering how
 mildly interactive and how low-traffic a typical bugzilla session is.
 You could do just fine over a 9600 bps modem.

It is more absurd to waste 5 minutes trying to establish login session
due to packet loss.

-- 
Best regards,
Michał Górny


signature.asc
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-08 Thread Nathan Phillip Brink
On Tue, Mar 08, 2011 at 03:53:01PM +0100, Micha?? G??rny wrote:
 On Tue, 08 Mar 2011 16:41:08 +0200
 Antoni Grzyma??a awa...@chopin.edu.pl wrote:
 
   On Tue, 8 Mar 2011 15:26:34 +0100, Micha? Grny wrote:
   On Mon, 07 Mar 2011 15:06:25 -0500
   Olivier Cr??te tes...@gentoo.org wrote:
  
   On Mon, 2011-03-07 at 20:47 +0100, Micha?? G??rny wrote:
Why does everyone assume it needs to be enforced? If user is
interested in protecting his/her data, he/she can simply use
https://. If he/she is not, there is no real reason to enforce
slower (and not always supported) SSL.
  
   Maybe it's not to protect the user, but to protect the Gentoo
   infrastructure.. And really, SSL has been supported by every
   browser for the last 15 years. And it is not in any way slow or
   slower than non-SSL.
  
   If you really think you need to force all users to use SSL, thus
   assuming they're unable to make their own decisions, why don't you
   restrict bugzie access completely?
  
   You don't seem to (or pretend not to) understand that using SSL 
   protects not *the user* (in which case, yes, a user is free to leave
  the door to *his own* house wide open), but the Gentoo infrastructure
  that is far from his own and that all of us are using.
 
 Please explain to me how not using SSL for a particular bugzie user is
 going to hurt Gentoo infra. Even if we're talking about a dev,
 and we're really assuming a dev is completely unaware of security
 issues he/she's dealing with, I'd say power outage could cause more
 damage.

If you access a bug which a user marked private/for devs only, or some
security bug, then the process of you viewing this information without
SSL would disclose this information to anyone listening on your
network. And disclosing your session cookie would allow anyone to find
any such private data they _want_ to find rather than just the content
you're viewing. Thus, by encrypting everything you are protecting
Gentoo users' data which is posted as private on bugzilla because they
trust that ``private'' actually means private.

  Besides, complaining about SSL being slow is absurd considering how
  mildly interactive and how low-traffic a typical bugzilla session is.
  You could do just fine over a 9600 bps modem.
 
 It is more absurd to waste 5 minutes trying to establish login session
 due to packet loss.

And if you have such a bad internet connection as you claim to have,
then perhaps there's a higher chance of people trolling your packets
anyways :-p.

-- 
binki

Look out for missing apostrophes!


pgpgNGileIJ7j.pgp
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Michał Górny
On Sun, 06 Mar 2011 23:55:31 +0100
Christian Ruppert id...@gentoo.org wrote:


 SSL is enabled by default now, so it's forced. Unfortunately the
 option to force SSL *only* for logged in user is no longer available
 in Bugzilla-4.x. It has been added in early 3.x AFAIR and later
 replaced by forcing SSL at all or not.
 If *anybody* can't use SSL for any reason please yell so that we can
 decide if we leave it as it is (plain + encrypted) or not.

Is there any *real* reason to force SSL? It is *hell* slow.

-- 
Best regards,
Michał Górny


signature.asc
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Dirkjan Ochtman
On Mon, Mar 7, 2011 at 10:12, Michał Górny mgo...@gentoo.org wrote:
 Is there any *real* reason to force SSL? It is *hell* slow.

Do you mean that SSL is slow or that bugs is slow? I also noticed that
Bugzilla is very slow right now, but it seems unlikely that it's due
to SSL.

Cheers,

Dirkjan



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Mike Frysinger
On Mon, Mar 7, 2011 at 4:12 AM, Michał Górny wrote:
 On Sun, 06 Mar 2011 23:55:31 +0100 Christian Ruppert wrote:
 SSL is enabled by default now, so it's forced. Unfortunately the
 option to force SSL *only* for logged in user is no longer available
 in Bugzilla-4.x. It has been added in early 3.x AFAIR and later
 replaced by forcing SSL at all or not.
 If *anybody* can't use SSL for any reason please yell so that we can
 decide if we leave it as it is (plain + encrypted) or not.

 Is there any *real* reason to force SSL? It is *hell* slow.

it should of course be force for logging in
-mike



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Michał Górny
On Mon, 7 Mar 2011 10:24:33 +0100
Dirkjan Ochtman d...@gentoo.org wrote:

 On Mon, Mar 7, 2011 at 10:12, Michał Górny mgo...@gentoo.org wrote:
  Is there any *real* reason to force SSL? It is *hell* slow.
 
 Do you mean that SSL is slow or that bugs is slow? I also noticed that
 Bugzilla is very slow right now, but it seems unlikely that it's due
 to SSL.

Both but yep, unrelated. I am just a personal SSL-forced-everywhere
hater.

-- 
Best regards,
Michał Górny


signature.asc
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Robin H. Johnson
On Mon, Mar 07, 2011 at 10:12:14AM +0100, Michał Górny wrote:
 On Sun, 06 Mar 2011 23:55:31 +0100
 Christian Ruppert id...@gentoo.org wrote:
  SSL is enabled by default now, so it's forced. Unfortunately the
  option to force SSL *only* for logged in user is no longer available
  in Bugzilla-4.x. It has been added in early 3.x AFAIR and later
  replaced by forcing SSL at all or not.
  If *anybody* can't use SSL for any reason please yell so that we can
  decide if we leave it as it is (plain + encrypted) or not.
 Is there any *real* reason to force SSL? It is *hell* slow.
The SSL forcing is temporarily disabled until we trace down why it's
causing slowness. Tracking is in bug 357711.

-- 
Robin Hugh Johnson
Gentoo Linux: Developer, Trustee  Infrastructure Lead
E-Mail : robb...@gentoo.org
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Robin H. Johnson
On Sun, Mar 06, 2011 at 11:55:31PM +0100, Christian Ruppert wrote:
 our Bugzilla (bugs.gentoo.org) will be unavailable for the next hours.
 We're going to migrate our old Bugzilla to Bugzilla-4.
 We expect our update to finish within the next hours.
All completed now. If you run into any problems, please file a new bug
under the Bugzilla product.

I'm sending this email as idl0r went to be after 8+ hours straight of
working on the new Bugzilla setup.

We apologize for this taking so extremely long.

Things didn't go so well at the database layer [1], and that hugely
increased the migration time. 

The Gentoo for the Bugzilla service went perfectly, a huge thanks to
idl0r for the years of work he has put into them.

 Some notes:
 SSL is enabled by default now, so it's forced.
The forcing is temporarily disabled now until we fix a possible related
performance issue, per bug 357711.

Footnotes:
[1] We discovered a potential MySQL bug with replication, where the
slaves end up truncating mediumtext columns to 1024 characters when done
with REPLACE INTO and GROUP_CONCAT. Will pursue with upstream this week.
This was only noted with mk-table-checksum, and we decided to just redo
replication from the master that was used for introducing the schema
changes. This added 3.5 hours onto the end of the migration :-(.

-- 
Robin Hugh Johnson
Gentoo Linux: Developer, Trustee  Infrastructure Lead
E-Mail : robb...@gentoo.org
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85


pgpWQInMnvLWm.pgp
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread justin
On 07/03/11 10:51, Robin H. Johnson wrote:
 The Gentoo for the Bugzilla service went perfectly, a huge thanks to
 idl0r for the years of work he has put into them.
 

Thanks for all your work on this.



signature.asc
Description: OpenPGP digital signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Jorge Manuel B. S. Vicetto
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On 07-03-2011 08:51, Robin H. Johnson wrote:
 On Sun, Mar 06, 2011 at 11:55:31PM +0100, Christian Ruppert wrote:
 our Bugzilla (bugs.gentoo.org) will be unavailable for the next hours.
 We're going to migrate our old Bugzilla to Bugzilla-4.
 We expect our update to finish within the next hours.
 All completed now. If you run into any problems, please file a new bug
 under the Bugzilla product.

Thank you both for all the work in the upgrade and for all the
maintenance work you've been doing for years on our Bugzilla.

- -- 
Regards,

Jorge Vicetto (jmbsvicetto) - jmbsvicetto at gentoo dot org
Gentoo- forums / Userrel / Devrel / KDE / Elections / RelEng
-BEGIN PGP SIGNATURE-
Version: GnuPG v2.0.17 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
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=vJNK
-END PGP SIGNATURE-



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Markos Chandras
On Sun, Mar 06, 2011 at 11:55:31PM +0100, Christian Ruppert wrote:
 Dear community,
 
 our Bugzilla (bugs.gentoo.org) will be unavailable for the next hours.
 We're going to migrate our old Bugzilla to Bugzilla-4.
 We expect our update to finish within the next hours.
 
 Some notes:
 SSL is enabled by default now, so it's forced. Unfortunately the
 option to force SSL *only* for logged in user is no longer available in
 Bugzilla-4.x. It has been added in early 3.x AFAIR and later replaced by
 forcing SSL at all or not.
 If *anybody* can't use SSL for any reason please yell so that we can
 decide if we leave it as it is (plain + encrypted) or not.
 
 All custom/Gentoo patches will be available *later* in a git repo[1].
 So if you'd like to fix something or improve the theme you can
 contribute patches.
 Thanks to Alex Legler (a3li) for the Bugzilla theme.
 
 [1]
 http://git.overlays.gentoo.org/gitweb/?p=proj/gentoo-bugzilla.git;a=summary
 
 -- 
 Regards,
 Christian Ruppert
 Role: Gentoo Linux developer, Bugzilla administrator and Infrastructure
 member
 Fingerprint: EEB1 C341 7C84 B274 6C59  F243 5EAB 0C62 B427 ABC8
 

Thank you very much. New bugzie looks pretty :)


Regards,
-- 
Markos Chandras / Gentoo Linux Developer / Key ID: B4AFF2C2


pgpqdtUPg7103.pgp
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Donnie Berkholz
On 09:51 Mon 07 Mar , Robin H. Johnson wrote:
 The Gentoo for the Bugzilla service went perfectly, a huge thanks to
 idl0r for the years of work he has put into them.

Thanks! One thing I've been very interested about in 3.x and 4.x is API 
access that's better than screen-scraping. I tried using the 
python-bugzilla client that accesses Bugzilla via XML-RPC but it didn't 
seem to work. Do we have anything available?

-- 
Thanks,
Donnie

Donnie Berkholz
Sr. Developer, Gentoo Linux
Blog: http://dberkholz.com


pgpThu2dFRGyd.pgp
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Tobias Klausmann
Hi! 

On Mon, 07 Mar 2011, Mike Frysinger wrote:
  If *anybody* can't use SSL for any reason please yell so that we can
  decide if we leave it as it is (plain + encrypted) or not.
 
  Is there any *real* reason to force SSL? It is *hell* slow.
 
 it should of course be force for logging in

If it is enforced for login, it should be enforced for logged
in sessions, cf. Cookie stealing (for a POC: Firesheep). And no,
restricting the login cookie to an IP is *not* safe enough.

Regards,
Tobias

-- 
Sent from aboard the Culture ship
GSV Zero Gravitas



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Dane Smith
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On 03/07/2011 09:48 AM, Tobias Klausmann wrote:
 Hi! 
 
 On Mon, 07 Mar 2011, Mike Frysinger wrote:
 If *anybody* can't use SSL for any reason please yell so that we can
 decide if we leave it as it is (plain + encrypted) or not.

 Is there any *real* reason to force SSL? It is *hell* slow.

 it should of course be force for logging in
 
 If it is enforced for login, it should be enforced for logged
 in sessions, cf. Cookie stealing (for a POC: Firesheep). And no,
 restricting the login cookie to an IP is *not* safe enough.
 
 Regards,
 Tobias
 

First off, a big thanks to infra and all involved in the migration. It
looks awesome!

As to the SSL bit, there is *no* reason not to be using SSL for anything
that requires a username / password. And I 100% agree with Tobias. If
it's necessary to use SSL to login, it's necessary to use it for the
duration of the session. I don't know how feasible it is to do, but if
normal viewing (no login) can be left SSL free, I see no issue there.
Otherwise however, SSL should be in use.

Regards,
- -- 
Dane Smith (c1pher)
Gentoo Linux Developer -- QA / Crypto / Sunrise / x86
RSA Key: http://pgp.mit.edu:11371/pks/lookup?search=0x0C2E1531op=index
-BEGIN PGP SIGNATURE-
Version: GnuPG v2.0.17 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
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=CxR9
-END PGP SIGNATURE-



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Mike Frysinger
On Mon, Mar 7, 2011 at 9:48 AM, Tobias Klausmann wrote:
 On Mon, 07 Mar 2011, Mike Frysinger wrote:
  If *anybody* can't use SSL for any reason please yell so that we can
  decide if we leave it as it is (plain + encrypted) or not.
 
  Is there any *real* reason to force SSL? It is *hell* slow.

 it should of course be force for logging in

 If it is enforced for login, it should be enforced for logged
 in sessions, cf. Cookie stealing (for a POC: Firesheep). And no,
 restricting the login cookie to an IP is *not* safe enough.

you're talking about two different things.  imo it's more important to
protect the credentials than spoofing/replay attacks.  the former is a
no brainer while the latter is fine to leave to the discretion of the
end user.
-mike



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Donnie Berkholz
On 16:35 Mon 07 Mar , Dirkjan Ochtman wrote:
 On Mon, Mar 7, 2011 at 15:13, Donnie Berkholz dberkh...@gentoo.org wrote:
  Thanks! One thing I've been very interested about in 3.x and 4.x is API
  access that's better than screen-scraping. I tried using the
  python-bugzilla client that accesses Bugzilla via XML-RPC but it didn't
  seem to work. Do we have anything available?
 
 Is that the one you get if you emerge pybugz?

No, pybugz is a screen-scraper. We previously had Bugzilla 2 so we 
couldn't do anything else.

 The Mozilla guys made a pretty nice REST API that can be installed as a
 plugin, I think. Maybe we could run that?

I've been somewhat following that too, but I don't know if anyone's 
written a CLI client for it yet, whereas python-bugzilla already exists 
(and has an ebuild in the sabayon overlay).

-- 
Thanks,
Donnie

Donnie Berkholz
Sr. Developer, Gentoo Linux
Blog: http://dberkholz.com


pgpXcCnIenCNR.pgp
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Michał Górny
On Mon, 7 Mar 2011 15:48:19 +0100
Tobias Klausmann klaus...@gentoo.org wrote:

 On Mon, 07 Mar 2011, Mike Frysinger wrote:
   If *anybody* can't use SSL for any reason please yell so that we
   can decide if we leave it as it is (plain + encrypted) or not.
  
   Is there any *real* reason to force SSL? It is *hell* slow.
  
  it should of course be force for logging in
 
 If it is enforced for login, it should be enforced for logged
 in sessions, cf. Cookie stealing (for a POC: Firesheep). And no,
 restricting the login cookie to an IP is *not* safe enough.

Why does everyone assume it needs to be enforced? If user is interested
in protecting his/her data, he/she can simply use https://. If he/she
is not, there is no real reason to enforce slower (and not always
supported) SSL.

It's like forcing everyone to have doors with semi-automatic locks.

-- 
Best regards,
Michał Górny


signature.asc
Description: PGP signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Christian Ruppert
On 03/07/2011 08:47 PM, Michał Górny wrote:
 On Mon, 7 Mar 2011 15:48:19 +0100
 Tobias Klausmann klaus...@gentoo.org wrote:
 
 On Mon, 07 Mar 2011, Mike Frysinger wrote:
 If *anybody* can't use SSL for any reason please yell so that we
 can decide if we leave it as it is (plain + encrypted) or not.

 Is there any *real* reason to force SSL? It is *hell* slow.

 it should of course be force for logging in

 If it is enforced for login, it should be enforced for logged
 in sessions, cf. Cookie stealing (for a POC: Firesheep). And no,
 restricting the login cookie to an IP is *not* safe enough.
 
 Why does everyone assume it needs to be enforced? If user is interested
 in protecting his/her data, he/she can simply use https://. If he/she
 is not, there is no real reason to enforce slower (and not always
 supported) SSL.
 
 It's like forcing everyone to have doors with semi-automatic locks.
 

*I* think it's ok if we're going to protect *our* data. Some user may
even benefit from it.
I don't see any disadvantages for our users.

-- 
Regards,
Christian Ruppert
Role: Gentoo Linux developer, Bugzilla administrator and Infrastructure
member
Fingerprint: EEB1 C341 7C84 B274 6C59  F243 5EAB 0C62 B427 ABC8



signature.asc
Description: OpenPGP digital signature


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Olivier Crête
On Mon, 2011-03-07 at 20:47 +0100, Michał Górny wrote:
 On Mon, 7 Mar 2011 15:48:19 +0100
 Tobias Klausmann klaus...@gentoo.org wrote:
 
  On Mon, 07 Mar 2011, Mike Frysinger wrote:
If *anybody* can't use SSL for any reason please yell so that we
can decide if we leave it as it is (plain + encrypted) or not.
   
Is there any *real* reason to force SSL? It is *hell* slow.
   
   it should of course be force for logging in
  
  If it is enforced for login, it should be enforced for logged
  in sessions, cf. Cookie stealing (for a POC: Firesheep). And no,
  restricting the login cookie to an IP is *not* safe enough.
 
 Why does everyone assume it needs to be enforced? If user is interested
 in protecting his/her data, he/she can simply use https://. If he/she
 is not, there is no real reason to enforce slower (and not always
 supported) SSL.

Maybe it's not to protect the user, but to protect the Gentoo
infrastructure.. And really, SSL has been supported by every browser for
the last 15 years. And it is not in any way slow or slower than non-SSL.


-- 
Olivier Crête
tes...@gentoo.org
Gentoo Developer


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


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Fabian Groffen
On 07-03-2011 15:06:25 -0500, Olivier Crête wrote:
 Maybe it's not to protect the user, but to protect the Gentoo
 infrastructure.. And really, SSL has been supported by every browser for
 the last 15 years. And it is not in any way slow or slower than non-SSL.

but the certificate security click-through-couple-of-times before you
can access bugzilla is sort of annoying

As outsider, I don't like to accept another certificate thing, just to
view a bugtracker.


-- 
Fabian Groffen
Gentoo on a different level



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Rich Freeman
On Mon, Mar 7, 2011 at 4:32 PM, Fabian Groffen grob...@gentoo.org wrote:
 As outsider, I don't like to accept another certificate thing, just to
 view a bugtracker.

When you think about it, this is a defect with your browser, and not
so much with SSL itself.

Your browser generally doesn't complain about unauthenticated
connections.  It accepts unauthenticated connections that aren't
encrypted without any issues, despite these being completely open to
numerous attacks.  However, your browser does complain when it makes
an unauthenticated connection that IS encrypted, even though this is
vulnerable to far fewer attacks.

Browsers shouldn't bug the user about self-signed certificates - they
should simply and clearly show that the user is connected to a host
that isn't authenticated by a trusted intermediate.

Oh, and browsers shouldn't come with root certs pre-installed by the
browser distributor either, but that is about as likely to get fixed
as the problem I just described.

In any case, I don't see poor browser design as a valid reason for
avoiding the use of SSL...

Rich



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Fabian Groffen
On 07-03-2011 16:52:23 -0500, Rich Freeman wrote:
 In any case, I don't see poor browser design as a valid reason for
 avoiding the use of SSL...

Please use a MUA that properly honours Reply-To: headers.  I'm on the
list.


-- 
Fabian Groffen
Gentoo on a different level



Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Mike Frysinger
On Monday, March 07, 2011 16:59:22 Fabian Groffen wrote:
 On 07-03-2011 16:52:23 -0500, Rich Freeman wrote:
  In any case, I don't see poor browser design as a valid reason for
  avoiding the use of SSL...
 
 Please use a MUA that properly honours Reply-To: headers.  I'm on the
 list.

subscribed != receiving.  there's no way of knowing who is.  get over it.
-mike


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


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Mike Frysinger
On Monday, March 07, 2011 16:32:55 Fabian Groffen wrote:
 On 07-03-2011 15:06:25 -0500, Olivier Crête wrote:
  Maybe it's not to protect the user, but to protect the Gentoo
  infrastructure.. And really, SSL has been supported by every browser for
  the last 15 years. And it is not in any way slow or slower than non-SSL.
 
 but the certificate security click-through-couple-of-times before you
 can access bugzilla is sort of annoying

i heard rumors the cacert is finally going into firefox ...

 As outsider, I don't like to accept another certificate thing, just to
 view a bugtracker.

if we're only forcing *login*, then this isnt an issue
-mike


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


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-07 Thread Hans de Graaff
On Mon, 2011-03-07 at 08:13 -0600, Donnie Berkholz wrote:

 Thanks! One thing I've been very interested about in 3.x and 4.x is API 
 access that's better than screen-scraping. I tried using the 
 python-bugzilla client that accesses Bugzilla via XML-RPC but it didn't 
 seem to work. Do we have anything available?

I've tried an ipad application that uses xmlrpc and that seemed to work
fine.

Kind regards,

Hans


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


Re: [gentoo-dev] Bugzilla 4 migration

2011-03-06 Thread Christian Ruppert
On 03/07/2011 01:00 AM, Jan Kundrát wrote:
 On 03/06/11 23:55, Christian Ruppert wrote:
 our Bugzilla (bugs.gentoo.org) will be unavailable for the next hours.
 We're going to migrate our old Bugzilla to Bugzilla-4.
 We expect our update to finish within the next hours.
 
 (Private reply, as I don't feel like flaming you in public. Feel free to
 re-send to a public list, or quote, preferably as a whole.)
 
 Hi Christian,
 I wanted to ask if I missed the announcement of the migration. I tried
 to imagine a case which would force people to go ahead and perform such
 an action without informing the world about the downtime in advance, but
 failed to find one. So, did I miss the announcement, or was that a lapse
 on some guy's side, or is it something else which warranted a swift
 action? Anyway, I'm looking forward to a nice, upgraded bugzie.
 
 Hm, so before sending this mail, I checked my gentoo-dev archive, and
 the first e-mail about Bugzilla migration is roughly 12 hours old.
 That's very different from how Infra has handled any other migration in
 the past (apart from dealing with unexpected emergencies, of course). I
 realize I'm in the armchair position in this case, but this looks like a
 rather dangerous move. When you add the workflow change, I wouldn't
 stick with an announcement 12 hours in advance myself (I do sysadmin
 stuff as a day job).
 
 With kind regards,
 Jan
 

Hey Jan,

I know I didn't announce it properly. It was *my* fault but this is
also a special case IMO.
We decided to migrate just a few hours ago because robbat2 and me having
enough time to do it now, finally. We're waiting since about 2007 for
Bugzilla upgrades and it's now 2011 so I thought it's ok to do it now
instead of waiting another few months (probably) or longer until we both
have enough time again etc.

We're not going to change the workflow, at least not now. We only do
that if you guys decided about it.

-- 
Regards,
Christian Ruppert
Role: Gentoo Linux developer, Bugzilla administrator and Infrastructure
member
Fingerprint: EEB1 C341 7C84 B274 6C59  F243 5EAB 0C62 B427 ABC8



signature.asc
Description: OpenPGP digital signature