Improving the SKS development model

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
25 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Improving the SKS development model

Yaron Minsky
As y'all have no doubt have noticed, I haven't had a ton of time to spend on SKS development of late.  But people on the list have started filling the gap, coming up with various patches to improve things.  I think we need to get a more organized development model to make it easier to put together releases and keep track of what's going on, and I think Jason's proposal is the right one: we should use mercurial for sharing development rather than just passing patches around.

If people would put up public hg repos containing their patches and emailed me when they thought they were ready for inclusion in the mainline, I would be happy to manage the process of pulling that in and of cutting releases.

So, what do people think?

A couple of other small changes I'm contemplating: start using google code as the hosting site for the main sks repository, and for other things like issue tracking and a wiki.  I've done the very basic setup, and you can find the site here: 

   http://code.google.com/p/sks-keyserver/

And I'm thinking of potentially moving the sks mailing list to google groups.  I'm also curious what people think about that.  I've already registered [hidden email] for that purpose.

y

_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Yaron Minsky
One other note: I finally cut a 1.1.1 release.  It is on the google code site.

y

On Wed, Aug 12, 2009 at 9:07 PM, Yaron Minsky <[hidden email]> wrote:
As y'all have no doubt have noticed, I haven't had a ton of time to spend on SKS development of late.  But people on the list have started filling the gap, coming up with various patches to improve things.  I think we need to get a more organized development model to make it easier to put together releases and keep track of what's going on, and I think Jason's proposal is the right one: we should use mercurial for sharing development rather than just passing patches around.

If people would put up public hg repos containing their patches and emailed me when they thought they were ready for inclusion in the mainline, I would be happy to manage the process of pulling that in and of cutting releases.

So, what do people think?

A couple of other small changes I'm contemplating: start using google code as the hosting site for the main sks repository, and for other things like issue tracking and a wiki.  I've done the very basic setup, and you can find the site here: 

   http://code.google.com/p/sks-keyserver/

And I'm thinking of potentially moving the sks mailing list to google groups.  I'm also curious what people think about that.  I've already registered [hidden email] for that purpose.

y


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

John Marshall-2
In reply to this post by Yaron Minsky
On Wed, 12 Aug 2009, 21:07 -0400, Yaron Minsky wrote:
> A couple of other small changes I'm contemplating: start using google code
> as the hosting site for the main sks repository, and for other things like
> issue tracking and a wiki.  I've done the very basic setup, and you can find
> the site here:
>
>    http://code.google.com/p/sks-keyserver/

I'm not a developer, but I think that it's a good idea for the
developers to share a central repository from which releases can be cut
and you've already done the legwork here.

> And I'm thinking of potentially moving the sks mailing list to google
> groups.  I'm also curious what people think about that.  I've already
> registered [hidden email] for that purpose.

Is there any need to move the mailing list from where it is?  I am not
in favour of mailing lists that encourage "rich" mail messages.  I don't
want to see list email full of HTML crud.

--
John Marshall

_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel

attachment0 (203 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Yaron Minsky
On Wed, Aug 12, 2009 at 10:06 PM, John Marshall <[hidden email]> wrote:
Is there any need to move the mailing list from where it is?  I am not
in favour of mailing lists that encourage "rich" mail messages.  I don't
want to see list email full of HTML crud.

I mostly want to move the list because savannah is an administrative disaster (I still can't get them to set up a damn redirect to my new website for sks), and I'd like high-quality list archives, which can not be had at savannah.

y

_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

John Marshall-2
On Wed, 12 Aug 2009, 22:09 -0400, Yaron Minsky wrote:

> On Wed, Aug 12, 2009 at 10:06 PM, John Marshall <
> [hidden email]> wrote:
>
> > Is there any need to move the mailing list from where it is?  I am not
> > in favour of mailing lists that encourage "rich" mail messages.  I don't
> > want to see list email full of HTML crud.
>
>
> I mostly want to move the list because savannah is an administrative
> disaster (I still can't get them to set up a damn redirect to my new website
> for sks), and I'd like high-quality list archives, which can not be had at
> savannah.
Fair enough.  Is it possible to configure a Google Groups mailing list
to be text only?  How much of one's life history (real or fake) does one
have to divulge to sign up to a Google Groups list?  At least with
Mailman all you need to give is an email address.  I refuse to use
Yahoo! groups but I haven't looked at Google groups.

--
John Marshall

_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel

attachment0 (203 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Yaron Minsky
Let's hold off a bit on the mailing list migration.  I'd like to focus first on getting the google code site in good shape...

y

On Wed, Aug 12, 2009 at 10:17 PM, John Marshall <[hidden email]> wrote:
On Wed, 12 Aug 2009, 22:09 -0400, Yaron Minsky wrote:
> On Wed, Aug 12, 2009 at 10:06 PM, John Marshall <
> [hidden email]> wrote:
>
> > Is there any need to move the mailing list from where it is?  I am not
> > in favour of mailing lists that encourage "rich" mail messages.  I don't
> > want to see list email full of HTML crud.
>
>
> I mostly want to move the list because savannah is an administrative
> disaster (I still can't get them to set up a damn redirect to my new website
> for sks), and I'd like high-quality list archives, which can not be had at
> savannah.

Fair enough.  Is it possible to configure a Google Groups mailing list
to be text only?  How much of one's life history (real or fake) does one
have to divulge to sign up to a Google Groups list?  At least with
Mailman all you need to give is an email address.  I refuse to use
Yahoo! groups but I haven't looked at Google groups.

--
John Marshall


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Phil Pennock-17
In reply to this post by John Marshall-2
On 2009-08-13 at 12:17 +1000, John Marshall wrote:

> On Wed, 12 Aug 2009, 22:09 -0400, Yaron Minsky wrote:
> > On Wed, Aug 12, 2009 at 10:06 PM, John Marshall <
> > [hidden email]> wrote:
> >
> > > Is there any need to move the mailing list from where it is?  I am not
> > > in favour of mailing lists that encourage "rich" mail messages.  I don't
> > > want to see list email full of HTML crud.
> >
> >
> > I mostly want to move the list because savannah is an administrative
> > disaster (I still can't get them to set up a damn redirect to my new website
> > for sks), and I'd like high-quality list archives, which can not be had at
> > savannah.
>
> Fair enough.  Is it possible to configure a Google Groups mailing list
> to be text only?  How much of one's life history (real or fake) does one
> have to divulge to sign up to a Google Groups list?  At least with
> Mailman all you need to give is an email address.  I refuse to use
> Yahoo! groups but I haven't looked at Google groups.

Disclaimer: speaking in a personal capacity.

It's a mailing-list.  It forwards on mails in the format in which they
were submitted.  Posting via the web UI generates plaintext emails -- I
just created a dummy list to double-check.

For a Google account, you need a name, a secondary email address (used
for password resets) and a country.

You do not need a Google account to join a Google Groups list; a list
manager can subscribe you directly, or send an invite, or you can follow
the advice at:
  http://onlygizmos.com/how-to-join-google-groups-without-a-google-id/2009/05/

Regards,
-Phil


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Kim Minh Kaplan
In reply to this post by Yaron Minsky
Yaron Minsky writes:

> If people would put up public hg repos containing their patches and emailed me
> when they thought they were ready for inclusion in the mainline, I would be
> happy to manage the process of pulling that in and of cutting releases.
>
> So, what do people think?

OK, I'll put up my patches on my public hg repository and post here when
it's done.

> A couple of other small changes I'm contemplating: start using google code as
> the hosting site for the main sks repository, and for other things like issue
> tracking and a wiki.  I've done the very basic setup, and you can find the site
> here: 
>
>    http://code.google.com/p/sks-keyserver/

If you go that route, it could be better if you reclaimed "sks" as a
project name.  Apparently it is blocked because sks is already a project
on Sourceforge but you seem to be its owner.

Kim Minh.


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Kim Minh Kaplan
In reply to this post by Yaron Minsky
Yaron Minsky writes:

> If people would put up public hg repos containing their patches and emailed me
> when they thought they were ready for inclusion in the mainline, I would be
> happy to manage the process of pulling that in and of cutting releases.

My repository (http://www.kim-minh.com/hg/sks/) now includes the recent
index patch (977e38781686) as well as a tail recursion fix
(c67b2f226c24).

Kim Minh.


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Re: Improving the SKS development model

John Clizbe-2
In reply to this post by Yaron Minsky
Yaron Minsky wrote:
> One other note: I finally cut a 1.1.1 release.  It is on the google code
> site.

What patches are included?

I didn't see any entries for 1.1.1 in the CHANGELOG

--
John P. Clizbe                      Inet:John (a) Mozilla-Enigmail.org
You can't spell fiasco without SCO. hkp://keyserver.gingerbear.net  or
     mailto:[hidden email]?subject=HELP

Q:"Just how do the residents of Haiku, Hawai'i hold conversations?"
A:"An odd melody / island voices on the winds / surplus of vowels"


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Yaron Minsky
In reply to this post by Kim Minh Kaplan
Wow!  You've been working hard.  Your repo contains 30-some patches consisting of 3k lines of diffs!  Are you proposing pushing all of that to the mainline tree?

y

On Thu, Aug 13, 2009 at 1:06 AM, Kim Minh Kaplan <[hidden email]> wrote:
Yaron Minsky writes:

If people would put up public hg repos containing their patches and emailed me when they thought they were ready for inclusion in the mainline, I would be happy to manage the process of pulling that in and of cutting releases.

My repository (http://www.kim-minh.com/hg/sks/) now includes the recent index patch (977e38781686) as well as a tail recursion fix (c67b2f226c24).

Kim Minh.


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Re: Improving the SKS development model

Yaron Minsky
In reply to this post by John Clizbe-2
I did not update the CHANGELOG properly.  I'll do that for the 1.1.2 release.

y

On Thu, Aug 13, 2009 at 3:52 AM, John Clizbe <[hidden email]> wrote:
Yaron Minsky wrote:
> One other note: I finally cut a 1.1.1 release.  It is on the google code
> site.

What patches are included?

I didn't see any entries for 1.1.1 in the CHANGELOG

--
John P. Clizbe                      Inet:John (a) Mozilla-Enigmail.org
You can't spell fiasco without SCO. hkp://keyserver.gingerbear.net  or
    mailto:[hidden email]?subject=HELP

Q:"Just how do the residents of Haiku, Hawai'i hold conversations?"
A:"An odd melody / island voices on the winds / surplus of vowels"


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Kim Minh Kaplan
In reply to this post by Yaron Minsky
Yaron Minsky writes:

> Wow!  You've been working hard.  Your repo contains 30-some patches consisting
> of 3k lines of diffs!  Are you proposing pushing all of that to the mainline
> tree?

No most of them are old commits in other branches that I did while
finding my way around mercurial and preparing the other patches that you
already included some months ago.  But I don't know enough about
Mercurial to hide them.

The only interesting patches are word index patch (977e38781686) as well
as a tail recursion fix (c67b2f226c24), the two last commits.  I still
have some other commits in a private repository but I do not thing they
should be published (except may be a reverse signature index but I have
to investigate more first).

On an other topic, I see that the preferred storage mecanism for keys is
as raw PGP files with modifications in the Berkeley DB.  Is there any
particular reason you chose not to dump all the keys in Berkeley DB and
then do without the dump directory?

I ask the question because it feels like over time the keys in the dump
directory will consist of many unused keys as they have been updated and
now leave in the Berkeley DB.

Kim Minh.


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Dinko Korunic
On Thu, Aug 13, 2009 at 05:40:04PM +0000, Kim Minh Kaplan wrote:
> as raw PGP files with modifications in the Berkeley DB.  Is there any
> particular reason you chose not to dump all the keys in Berkeley DB and
> then do without the dump directory?

I reckon this is because of initial import duration (build vs pbuild)
only. Personally, I am against using pbuild exactly for reasons you said.

--
NAME:Dinko.kreator.Korunic   DISCLAIMER:Standard.disclaimer.applies
ICQ:16965294        JAB:[hidden email]        PGP:0xEA160D0B
HOME:http://dkorunic.net    QUOTE:Eat.right.stay.fit.and.die.anyway


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Jason Harris
In reply to this post by Kim Minh Kaplan
On Thu, Aug 13, 2009 at 05:40:04PM +0000, Kim Minh Kaplan wrote:

> Yaron Minsky writes:
>
> > Wow!?? You've been working hard.?? Your repo contains 30-some patches consisting
> > of 3k lines of diffs!?? Are you proposing pushing all of that to the mainline
> > tree?
>
> No most of them are old commits in other branches that I did while
> finding my way around mercurial and preparing the other patches that you
> already included some months ago.  But I don't know enough about
> Mercurial to hide them.
>
> The only interesting patches are word index patch (977e38781686) as well
> as a tail recursion fix (c67b2f226c24), the two last commits.  I still
To keep unwanted history out of one's repo., specify the changeset(s)
when pulling from remote repos:

  %hg clone https://sks-keyserver.googlecode.com/hg/ sks
  %cd sks
  %hg pull http://www.kim-minh.com/hg/sks#977e38781686  <-- optional
  %hg pull http://www.kim-minh.com/hg/sks#c67b2f226c24
  ...
  %hg push

Then just continue to selectively import patches found via:

  %hg in http://www.kim-minh.com/hg/sks

Without the selective pull(s), you still arrive at version c67b2f226c24,
but also copy/preserve all of K-M's intermediate work, get the multiple
heads "problem," and clutter the commit log with entries you don't care
about (unless you use "hg log -f").

> have some other commits in a private repository but I do not thing they
> should be published (except may be a reverse signature index but I have
> to investigate more first).

Sounds cool, mind putting it in your repo. somewhere?
 
--
Jason Harris           |  NIC:  JH329, PGP:  This _is_ PGP-signed, isn't it?
[hidden email] _|_ web:  http://keyserver.kjsl.com/~jharris/
          Got photons?   (TM), (C) 2004

_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel

attachment0 (322 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Kim Minh Kaplan
In reply to this post by Dinko Korunic
Dinko Korunic writes:

> I reckon this is because of initial import duration (build vs pbuild)
> only. Personally, I am against using pbuild exactly for reasons you said.

Are you talking of build vs fastbuild?  You can not do without pbuild
for the recon process.  Note that one reason of the excessive slowness
of build is because of a strange tuning of database, namely pagesize.
By default it is set to 2Kbytes.  Changing to 8Kbytes yields better
results (most keys are above 1Kbytes in size and well below 2Kbytes).

Kim Minh


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Dinko Korunic
On Fri, Aug 14, 2009 at 06:45:53PM +0000, Kim Minh Kaplan wrote:
> > I reckon this is because of initial import duration (build vs pbuild)
> > only. Personally, I am against using pbuild exactly for reasons you said.
>
> Are you talking of build vs fastbuild?  You can not do without pbuild

Fastbuild, yes, my mistake. I appologize.

> for the recon process.  Note that one reason of the excessive slowness

OFC, yes.

> of build is because of a strange tuning of database, namely pagesize.
> By default it is set to 2Kbytes.  Changing to 8Kbytes yields better
> results (most keys are above 1Kbytes in size and well below 2Kbytes).

Good to know, thanks.

--
NAME:Dinko.kreator.Korunic   DISCLAIMER:Standard.disclaimer.applies
ICQ:16965294        JAB:[hidden email]        PGP:0xEA160D0B
HOME:http://dkorunic.net    QUOTE:Eat.right.stay.fit.and.die.anyway


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Yaron Minsky
In reply to this post by Kim Minh Kaplan
Sounds like another good patch!

On Fri, Aug 14, 2009 at 2:45 PM, Kim Minh Kaplan <[hidden email]> wrote:
Dinko Korunic writes:

> I reckon this is because of initial import duration (build vs pbuild)
> only. Personally, I am against using pbuild exactly for reasons you said.

Are you talking of build vs fastbuild?  You can not do without pbuild
for the recon process.  Note that one reason of the excessive slowness
of build is because of a strange tuning of database, namely pagesize.
By default it is set to 2Kbytes.  Changing to 8Kbytes yields better
results (most keys are above 1Kbytes in size and well below 2Kbytes).

Kim Minh


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Yaron Minsky
In reply to this post by Kim Minh Kaplan
On Thu, Aug 13, 2009 at 1:40 PM, Kim Minh Kaplan <[hidden email]> wrote:
The only interesting patches are word index patch (977e38781686) as well
as a tail recursion fix (c67b2f226c24), the two last commits.  I still
have some other commits in a private repository but I do not thing they
should be published (except may be a reverse signature index but I have
to investigate more first).

Neither of those revisions appear to be in the hg repo you mentioned:

 [yminsky@minskyprimus sks]$ hg log http://www.kim-minh.com/hg/sks/ -r 977e38781686
abort: unknown revision '977e38781686'!
[yminsky@minskyprimus sks]$ hg log http://www.kim-minh.com/hg/sks/ -r c67b2f226c24
abort: unknown revision 'c67b2f226c24'!

Are you sure the changes you're talking about are in there?

You might want to clone my main repo, apply your patches, and serve that repo up for me to pull from.  That way, you'll get rid of all the cruft that you don't care about anymore.

y

_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
Reply | Threaded
Open this post in threaded view
|

Re: Improving the SKS development model

Kim Minh Kaplan
Yaron Minsky writes:

> Neither of those revisions appear to be in the hg repo you mentioned:
>
>  [yminsky@minskyprimus sks]$ hg log http://www.kim-minh.com/hg/sks/ -r
> 977e38781686
> abort: unknown revision '977e38781686'!
> [yminsky@minskyprimus sks]$ hg log http://www.kim-minh.com/hg/sks/ -r
> c67b2f226c24
> abort: unknown revision 'c67b2f226c24'!
>
> Are you sure the changes you're talking about are in there?

Ouch, I was trying to cleanup the repository and overdid it a bit.  I
still have to get used tu Mercurial.  The patches are back in.

Kim Minh.


_______________________________________________
Sks-devel mailing list
[hidden email]
http://lists.nongnu.org/mailman/listinfo/sks-devel
12