Google Open Source Blog: Bidding farewell to Google Code

[Greg Stein]

Any thoughts on where to go?

http://google-opensource.blogspot.com/2015/03/farewell-to-google-code.html

[Justin Erenkrantz]

GitHub?  Or, back to ASF?

Either is fine by me.  -- justin

On Mar 12, 2015 12:37 PM, "Greg Stein" <gst...@gmail.com> wrote:

Any thoughts on where to go?

http://google-opensource.blogspot.com/2015/03/farewell-to-google-code.html

--
You received this message because you are subscribed to the Google Groups "Serf Development List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to serf-dev+u...@googlegroups.com.
To post to this group, send email to serf...@googlegroups.com.
Visit this group at http://groups.google.com/group/serf-dev.
For more options, visit https://groups.google.com/d/optout.

[Ivan Zhakov]

On 12 March 2015 at 22:37, Greg Stein <gst...@gmail.com> wrote:
> Any thoughts on where to go?
>
> http://google-opensource.blogspot.com/2015/03/farewell-to-google-code.html
>

I am in favor go to ASF if possible (if we're able to satisfy ASF
requirements for project):
1. Most (all ?) of serf committers are ASF committers or members, so
we're familiar with process, infrastructure etc.
2. ASF supports Subversion, downloads and mailing lists
3. ASF will be serving project's interested, while other project
hostings may change their policy in the same way as Google did

--
Ivan Zhakov

[Bert Huijben]

+1 on moving into ASF.

I'm just not sure how.
I don't think we should really spend all the time to start a new TLP here for such a small project, if we don't need to.
A subproject of APR or Subversion, or ???, with its own release schedule etc. would probably work good enough and avoid the administrative and incubation work for a new TLP.

+0.5 on moving to GitHub.

(I see we already have an old project page on Source forge?)

Bert

[Greg Stein]

On Thu, Mar 12, 2015 at 3:07 PM, Bert Huijben <be...@qqmail.nl> wrote:

> -----Original Message-----
> From: serf...@googlegroups.com [mailto:serf...@googlegroups.com] On
> Behalf Of Ivan Zhakov
> Sent: donderdag 12 maart 2015 21:02
> To: serf...@googlegroups.com; Greg Stein
> Subject: Re: [serf-dev] Google Open Source Blog: Bidding farewell to Google
> Code
>
> On 12 March 2015 at 22:37, Greg Stein <gst...@gmail.com> wrote:
> > Any thoughts on where to go?
> >
> > http://google-opensource.blogspot.com/2015/03/farewell-to-google-
> code.html
> >
> I am in favor go to ASF if possible (if we're able to satisfy ASF
> requirements for project):
> 1. Most (all ?) of serf committers are ASF committers or members, so
> we're familiar with process, infrastructure etc.
> 2. ASF supports Subversion, downloads and mailing lists
> 3. ASF will be serving project's interested, while other project
> hostings may change their policy in the same way as Google did

+1 on moving into ASF.

I'm just not sure how.
I don't think we should really spend all the time to start a new TLP here for such a small project, if we don't need to.

There is no "minimum size" for a TLP. It would be just fine for us to be a TLP, and the Board has been talking about accepting projects directly into a TLP (no incubation). Being a TLP means that we report once a quarter -- an hour for a writeup every three months isn't a burden.

 

A subproject of APR or Subversion, or ???, with its own release schedule etc. would probably work good enough and avoid the administrative and incubation work for a new TLP.

Apache OpenOffice also uses serf, for AOO's WebDAV features. ... but I think we just start as a TLP.

+0.5 on moving to GitHub.

I really do like GitHub's feature set, but not all that excited about switching to git.

 

(I see we already have an old project page on Source forge?)

Euh... no. Where are you seeing that???

Cheers,

-g

[Ivan Zhakov]

I meant TLP of course. Serf is separate project and should be remain
the same. Administration/incubation work should not be significant
given that most of us ASF committers/members. Otherwise we should just
merge serf to ra_serf, but I'm against about this.

--
Ivan Zhakov

[Bert Huijben]

This is what I found on Sourceforge, but it appears I didn’t look at the url yet:

 

http://sourceforge.net/projects/serf.mirror/

 

I think the last part already says enough. It looks like it is just a mirror of what used to be at google code somewhere in December 2013.

 

                Bert

 

--

[Greg Stein]

On Thu, Mar 12, 2015 at 3:02 PM, Ivan Zhakov <iv...@visualsvn.com> wrote:

On 12 March 2015 at 22:37, Greg Stein <gst...@gmail.com> wrote:
> Any thoughts on where to go?
>
> http://google-opensource.blogspot.com/2015/03/farewell-to-google-code.html
>
I am in favor go to ASF if possible (if we're able to satisfy ASF
requirements for project):

Easily satisfied. All committers for serf have been required to have an Apache ICLA on file with the ASF. ie. we've been abusing their Secretary and record-keeping :-P ... but it was done for exactly this purpose: a potential return to the ASF one day. (for those who didn't know, serf actually started at the ASF but moved out because it didn't have a community beyond me/Justin) 

Second, we'd start as a TLP with a half-dozen ASF Members (of our 11 committers, 10 are Members). I'm pretty positive the Board would sign off on that without any concerns, based on the recent "pTLP" discussions.

Cheers,

-g

[Ivan Zhakov]

That's great!

I added my disclaimer because I remember some concerns that serf
project will be able to satisfy three +1 policy for release.

--
Ivan Zhakov

[Justin Erenkrantz]

Part of the reason I have been quiet on-list is to foster ownership from others.  I am confident that the community is broader than just Greg and I now.  -- justin

[Lieven Govaerts]

On Thu, Mar 12, 2015 at 8:41 PM, Justin Erenkrantz
<jus...@erenkrantz.com> wrote:
> GitHub? Or, back to ASF?

The question is, what do we gain from going the ASF route?

I'd say, whatever route that helps us find the resources (or sponsors)
for a http 2.0 implementation is the one in the best interest of the
project.

If that's likely within ASF then fine by me, otherwise we might as
well go for the easiest hosting solution.

Lieven

[Ben Reser]

On 3/12/15 4:23 PM, Lieven Govaerts wrote:
> The question is, what do we gain from going the ASF route?

That's easy. Infrastructure supported by multiple corporate and individual
sponsors as opposed to infrastructure supported by a single entity that might
discontinue the infrastructure at any point.

A lot of people had that attitude early on that Google Code would never go away
because Google would always support it (not saying anyone here thought that).
I find the likelihood that Github is around essentially forever far less likely
than Google.

Google is winding this down fairly quickly, though with reasonable time for
people to migrate. I wouldn't count on that happening elsewhere. I hope for
the open source communities sake that it does, but I wouldn't bet on it. See
what happened with Twitpic for an example of how things can fall apart quickly.

With the ASF you have so many different projects that depend on the
infrastructure that I find it hard to believe that we'd have so little support
as to be able to keep it going.

This of course is not to say that the ASF is perfect. I just think long term
the foundation with support from many sources is superior to the charity of a
single organization.

> I'd say, whatever route that helps us find the resources (or sponsors)
> for a http 2.0 implementation is the one in the best interest of the
> project.

I don't think where the project is hosted has any impact on this.

[Stefan Sperling]

On Thu, Mar 12, 2015 at 02:37:28PM -0500, Greg Stein wrote:
> Any thoughts on where to go?
>
> http://google-opensource.blogspot.com/2015/03/farewell-to-google-code.html

+1 ASF

[Greg Stein]

ALL: please respond before Monday night. … the trend seems to be "return to ASF". Please respond to this specifically. The Board is meeting on Wednesday. If the project *does* want to go that direction, then I can put a resolution up for the Board. … Google Code is NOT going anyway "now". We can also wait a month or two. … and/or we can start the move and straddle old/new style releases. … we could release 1.4 old-style, or as an Apache project.

Please provide some guidance/thoughts over the next few days. No rush, but we *do* have an opportunity to move next week.

Cheers,

-g

[Lieven Govaerts]

On Fri, Mar 13, 2015 at 1:36 AM, Ben Reser <b...@reser.org> wrote:
> On 3/12/15 4:23 PM, Lieven Govaerts wrote:
>> The question is, what do we gain from going the ASF route?
>
> That's easy. Infrastructure supported by multiple corporate and individual
> sponsors as opposed to infrastructure supported by a single entity that might
> discontinue the infrastructure at any point.

That's it? That's our message to our fellow ASF'ers? "Serf wants to
join ASF because we trust you as the best hosting solution?"

I'm -0.5 in this.

>
> A lot of people had that attitude early on that Google Code would never go away
> because Google would always support it (not saying anyone here thought that).
> I find the likelihood that Github is around essentially forever far less likely
> than Google.
>
> Google is winding this down fairly quickly, though with reasonable time for
> people to migrate. I wouldn't count on that happening elsewhere. I hope for
> the open source communities sake that it does, but I wouldn't bet on it. See
> what happened with Twitpic for an example of how things can fall apart quickly.
>

We don't need a hosting provider that's around forever. If the open
source communities around the world migrate to the next cool hosting
solution in 5 years or so, and serf is still actively maintained at
that time, we can do the tiny effort to move to something else as
well.

> With the ASF you have so many different projects that depend on the
> infrastructure that I find it hard to believe that we'd have so little support
> as to be able to keep it going.
>
> This of course is not to say that the ASF is perfect. I just think long term
> the foundation with support from many sources is superior to the charity of a
> single organization.
>
>> I'd say, whatever route that helps us find the resources (or sponsors)
>> for a http 2.0 implementation is the one in the best interest of the
>> project.
>
> I don't think where the project is hosted has any impact on this.
>

Exactly.

If this project wants to be relevant in a couple of years for anyone
else besides Subversion, it needs to have HTTP 2.0 support.

My hope is that either:
- someone finds the time to integrate an existing library (e.g.
MIT-licensed nghttp) in serf.
- we can pool resources with another project to write a ASLv2-licensed
HTTP 2.0 client library

A move to ASF would make sense if we actively work on the second
option: talk with the httpd/apr guys (i.e. you guys) to see how we can
work together on this, in order to give the serf project a future (at
least for the next 5 years).

If we are only looking for a hosting solution - which is what I
understand the majority here wants, from your reply and all the quick
"+1 for ASF" replies I saw earlier - I propose we go to GitHub for the
better integrated tooling and without the (for us unneeded) ASF
overhead.

Lieven

[Jim Jagielski]

The ASF needs, imo, projects that really understand and grok the Apache Way. Serf would be a shining example of that, ala Subversion and httpd, within the ASF.

[Ivan Zhakov]

On 13 March 2015 at 13:24, Greg Stein <gst...@gmail.com> wrote:
> ALL: please respond before Monday night. … the trend seems to be "return to
> ASF". Please respond to this specifically.

Just to confirm my opinion: +1 ASF.

--
Ivan Zhakov

[Greg Stein]

On Fri, Mar 13, 2015 at 5:42 AM, Lieven Govaerts <l...@mobsol.be> wrote:

>...

If this project wants to be relevant in a couple of years for anyone
else besides Subversion, it needs to have HTTP 2.0 support.

Agreed. … and I believe we can all agree our hosting solution will not solve this problem. That it is orthogonal.

 

My hope is that either:
- someone finds the time to integrate an existing library (e.g.
MIT-licensed nghttp) in serf.
- we can pool resources with another project to write a ASLv2-licensed
HTTP 2.0 client library

A move to ASF would make sense if we actively work on the second
option: talk with the httpd/apr guys (i.e. you guys) to see how we can
work together on this, in order to give the serf project a future (at
least for the next 5 years).

httpd is likely our best friends for HTTP/2. That's really outside the bounds of APR. (and also looking at activity levels on both)

 

If we are only looking for a hosting solution - which is what I
understand the majority here wants, from your reply and all the quick
"+1 for ASF" replies I saw earlier - I propose we go to GitHub for the
better integrated tooling and without the (for us unneeded) ASF
overhead.

I've already said that I like the GitHub tooling, but dislike the use of git. So not opposed, but hope for something else.

….

Just to be clear: this community can *always* choose to exit the ASF. We did it once before. None of our options laid out so far (GitHub, ASF) are permanent.

Are there other SVN hosting providers to consider? I use Assembla for some private svn hosting. They are also friends of Subversion.

Does Bitbucket provide svn hosting?

Cheers,

-g

[Jim Jagielski]

On Fri, Mar 13, 2015 at 7:10 AM, Greg Stein <gst...@gmail.com> wrote:

httpd is likely our best friends for HTTP/2. That's really outside the bounds of APR. (and also looking at activity levels on both)

 

++1 

[C. Michael Pilato]

+1 to return to ASF, FWIW from but an occasional committer.

[Justin Erenkrantz]

+1 for returning to ASF.  -- justin 

[Bert Huijben]

> -----Original Message-----
> From: serf...@googlegroups.com [mailto:serf...@googlegroups.com] On
> Behalf Of Lieven Govaerts
> Sent: vrijdag 13 maart 2015 11:42
> To: serf...@googlegroups.com
> Subject: Re: [serf-dev] Google Open Source Blog: Bidding farewell to Google
> Code
>

I'm probably able to spend time on this, but as long as Subversion's http server (Apache Httpd) doesn't have HTTP 2.0 support, I don't think it is worth investing now.

But that really makes it a "chicken and egg problem"...


I'm not sure how this all fits the httpd design. In most cases a thread or worker is responsible for a connection today.

Perhaps we should couple this to trying to get Subversion support working as fastcgi process, as that would give the option to test against many more servers, some of which already have an HTTP 2.0 implementation.

One of the reasons for mentioning this here is that it might be a good idea to implement the fastcgi protocol in serf (or something serf-like) for that.
(Just like there was some discussion of using serf as http server a few years ago)

> A move to ASF would make sense if we actively work on the second
> option: talk with the httpd/apr guys (i.e. you guys) to see how we can
> work together on this, in order to give the serf project a future (at
> least for the next 5 years).

Bert

>
> If we are only looking for a hosting solution - which is what I
> understand the majority here wants, from your reply and all the quick
> "+1 for ASF" replies I saw earlier - I propose we go to GitHub for the
> better integrated tooling and without the (for us unneeded) ASF
> overhead.
>
> Lieven
>

[Lieven Govaerts]

On Fri, Mar 13, 2015 at 12:10 PM, Greg Stein <gst...@gmail.com> wrote:
> On Fri, Mar 13, 2015 at 5:42 AM, Lieven Govaerts <l...@mobsol.be> wrote:
>>...
>>
>> If this project wants to be relevant in a couple of years for anyone
>> else besides Subversion, it needs to have HTTP 2.0 support.
>
>
> Agreed. … and I believe we can all agree our hosting solution will not solve
> this problem. That it is orthogonal.
>

My hope was to use this occasion to discuss how we see the future of
serf. Based on the outcome of that discussion we should be able to
decide if a simple hosting provider or the ASF best matches our needs.

Given that most serf devs seem to want a quick transfer to the ASF,
I'll join with a "+0 ASF", but I'm still wondering what we expect more
from ASF than just reliable hosting.

Lieven

[Greg Stein]

On Tue, Mar 17, 2015 at 5:59 AM, Lieven Govaerts <l...@mobsol.be> wrote:

>...

Given that most serf devs seem to want a quick transfer to the ASF,

I'll join with a "+0 ASF", but I'm still wondering what we expect more
from ASF than just reliable hosting.

Honestly, I expect a little bit of "closeness" to Apache httpd, where we can coordinate with them on http/2 development. But we could do that without being part of the Foundation.

Second would be the Foundation's legal umbrella. It is a remote point.

But... third, and (IMO) most important: we don't have to switch to git. ... I think GitHub is awesome. Truly. But it really means using git. I don't like that for day-to-day. It's a hard tradeoff. To get the utility of GitHub, at the cost of daily work.

....

We have more to discuss. There is no rush, as Google Code is not disappearing soon. We don't have to push a Resolution before the ASF Board in two days.

...

So: we know we can operate as a successful community. Maybe that means hosting is a commodity for us. Given that most of us are ASF Members, we have a tendency to land there. Do we have a rationale? Or is it a comfortable default?

(I know that git is a not-comfortable default)

...

We'll be successful wherever we go (yay, community!). But with our tooling dropping out from under us, which toolset do we shift to?

And no... I'm not drawing this out into a three-week-long lovey-dovey koombaya. Lieven has great questions, and I'd like to respectfully find some answers.

Cheers,

-g

[Stefan Sperling]

On Tue, Mar 17, 2015 at 06:41:36AM -0500, Greg Stein wrote:
> We'll be successful wherever we go (yay, community!). But with our tooling
> dropping out from under us, which toolset do we shift to?

I know the ASF is about more than just tooling, but I wouldn't dismiss
the ASF as an option even if this was just a tooling question.
Because my personal experience with ASF infra has been great. We know
the people there and they already work with us on a daily basis.
The trust we have built there would need to be built from scratch
with any other organisation.

So I don't see any reason to prefer some arbitrary hosting company's
tooling over ASF. I'd expect the ASF to survive e.g. github's lifetime by
any margin because it is run by a strong community rather than for profit.
Perhaps I'm wrong on that point, but well, consider why we're having this
discussion in the first place...

[Lieven Govaerts]

So,

google code goes read-only in two weeks. Thanks for the early warning Greg ;)

From this thread it's clear that almost all serf devs prefer to move
to ASF, or at least have no objections against such a move.

We are now 5 months further, is this still the preferred way to go? If
yes, how shall we / who wants to proceed?

regards,

Lieven

On Thu, Mar 12, 2015 at 8:37 PM, Greg Stein <gst...@gmail.com> wrote:
> Any thoughts on where to go?
>
> http://google-opensource.blogspot.com/2015/03/farewell-to-google-code.html
>

[Justin Erenkrantz]

+1 to going back to ASF.  -- justin

[Greg Stein]

Woah! Didn't realize it was coming up "now".

Some of us chatted about moving to the ASF in April, during ApacheCon, and yeah: ASF was still everybody's choice. Let's let this stew for a few days, in case anybody offers another choice.

In a separate thread, I'll put a draft of the resolution that we would submit to the Board. We need a volunteer for VP. I'll do it, but maybe somebody else wants to try their hand? I believe we have 11 committers last time I checked, and 10 are ASF Members. The Board should not even blink at that. I'll list all on the resolution, ping each, and can remove as needed (some people became a committer for a patch or two, and may not wanna be on the PMC).

Any options other than ASF?

Cheers,

-g

[Bert Huijben]

+1 for ASF.

Let me know if you need any help, to get things handled.

Bert

---

From: Greg Stein
Sent: ‎12-‎8-‎2015 23:27

To: serf...@googlegroups.com
Subject: Re: [serf-dev] Google Open Source Blog: Bidding farewell to GoogleCode

[Greg Stein]

Easy. Volunteer for VP ;-)

[be...@qqmail.nl]

LOL.

 

But if non of the original core serf developers would like to volunteer… sure.

 

I would guess it would take you a tiny bit less time than it would take me, given your involvement & experience with other projects… but I would like to try.

 

Bert

 

From: Greg Stein
Sent: donderdag 13 augustus 2015 00:42
To: serf...@googlegroups.com
Subject: Re: [serf-dev] Google Open Source Blog: Bidding farewell toGoogleCode

 

 

Easy. Volunteer for VP ;-)

 

On Wed, Aug 12, 2015 at 5:41 PM, Bert Huijben <be...@qqmail.nl> wrote:

+1 for ASF.

Let me know if you need any help, to get things handled.

Bert

[Jim Jagielski]

+1

--
Jim Jagielski
Brief? Mobile

[Lieven Govaerts]

On Thu, Aug 13, 2015 at 2:01 AM, <be...@qqmail.nl> wrote:

LOL.

 

But if non of the original core serf developers would like to volunteer… sure. 

Me, ​only if no one else would want to take the role.​

 

I would guess it would take you a tiny bit less time than it would take me, given your involvement & experience with other projects… but I would like to try.

​Yay, the Dutchman for VP! :)

Lieven