Plans to change to Git?

classic Classic list List threaded Threaded
50 messages Options
123
Reply | Threaded
Open this post in threaded view
|

Re: Plans to change to Git?

laurence taylor
Nobody has mentioned git-svn I don't think.

On Thu, Sep 1, 2011 at 7:59 PM, Marcelo de Moraes Serpa <[hidden email]> wrote:
> It's true that GitHub helps people to keep their own forked version of
> the compiler up-to-date, but I don't think that's a good thing
> community-wise to have people having their own modified version of haXe,
> which might prevent them from reusing/sharing compatible code with other
haXe 
developers.

Wow! I'm surprised you think like this. HaXe is already open source and there *are* modified versions of the compiler lying around. *The official version will always be the official version*. People can judge and choose their best option, and new contributions will eventually be backported to the official one.

Fact is, SVN is the big elephant in the room. It's just ugly, slow and awkward. It works, yes, but is far from optimal if you compare it with git (or even bazaar and HG).

I'm surprised how closed-minded some developers can be. Take some time to read and try it, Start using it, and you'll eventually make the switch. 

Another thing: The switch to git shouldn't be painful. You can pretty much use *the same workflow* you use for SVN. That's how I learned it.

- Marcelo.




2011/9/1 Heinz Hölzer <[hidden email]>
Am 01.09.2011 04:22, schrieb Nicolas Juneau:

it's not only happy land with unicorns and rainbows
awesome metaphor ;)

--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: Plans to change to Git?

Nicolas Juneau
In reply to this post by Marcelo de Moraes Serpa
The forks issue can be good or bad depending on how you see it. If forks are
born out of dispute or disconnect in the community (which can lead to
community fragmentation), maybe it's not such a good outcome if it could have
been worked out. If the forks are made to experiment or to fix problems that
cannot be adressed by the core team, it can be a good thing. Like you
mentionned, the forks can be later integrated into the main tree.

I may be wrong here, but I don't think it's a matter of being close minded.
Git hasn't been ruled out completely - it's being discussed. It's not like
it's our decision anyway. The switch _should_ not be painful (in a perfect
world), should such a switch occur. That's in theory. In practice, it can be.
In the end, should I ever have to contribute to haXe's code, I'd like the VCS
to be Git, but if it's not, then it's because it has been judged that
maintaining the current VCS is a better way to go for the moment.

Le 1 septembre 2011 13:59:14, Marcelo de Moraes Serpa a écrit :

> > It's true that GitHub helps people to keep their own forked version of
> > the compiler up-to-date, but I don't think that's a good thing
> > community-wise to have people having their own modified version of haXe,
> > which might prevent them from reusing/sharing compatible code with other
> > haXe developers.
>
> Wow! I'm surprised you think like this. HaXe is already open source and
> there *are* modified versions of the compiler lying around. *The official
> version will always be the official version*. People can judge and choose
> their best option, and new contributions will eventually be backported to
> the official one.
>
> Fact is, SVN is the big elephant in the room. It's just ugly, slow and
> awkward. It works, yes, but is far from optimal if you compare it with git
> (or even bazaar and HG).
>
> I'm surprised how closed-minded some developers can be. Take some time to
> read and try it, Start using it, and you'll eventually make the switch.
>
> Another thing: The switch to git shouldn't be painful. You can pretty much
> use *the same workflow* you use for SVN. That's how I learned it.
>
> - Marcelo.
>
>
>
>
> 2011/9/1 Heinz Hölzer <[hidden email]>
>
> > Am 01.09.2011 04:22, schrieb Nicolas Juneau:
> >  it's not only happy land with unicorns and rainbows
> >
> > awesome metaphor ;)
> >
> > --
> > haXe - an open source web programming language
> > http://haxe.org
--
Nicolas Juneau

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: Plans to change to Git?

Johann Borck
In reply to this post by Marcelo de Moraes Serpa
On 09/01/2011 07:59 PM, Marcelo de Moraes Serpa wrote:
[...]
I'm surprised how closed-minded some developers can be. Take some time to read and try it, Start using it, and you'll eventually make the switch. 

Yes, sure, but why don't you just put up a git repo somewhere, keep it up to date, manage experimental and feature branches, pull requests, and all that stuff, keep track of who does what and create nice reports about it, take community polls on the features, and finally offer patches based on all that? And over time you automate some of those workflows, your setup is working and everything is fine... only everytime someone shows up and calls you closed-minded you throw it all away and start over, 'cause hey that's just an impeccable argument, isn't it?

regards,
Johann



--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: Plans to change to Git?

Marcelo de Moraes Serpa
If someone wants to fork haXe, they can do right now. That's protectionism. This mindset is not compatible with open source.  Anyway, this is turning into a flame-war. Fact is, git is way better than SVN, no one could argue about that ;)

On Thu, Sep 1, 2011 at 1:40 PM, Johann Borck <[hidden email]> wrote:
On 09/01/2011 07:59 PM, Marcelo de Moraes Serpa wrote:
[...]
I'm surprised how closed-minded some developers can be. Take some time to read and try it, Start using it, and you'll eventually make the switch. 

Yes, sure, but why don't you just put up a git repo somewhere, keep it up to date, manage experimental and feature branches, pull requests, and all that stuff, keep track of who does what and create nice reports about it, take community polls on the features, and finally offer patches based on all that? And over time you automate some of those workflows, your setup is working and everything is fine... only everytime someone shows up and calls you closed-minded you throw it all away and start over, 'cause hey that's just an impeccable argument, isn't it?

regards,
Johann



--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: Plans to change to Git?

danielku15
In reply to this post by laurence taylor
laurence taylor wrote
Nobody has mentioned git-svn I don't think.
That's probably like working with Office 2010 (docx) and only save to Office 2003 Documents (doc). Of course it works, but you'll never can use the "awesomeness" (thanks barney for this word xD) of the new features.


Each SVN user should have tried out git once. Once you've tried it you'll love it and won't switch back to SVN.  Believe me.

And in addition to the discussion: It's not only about to change from SVN to Git. It's also about to move from Google Code to Github. Also with Github you get a lot of community influence you won't get on Google Code. While Google Code is some kind of my-private-project-in-public hosting, Github is more a social network for developers. You easily find new developers and the platform simply motivates you to follow and contribute to projects. All of you should give it a try and check it out. ;)
Reply | Threaded
Open this post in threaded view
|

Re: Plans to change to Git?

Simon Asselbergs
In reply to this post by Marcelo de Moraes Serpa
Hi List,

I also am surprised this forks or about repositories for that matter.

Forks are always useful, maybe not from the perspective of those who prefer the original. But forks too are also the result of open source projects, offer choice (choice is always good), a sign of evolution of a concept and also help innovation. So if the majority of the original community feels confident about the how contributions of other community members are dealt with, the direction the project is taken to (etc) the fear for damage because of people forking the project is unnecessary and unrealistic. There is always chance someone will fork so it will happen. I would rather call that rather innovation than the law of Murphy in effect. For if there wasn't the ability to fork e.g. OpenOffice or Mambo would be abandoned without offering an (innovative) alternative, thus less to choose from.

If someone - in disregard of the reason - forks a project nobody can stop her/him for doing that as well (if there are no license issues of course). To expect that nobody will fork, or exorcizing forks doesn't work isn't realistic or sustainable. I think it is more productive to always strive to improve leadership, the effectiveness dealing with proposed contributions (not only being sensitive rational but also the tone) to forward an open source project. If the latter doesn't get enough attention in any project than that is the real reason projects get abandoned, not the other way around: not forks. Forking is an importing feature of git. The popularity of git didn't kill any open source project because of using it so it seems. In fact quite some tend to use it to encourage it innovation and depend of people forking. And that doesn't seem to hurt either. Despite of well known discussion regarding Linus not scaling (and maybe SVN also doesn't :p ), using his git does.

Simon

On Thu, Sep 1, 2011 at 7:59 PM, Marcelo de Moraes Serpa <[hidden email]> wrote:
> It's true that GitHub helps people to keep their own forked version of
> the compiler up-to-date, but I don't think that's a good thing
> community-wise to have people having their own modified version of haXe,
> which might prevent them from reusing/sharing compatible code with other
haXe 
developers.

Wow! I'm surprised you think like this. HaXe is already open source and there *are* modified versions of the compiler lying around. *The official version will always be the official version*. People can judge and choose their best option, and new contributions will eventually be backported to the official one.

Fact is, SVN is the big elephant in the room. It's just ugly, slow and awkward. It works, yes, but is far from optimal if you compare it with git (or even bazaar and HG).

I'm surprised how closed-minded some developers can be. Take some time to read and try it, Start using it, and you'll eventually make the switch. 

Another thing: The switch to git shouldn't be painful. You can pretty much use *the same workflow* you use for SVN. That's how I learned it.

- Marcelo.




2011/9/1 Heinz Hölzer <[hidden email]>
Am 01.09.2011 04:22, schrieb Nicolas Juneau:

it's not only happy land with unicorns and rainbows
awesome metaphor ;)

--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: Plans to change to Git?

Nicolas Cannasse
Le 02/09/2011 14:56, Simon Asselbergs a écrit :

> Hi List,
>
> I also am surprised this forks or about repositories for that matter.
>
> Forks are always useful, maybe not from the perspective of those who
> prefer the original. But forks too are also the result of open source
> projects, offer choice (choice is always good), a sign of evolution of a
> concept and also help innovation. So if the majority of the original
> community feels confident about the how contributions of other community
> members are dealt with, the direction the project is taken to (etc) the
> fear for damage because of people forking the project is unnecessary and
> unrealistic. There is always chance someone will fork so it will happen.

I'm not afraid of forks, and I think there's definitely a lot of cases
where they make senses (when the software is no longer updated or when
you want to push it a very specific direction).

At the same time, I think that's hurting the community as a whole. Since
people forking are often good developers, there would contribute more by
running the main version than having their own incompatible version.

Forking does not hurt that much for huge communities, but haXe is not
there yet so I prefer if it doesn't happen.

That doesn't have that much to do with Git BTW, I was just replying to
the previous poster which was pointing that Git was helping forking,
which I didn't see as a good point.

Nicolas

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: Plans to change to Git?

Edward Middleton
On 09/03/2011 01:20 AM, Nicolas Cannasse wrote:

> I'm not afraid of forks, and I think there's definitely a lot of cases
> where they make senses (when the software is no longer updated or when
> you want to push it a very specific direction).
>
> At the same time, I think that's hurting the community as a whole. Since
> people forking are often good developers, there would contribute more by
> running the main version than having their own incompatible version.
>
> Forking does not hurt that much for huge communities, but haXe is not
> there yet so I prefer if it doesn't happen.
>
> That doesn't have that much to do with Git BTW, I was just replying to
> the previous poster which was pointing that Git was helping forking,
> which I didn't see as a good point.

With git everyone is a forker.  Thats the model git takes.  In general
it doesn't hurt because very few people actually want to own a project,
and they tend to be the ones that are followed.

I think the bigger risk for a project is that people don't push back
good ideas and code rather then the community diverging.

Most of the contributions I have made to projects came about because I
needed something and someone else (often the maintainer) was interested
in using it.  Other times people saw how I had solved a problem and
suggested or implemented a better approach.

If the project was on a centralized VCS I probably wouldn't post my
changes because it might be seen as an attempt to usurp the one true
code base, but with something like github thats just the way its done.

Edward

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: Plans to change to Git?

Simon Asselbergs
In reply to this post by Nicolas Cannasse
I'm not afraid of forks, and I think there's definitely a lot of cases where they make senses (when the software is no longer updated or when you want to push it a very specific direction).

At the same time, I think that's hurting the community as a whole. Since people forking are often good developers, there would contribute more by running the main version than having their own incompatible version.

Those with most control on the project better start figuring out why good developers in the haxe community are forking maybe you can indeed try to find common ground.. I would be surprised if these reasons would be "not being updated", because that is very clearly not the case. Nicolas and other clearly are very busy updating it. As far as the direction of the project I get the strong notion you are the strongest driver Nicolas.

The project has great value I think and is fairly well known amongst "hardcore" actionscripters. I don't know how large the community is, but if it is a bit too small (like you are saying) than it comes to me at a surprise? Why should this project have a small community?

How adopted is Ocaml (I have no idea). I understand it is a language very much fitting a project like haXe. Being cross platform, performant and typically a well suited tool for for (byte)code generation, most probably. If it's a niche language then it doesn't come as a surprise that it is harder to find people who want to pickup the source code of haXe and provide patches etc. If so then it's a bit of a pitty the optimal used technology happens to be a niche.

Forking does not hurt that much for huge communities, but haXe is not there yet so I prefer if it doesn't happen.

Haxe will definitely not move in the direction of Functional Programming. It makes sense haXe isn't moving into that direction as functional programming is a niche, but that niche is where my interest lie. I am very willing to learn Ocaml just to start forking haXe. Ocaml is also very interesting to learn as a general purpose language since it supports functional, imperative, and object-oriented programming styles. I want more a kind-of-Scala-looking-a-bit-like-haxe then the other way around. But the haxe code generation targets now are exactly what I need, haxe is leaner than Scala and also I like Ocaml. As I am busy trying to contribute to Stax (a FP haXe lib) I constantly hit walls, because haXe doesn't fit well with modern FP. Implementing FP with haxe means hacky boilerplate code, very logical because haXe isn't designed for this kind of stuff. But I would still like to compliment for having haxe, although it's doesn't fit me enough it is a very impressive compiler+language.

Greetings

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: Plans to change to Git?

Chris Ochs
In reply to this post by Nicolas Cannasse
My bet is that you will see more people contribute to haxe by changing
to git.  I hate having to work with svn on open source projects.  The
whole worklfow is extremely cumbersome compared to git.

Git makes it easy to work with open source projects.  Svn gets in your way.

Chris

--
haXe - an open source web programming language
http://haxe.org
123