On Thu, Nov 27, 2014 at 3:37 PM, Esteban Lorenzano <estebanlm@gmail.com> wrote:

On 27 Nov 2014, at 14:43, phil@highoctane.be wrote:

At times, mczs still come handy for some merges...

And Smalltalkhub is good as a safe heaven for collecting packages that are otherwise scattered all around.

nothing that a real catalog/centralised package manager alla npm/apt-get/etc cannot do it.��
using a package manager as a catalog feels to me like hammering a screw.

Sure, ��I am with you on that. I am yum - ing all day long :-)

But as you know, the Pharo Catalog descriptions are quite lonely.

I'd like to take a moment to tackle that one.

In fact, there are several concerns that I find (and other members on the list share the feeling) must be addressed for commercial development.��

e.g.: Stronger RDBMS support, AAA, Logs, ...

But this not really Pharo per se, but a onion ring around it.

As you guys are doing 4.0, that onion ring is making 3.0 work for commercial stuff.

I think that we'll always be one version behind, which is ok.

Is there a reason why Smalltalkhub would not stay working nicely?

because we do not have the strength/willing to maintain it.��
and each day we are: farther from ��state of the art�� source management, and farther from state of the art javascript, etc. (which is the reason why sthub��

Ok, thx.

��

Are we talking about using bigger boxes here? Or is it a deeper issue?

keeping up-to-date a system like sthub is a lot of work, and takes a lot of time.��
personally, I prefer way more to spend my time in things that will have a real impact in the community (like having a modern vm) than lose it trying constantly to catch up with what others (with a lot more resources) already did.��

Sure, modern vm anytime!

The git worklflow is superpowerful and that's what people use these days.

In a MOOC I do, all exercises and slideware are in Github...

��

each time I start a sub-project, my first question is: ��this stuff will have a multiplier effect in the community?�� and second question is ��do we have to doit from scratch, or can we take advantage of other projects?��. As a maintainer, and being conscious of our limitations, this are the driving forces I find positive to work.
(and of course, most times I do not start sub-projects at all, I just jump into a burning place and try to do my best to fix it�� not always very successfully :P)

Nah, you rule. I wish I was as good as you are.

��

so, coming back to less ��philosophical�� question:

1) do we need a state-of-the-art source code management? YES.
2) do we want to spend the few manpower we have on running into a worst solution of what is already around? I think no.��

libgit2 would provide that without going through Github driver hoops, which is sweet and more welcoming to newcomers.

��

Esteban

Phil

On Thu, Nov 27, 2014 at 1:52 PM, kilon alios <kilon.alios@gmail.com> wrote:
I forgot to add that git comes with excellent gui clients that are far more powerful and elegant that what Pharo offers currently .��

If you are user of emacs there is magit , really powerful gui client and very popular among emacs users.

For gui client I have used quite a lot SmartGit��

http://www.syntevo.com/smartgit/

and recently a fellow python developer introduced me to Sourcetree

http://www.sourcetreeapp.com/

Both are free for non commercial projects. They require a license for commercial use but they are relative cheap. They come with diff tools, easy commit access , branching, merging and tons of stuff to make life easier for complex scenarios and they integrate well with bitbucket and other online repositories besides github.��

But even from command line there is a lot of room for automation by creating bash scripts to make commits one step process.��