Frank, On 22 May 2013, at 14:55, Frank Shearar <frank.shearar@gmail.com> wrote:
On 22 May 2013 13:44, Camillo Bruni <camillobruni@gmail.com> wrote:
On 2013-05-22, at 14:36, Frank Shearar <frank.shearar@gmail.com> wrote:
On 22 May 2013 13:20, Camillo Bruni <camillobruni@gmail.com> wrote:
On 2013-05-22, at 14:02, Frank Shearar <frank.shearar@gmail.com> wrote:
On 22 May 2013 12:49, stephane ducasse <stephane.ducasse@free.fr> wrote:
./pharo Pharo.image config filetree://`pwd`/../src/
Fails on trying to execute #mcRepositoryAsUser:withPassword: on a GenericUrl instance inside Gofer.
Yep, I'd agree with your assessment. Has Pharo forked Gofer, or does it still use Lukas Renggli's repository?
We forked it. We cannot built a part of our infrastructure on a project that can vanished in nature.
Well. What I should have asked was "have you cloned the repository with the intention of pushing improvements upstream, or have you forked the project with no intention of pushing changes upstream?"
well, check on how much we changed in detail and them come back.
Sorry, was that supposed to be an answer? I must go find out how much work you've done before guessing whether you've forked and abandoned upstream?
abandon or progress that's the question here.
That's a false dichotomy. There's a middle path, where you do the MC equivalent of a pull request. Now there's something useful to aim for: GitHub works so very, very well because it's utterly trivial to make upstream repositories know you have improvements lying around. Because fragmentation of your committer base _does not help you_ in the long run.
Now if the upstream maintainer _dies_, sure, that's a different story.
Like I said: What I want to know is this: if I want to make a change to infrastructure common to multiple Smalltalks, does Pharo expect me to submit special unique snowflake additions to Pharo AND to upstream?
well if you expect us to push anything upstream then have a look at the massive amount of changes that happened since pharo 1.0.
I don't care about Pharo 1.0 in the context of this discussion. I care about one particular package, one that should have no UI elements in it or similar nasty sources of entanglement. (I have attempted to port quite a few Pharo packages to Squeak. Lack of separation between UI and core functionality has been the only reason I've had to abandon a port, thus far.) And certainly Gofer as it's pulled in via Metacello (er, which one? The upstream one, not Pharo's fork of same) works fine in Pharo, Squeak and Gemstone, and wherever else Metacello and Gofer run. I should _not_ have to submit to more than one repository to add functionality to Gofer.
There IS a reason pharo forked from squeak...
I'm not pursuing this thread: it's my understanding that the fork was largely political, and personal, and that's as far as I'm going to talk about it.
Well, you are wrong about the reasons for the fork, IMO. But it is a done deal. (See my answer to Janko). Now, for library/framework developers/maintainers it is hard to be compatible with different Smalltalk implementations. It is not impossible, but requires manpower, interest and will. Recently, only Fuel managed to do that (of the bigger projects). Sven PS: I occasionally read the Squeak mailing and I see the work that you are doing and it is impressive. Your attitude, approach and drive remind me of the many Pharo developers. I would not dream of convincing you, but IMHO you would fit perfectly in the Pharo community, where you would find more interaction and feedback - maybe you would even reconsider using Smalltalk in business. -- Sven Van Caekenberghe http://stfx.eu Smalltalk is the Red Pill