I really appreciate getting quite timely reviews, and getting my patches merged to trunk fairly quickly. That's at least as important as the mechanics and you're getting it more right than many projects do.
I do agree it's fairly hard to know how to contribute changes. For instance, within one review thread, one person asked for patches as attachments and someone else asked for them inline. Some documentation says to send patches and some says to send a pull request. There are some code reviews on logilab and maybe some on bitbucket? There is some mention of hg mutable history, which sounds interesting, but also sounds pretty immature and probably not something drive-by contributors would want to have to learn. Even suggesting mq may be too much, and if a project wants sanitized history, I think the core maintainers are probably best placed to do that themselves. I think the key thing is to have a consistent, simple, story for basic contributions. On Fri, Sep 21, 2012 at 6:00 PM, afayolle <afayolle...@free.fr> wrote: > Hello, > > Just a quick note, I've read > http://jakevdp.github.com/blog/2012/09/20/why-python-is-the-last/ The > article is interesting in its opinion about Python in the scientific > computation world. But I found the graph in the "Github" section very > interesting, especially since I'm now in the position of an external > contributor. > > To make a long story short, contributing to Logilab's projets from the > outside is not easy as in could be in these days of github / bitbucket > and all. IMO, if the goal is to gain external contributors to projects > such as pylint (+ astng + lgc) or Cubicweb, then some effort should be > put in setting an official bitbucket repository from which forking a > branch would be easy (it's already not very complicated to fork from > hg.logilab.org) and more important, making a pull request when a new > feature is implemented would be a breeze. > > The recent discussions on this list make it quite obvious that, although > the current process may work fine for insiders, it is very painful for > outsiders. It is fine with me if tickets are not tracked on bitbucket. I > can live with code reviews happening on logilab.org too. But please, > make it a breeze for us to publish our changes on bitbucket and notify > you of these. When contributing on spare time to an free software > project, having to spend precious time adapting to the idiosyncrasies of > the organization in charge for managing the releases of the software > (which is definitely work for which we are all grateful) rather than > hacking can be a deterrent, especially if: > > * the process is not visible (I know about > http://www.logilab.org/card/contributing, but the link on the home page > is not very visible, and not displayed at all if I am logged in on the > site) > > * the process adds steps which could be left out (and indeed are left > out in other projects), namely the required use of mq. > > I'd like to hear some feedback on this from Logilab folks, of course, > and also from other external contributors. > > Regards, > > -- > Alexandre > > > _______________________________________________ > Python-Projects mailing list > Python-Projects@lists.logilab.org > http://lists.logilab.org/mailman/listinfo/python-projects > -- Martin
_______________________________________________ Python-Projects mailing list Python-Projects@lists.logilab.org http://lists.logilab.org/mailman/listinfo/python-projects
