From: Jeff A. <ja...@fa...> - 2017-10-25 20:00:36
|
Just playing this back to get it straight ... the idea would be to publish the devguide as a document at readthedocs (thanks for the access), and link it, and so much else, from a website that develops in this repo: https://github.com/jython/jython.github.io . There are help pages about it, I know, but as a living example, it would be like the PyInstaller site where https://github.com/pyinstaller/pyinstaller.github.io is instantiated as http://www.pyinstaller.org/ This sounds good to me, and lets us share this kind of work more easily. Jeff Jeff Allen On 23/10/2017 19:19, Jim Baker wrote: > Some top posting first: > > We should just use read the docs; and point our CNAME for jython.org > <http://jython.org> (and related domains) to it > (http://docs.readthedocs.io/en/latest/alternate_domains.html). > Currently there are too many bottlenecks to getting stuff done in > Jython, and that's super frustrating. Automation can resolve some of > these bottlenecks. Also I don't want to be one of the bottlenecks > myself!!! Did I say I wanted to work on docs earlier this summer? I > did work a bit on it privately, but then it got buried under my > general pile of stuff. This should not be stopping anyone else working > on this issue however. > > Anyway... > > I have an existing RTD project, > https://readthedocs.org/projects/jython/ Rather than having it point > to the Jython book, it might make more sense to have be the overall > project or at least to https://github.com/jython/docs (not there > yet!). See subproject support: > http://docs.readthedocs.io/en/latest/subprojects.html > > Jeff, and others here, I can add you to the jython project on RTD as > admins. Just need your username. > > There may be better approaches, but nothing is as simple as RTD given > the source is reStructuredText. Also IIRC, RTD does allow some degree > of customization of the look, although I personally would be so much > happier with fresh and current content, vs any customization! > > > On Mon, Oct 23, 2017 at 1:40 AM, Jeff Allen <ja...@fa... > <mailto:ja...@fa...>> wrote: > > I've finished my preliminary restructuring of the developers' > guide at https://github.com/jython/devguide > <https://github.com/jython/devguide>. > > +1 > > You'll recall that one of the prompts for this was lessening > support for hg.python.org <http://hg.python.org>. We want to > follow CPython and get the Jython source over to GitHub. For that > we need a changed core development and contribution process, like > CPython's, where everyone can read it. So far, I have: > > 1. Stolen CPython's text and structure. > 2. Assessed which pages work for us unchanged or with only minor > change. > 3. Made parallel pages (name_jy.rst) where the content differs > widely. > 4. Integrated Frank's Jython-specific text from the current > devguide into those and some unique pages. > 5. Added a bit of content myself. > 6. Contributed a few technical improvements upstream to the > CPython devguide. > 7. Merged changes from the upstream CPython devguide (not just > mine) into the Jython devguide. > > The merge was not totally painless, but only 3 files gave me > conflicts. It took a couple of hours with kdiff3, and working > carefully from the Git Book. It's feasible, but there may be a > better way. Two of these files had a lot of differences. Somehow > kdiff3 seemed to guess correctly nearly everywhere. > > Yay, tools! Given the scope of just the devguide, it's good to hear > that it worked so well. > > Could we think about how we publish this so it becomes a > functional replacement for http://www.jython.org/devguide ? It is > not ready yet, but if we all think it will be, then the question > becomes how. I could set up publishing to readthedocs (on my own > account), but do we have a Jython/PSF one? Or is it better to take > the html to jython.org <http://jython.org> as we have previously? > > It was nice to see my PRs to the CPython guide, once accepted, go > live on docs.python.org <http://docs.python.org> within hours. I > assume this happens automatically. That's to aim for, I think, > creating one decision point (the PR acceptance) and giving the > rest of the job to machines. > > > +100 > |