On Wed, Dec 31, 2014 at 11:16 AM, Stephen J. Turnbull email@example.com wrote: >
Fetchinson . writes:
I'm not sure exactly what would be discussed [on mobile-sig], especially in the short term when the big ticket item is getting the patches into the main Python tree.
It probably matters more that there is substantive discussion than what is discussed. There are probably a lot of people who are far more interested in "how to do 'it' now" than "get 'it' into the main Python tree 'someday'", for example.
I'm not sure if a mailing list would help with this as much as a good, simple set of docs. It's not like there's unknown territory here; once you've got Python installed on a mobile device, the "how" is reasonably easy. The catch is that someone with Python experience won't necessarily have experience managing a cross-platform C/Automake build, which is what you need to get Python installed in the first place.
However, once someone provides a binary for Python, others can just use that binary. That's what I've done with my platform support libraries and templates; they're a set of step-by-step instructions for going from a clean sheet to a running Python install.
Again, I'm not claiming my code is complete and production ready - just that this is an achievable goal IMHO, and my intention is to document "getting Python working" independent of getting any particular platform working. To my mind, while this could be maintained as an external set of patches etc, it makes sense to get the patches into the main source tree.
I'd think the core devs would be far more likely to include a patch
if it comes with a big red note on it saying "it was discussed on mobile-sig and was blessed by all participants" as opposed to a random patch by a random person.
Nope. It helps somewhat to have consensus backing for a particular patch by the (random) interested parties, but not that much. What helps a lot is to get someone trusted to have a good sense for "Pythonicity" and known to have experience in getting patches committed as spokesman.
IIUC correctly it is Russ's opinion that his set of patches is not very invasive and wouldn't fall afoul of disturbing some other applications or platforms. If so, the real problem is to get core attention.
That is correct. There are a couple of tweaks needed to improve cross platform build support, a couple of tweaks introducing sys.platform support for iOS and Android, and some other miscellaneous patches to correct support for individual features. There's a bunch of little changes required, but it's not a big patch by any measure.
Although it's a couple of months off, I would suggest a talk (if it's
still possible to schedule), a BOF, and a sprint at Pycon North America in March. That should get things moving, and several of the "important" folks who have already expressed interest should be there, too.
Unfortunately, due to the tyranny of distance (I'm based on the west coast of Australia), plus work commitments, it's unlikely that I'll be attending PyCon in Montreal.
Yours, Russ Magee %-)