On Thu, Oct 10, 2013 at 12:37 PM, Steve Waterbury <water...@pangalactic.us>wrote:
> On 10/10/2013 04:54 AM, Łukasz Mach wrote: > >> W dniu 10.10.2013 03:43, Steve Waterbury pisze: >> >>> I'd have no expectation that after 'pip install' it would be usable as a >>>> regular platform Python library though. >>>> >>> >>> No, that would make absolutely no sense at all. >>> >> >> I cannot agree with your anti-pip-install-ism :). >> > > I'm not anti-it, I just don't see any advantage it gives. > Using pyjs in a directory has always worked for me, > no installation necessary. > > In what you quoted above, what I was saying was that > using pyjs as a regular Python library would make no > sense at all, not that developing a pip install capability > makes no sense at all -- it makes some sense, just not much > ROI that I can see ... although I admit it would > have psychological attractiveness, which shouldn't be > underrated. ;) > > > It should be fairly easy to avoid conflicting with core libraries and >> pyjs core libraries. >> >> setup.py can have some trigger which won't allow to install if not in >> virtualenv. >> >> Also, some part of pyjs/pyjamas library can be used by native python >> code. Eg. in pyjd, when is fixed, or in new style of desktop pyjamas >> (that one which uses qt/wxgtk...) >> > > I don't understand this idea -- there are way more mature and > powerful python desktop gui technologies that use wx/qt, > such as pyface / enaml. I don't see how pyjd or any new style of > desktop pyjamas could possibly offer any benefits over those ... > if anything, it would make sense to work on adding to pyjs a > capability for translating code from *them* into javascript ... > > Also, I think pyjd is a distraction from the unique strength of > pyjs: using Python to generate javascript for web applications. > IMO, if pyjs simply focuses on that alone, it will be an > excellent library. > the GI backend is 95% feature complete (save addEventListener, which i must still implement via ctypes/cffi), and the MS one is pretty much a non-issue. the GI backend supersedes all other linux backends (hulahop/pythonwebkit/etc), works fantastic, and once feature complete should work without issue for a very long time... it *should* even work with webkit2 variants, though i have not confirmed. while i don't think pyjd is necessary for success, it's absolutely wonderful for development... and that alone is enough to motivate me at least. as for pip, my goal is to consolidate all pyjs* commands into a single `pyjs [sub] [--arg]` style (eg. like git), which would encompass everything from translation to switchable pyjd backends, so being able to install via pip is a good thing IMO. the fact that all libraries are not useable by python is not an issue, because they can just be put outside PYTHONPATH, and only included as needed (the meta_path importer mentioned later will be able to do this transparently). additionally, and this is probably one of the better reasons, once i pull out the importlib stuff in favor of a meta_path impl, we'll be able to cache BOTH translated JS AND compiled python files for packaging... this means we can properly pre-translate/compile the entire library (with a couple common variants, eg. with/without --debug, etc), resulting in a SIGNIFICANT speed boost, due to the fact that the translator need only translate your code, and can include the stdlib verbatim (additionally, the waf builder properly caches your code locally, so only changed code is re-translated). ...this combined with the waf builder should result in rebuild that literally complete in less than 5-10 seconds, if that; i've already done massive amounts of development in distutils/waf/meta_path and all the other pieces needed to implement these things for related projects at $DAYJOB, i mainly just need to port them over to pyjs. SIDENOTE... i know i've been pretty silent for much longer than i'd like... since last year I've switched employers, moved twice (just last month!), got married (again last month!), my kid started school, my [now] wife completed grad school, and she just landed her job less than 2 weeks ago... needless to say i've barely found time whatsoever for devel outside work (which frankly, depresses me), but all along i've been trying to make pyjs relevant at $DAYJOB, which would allow me to do some devel on the clock... i very much think this is a real possibility rather soon, but as always, it's just a matter of time. do note, i don't offer any of this as an excuse/justification of anything (or even expect anyone to care for that matter ;), i'm simply relaying my situation which hopefully highlights some reasons for my lack of activity and/or presence. the good news is that, after 7+ years of schooling for my wife, moving around 2 states, and in general, dealing with non-stop upheaval, i am literally a month or 2 from finally being on a path to financial (and general) stability... to the point that i am now looking into buying the house i recently moved to. tl;dr... i'll brb :) -- C Anthony -- --- You received this message because you are subscribed to the Google Groups "Pyjs.org Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to pyjs-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.
