On Sun, Jan 29, 2012 at 10:03 AM, C Anthony Risinger <[email protected]> wrote: > On Sun, Jan 29, 2012 at 3:35 AM, lkcl luke <[email protected]> wrote: >> On Sat, Jan 28, 2012 at 11:47 PM, C Anthony Risinger <[email protected]> wrote: >> >> splitting the files up - badly - and making a dog's dinner out of >> things by duplicating files and having others that maybe import from >> some, have the same class name but extended functionality, such that >> even the lead developer doesn't know what the fuck is going on, is a >> bad idea. >> >> so, chalk that one down to experience: try a different route to >> achieve the same goal. > > if we can make it all work from one file that would be fantastic. i > don't like how fragmented it is either, **especially** the overrides > ... i didn't realize what was really going on until hours of wasted > time (back when). but i also don't like the complete impossibility of > *pretending* GWT is upstream, while simultaneously sprinkling > modifications all over the place. we don't have to use the raw > translation as-is, but we need to adhere to a process/toolchain.
*nods vigorously*. > i 999% believe we need commit full-force to GWT as an upstream, or > just walk away, now, because it's difficult enough already without > internal chaos. *thinks*... i'm kinda nervous about throwing away what's been tried-and-tested [the existing pyjamas ui codebase] even though it's not "perfect". if you recall, in off-list conversations when there was the possibility of funding pyjamas development full-time for several months, i advocated a piecemeal conversion process. however, now that the library/gwt/* mess is gone, i see no reason why not to just do an auto-conversion of the gwt code and just commit it, and go from there.... leaving the existing pyjamas/ui code entirely alone. that would at least give a starting point, and allow people to take a look, see what needs to be done, and what's practical. i think also it will clearly show people just quite how shit java really is. i don't think people who use pyjamas - except those who've come from gwt originally - realise just how subtle some of the problems are. > there are too many engines, too many browsers, and too much cruft and > obsolescence strewn about. nobody is really testing changes because > it's just not feasible (btw, the mail app is completely busted when > built with a non-2010 version of pyjs) ... yes - that was around the time where i took more of a back seat (more specifically: changed leadership strategy), having spent vast amounts of time keeping things constantly up-to-date, constantly running regression tests, i decided to create the DEVELOPER.RULES and asked people to follow it. ... unfortunately, that didn't happen. now - as a community - the community sees the consequences of not following the guidelines that i set: you - _you_ as a community - end up with code that is of poorer quality. me? i primarily use pyjd, for personal work. and *if* i get paid to do some work using pyjs, i fix the problems encountered. otherwise - i'm letting _you_ - as a community - take an active role in sorting things out. l.
