On Tue, 27 Sep 2011 23:44:06 +0200 Thomas Gstädtner <tho...@gstaedtner.net> said:
if we take your view that the only thing that matters in a release is sitting and waiting 6+ months for a distro to package it, then why do we care about release at all? reality is this: we release as source tarballs "day 1" and there will be the first major group of users that take this and use that and need it to work. that means it must work TODAY. if it doesn't work today no devs can even test it. no PA 1.0 here - no dbus API. if it wasn't for me testing it against an installed and working "today" distro and changing and fixing code - the PA support wouldn't work right. mike did some awesome work. he missed some critical small things and some integration work. i CAN'T test and fix against a PA 1.0 today... and mike can't because he doesn't have it either - that's WHY he "reverse engineered" the PA protocol. looking at the number of other people working on the e17 release todo list... actually getting things finished - well... those DOING the work that call the shots. those DOING the code to support PA don't have a dbus API... TODAY - when the code NEEDS TO BE DONE... and so... it didn't get done and won't be worked on by those people. if others wish to step up and work on the release todo - then maybe they can implement it if they have a dbus API available to test against. i'd welcome people stepping up and doing things. but the voices of those who worked on it says "no dbus API here.. so moot. not going to bother. what is there now works. what's the fuss?". if the issue is that "OMG!~ they may break internal protocol!!! use dbus!!! it won't break". please tell me that the networkmanager dbus api never broke. or connman... or... yeah right. i thought so. :) (it seems to me that people who use dbus love to break api's much more as the effects are less drastic than a c api break - where u get crashes or simply apps not starting at all, but in dbus land features just stop working or malfunction, so the error state degredation is less). so even if distros pick e up on 6+ months after release.. the PA code there will still work. or should. i don't see that going away in a hurry. now even after source tarballs we will have packages for fedora, ubuntu, debian etc. that are made within weeks of src release (or days) and so are available via extra repositories (or in debian testing/unstable which is what people actually use), so let's be realistic... e needs to work HERE. in this situation. we can't just say "oh well it doesn't work for 90% of potential users today but as long as everyone upgrades their distribution the instant they come out from distro vendors AND those distros ALL pick up PA 1.0 - sure over 3, 6, 9, 12 months e may being to be usable... so we just will ignore the problem and throw out work already done so we can ensure we get a reputation for releasing stuff that is not usable so maybe in 12 months 50% of possible users might have e work right (with audio)". why 50%? u REALLY think people upgrade their distribution instantly? i don't know what bizarre world you live in but they don't. they often lag behind the actual release (in fact the majority do i'd say) by at least 3-6 months or more. the majority of possible users may very conceivably be on a distro released 6 or 12 months ago. reality is we have to work for them. so that means working with what is out TODAY (or even last year). > Ok, let's get realistic: > _If_ e17 should go stable later this or early next year, it will take > a while until distros include it in their trees at all, lat alone > stabilize it (for the distros that maintain stable branches). > So here's a list of the major distros including a) the current release > with its PA version b) the first release e17 could be included in > should it be release in the next 6 months c) the first release that > will have PA with dbus support > > Debian > a) stable/squeeze: PA 0.9.21 b) unstable/sid: PA 0.9.23 c) unkown > (likely not before 7.0, but then again, e17 won't make it either.) > > Fedora > a) f15: PA 0.9.22 b) f16: PA 0.9.23 c) unknown (probably fast-tracked > to f16, more likely coming in f17) > > Gentoo > a) current: PA 0.9.23 b) ~testing: PA 1.0 c) ~testing (PA 1.0 will be > stable before EFL, let alone e17) > > OpenSuse > a) 11.4: PA 0.9.22 b) 12.1: PA 0.99/1.0 c) 12.1 (next release will be > early 2012 and include PA 1.0) > > Ubuntu > a) 11.04: PA 0.9.22 b) 11.10: PA 0.99/1.0 c) 11.10 (upcoming release > will include PA1.0) > > In conclusion: even if the e17 release happens very soon, it will not > be included in any stable tree this year, and very likely not even > next year. In the non-stable trees, of the major distros Ubuntu, > OpenSuse, Gentoo all already have PA 1.0 (or a 0.99 prerelease) and > likely have it stabilized before 2012. Fedora might, as they usually > stabilize core-packages much more quickly than other distros, Debian > of course won't, but sid will have 1.0 support soon. > > P.S.: List, please don't be insulted if I didn't include your distro, > it was not my goal to do you any psychological harm. I included those > 5 because I assume that they are the original distros which cover the > largest userbase. > > ------------------------------------------------------------------------------ > All the data continuously generated in your IT infrastructure contains a > definitive record of customers, application performance, security > threats, fraudulent activity and more. Splunk takes this data and makes > sense of it. Business sense. IT sense. Common sense. > http://p.sf.net/sfu/splunk-d2dcopy1 > _______________________________________________ > enlightenment-devel mailing list > enlightenment-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- The Rasterman (Carsten Haitzler) ras...@rasterman.com ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 _______________________________________________ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
