Hi, 2016-12-03 16:36 GMT+01:00 Raoul Snyman <[email protected]>:
> The time between major releases is too long. I fixed 6 major video bugs a >> year ago and with the Beta and RC releases they will not be out till >> Easter. This is not a complaint but a fact of life, when we need to >> release good bug-free software. >> > > Very true. This has been the most chaotic year of my life, to be honest. > Aye, limited resources and time is an issue, but such is life. > > > OpenLP is getting more stable and the bug re-writes should be getting less >> but we need to be able to continue to release so here is a proposal. >> >> >> 1. Release 2.6 with current functionality (maybe we fix up some of the >> small improvements) and target it for Easter >> 2. Add web updates into 2.8 >> 3. Target renderer to 2.8 or 3.0 depending on how long it takes to make >> stable >> > > Yes, I think this is a good idea. My only proviso is this: WebKit is going > the way of the dodo, WebEngine is supposed to be the new cool, but between > the distros and the macOS and Windows builds, I'm not sure how much longer > we can use WebKit for. > It would be really great to do the switch from QtWebkit to QtWebEngine before 2.6, but I guess it isn't realistic to get it done unless someone suddenly gets a lot of time on his hands. But since QtWebEngine still hasen't made it into debian and ubuntu yet, maybe that is ok for now - though we could probably use snap-packages as a workaround. If we stay on QtWebkit we are also stuck on python 3.4 and PyQt 5.5 on windows (since later PyQt versions for win doesn't contain QtWebkit). But that could actually be good, since PyInstaller currently has an issue with Python 3.5 that causes apps built on win10 not to work on older windows versions (https://github.com/pyinstaller/pyinstaller/issues/1566). So maybe it's a good idea to stick with QtWebkit for one more release. > > Maybe change the naming OpenLP 2016 and OpenLP 2017 targetted for the 1st >> quarter each year would be a way forward. with bug fixes targetted at Sept >> and Dec unless there are major showstoppers like HTML changes at >> SongSelect >> or Bible Gateway. >> > > Urgh, personally I hate year numbers instead of version numbers. But > targetting a release for the first quarter of each year sounds like a good > idea. > > I do think we've gotten the release procedure down fairly well (especially > when I follow the instructions I wrote for myself on the wiki), so once we > have a regular release cycle for new versions, maybe we can speed it up to > 2 per year. > > I would also like to stick to the current numbering scheme, but targeting a feature release for each spring sounds good to me. Best regards, Tomas
_______________________________________________ openlp-dev mailing list [email protected] https://lists.openlp.io/mailman/listinfo/openlp-dev
