Beta 5 Release Plan

edited 2006 Dec 23 in Developers
<i>This post was originally made by <b>danij</b> on the dengDevs blog. It was posted under the categories: Beta 5, Engine, Games, jDoom, jHeretic, jHexen, Releases, Version 1.9.</i>

We a rapidly approaching 5 months since beta4 was released. I think we should now be looking to release beta5 ASAP.

There has been some significant changes (and many bug fixes) since then and it would be a good idea to get them in the hands of the community.

So, what work is there remaining which must be completed before beta5 can be released? The rest we should postpone until a later beta.

Comments

  • I will need to do a few minor checks/updates with the build scripts (I don't have much time at the moment to check other things.) As far as I can see, everything is running well so far - with the caveat that I have not tested multiplayer at all, and demo playback is still very much broken (both wonky playback) and last time I tried - crashes on finishing. Oh - and snowberry svn really needs work.

    I have public build logs of the last few svn commits, and will continute to provide them at http://eyagi.bpa.nu/~jamie/deng/buildlogs/ as I build new revisions. Online source is available here http://eyagi.bpa.nu/~jamie/deng/sourcebrowser/ to browse.

    Personally I favour the release early, release often strategy, so I can get my stuff finished by this weekend.

    I do need Win32 and OSX binaries, and I need to hit up the wiki and jfiles to prepare the torrent, and I'll need to look for a torrent tracker before we make the release.

    Once we finalise the release, and make the source tarballs, I'll need 1-2 days to make the Ubuntu packages, otherwise, it should be all fine.
  • As most of you will be aware; skyjake is currently missing in action (still)...

    However, work has still been progressing fairly steadly (but slowly due to the developers' real life commitments).

    As I mentioned on the forums recently, it is my intention to have Doomsday 1.9.0-beta5 released on or around Christmas.

    To that end the team have decided to move all non-critical work previously scheduled for beta 5 to beta 6 (see the DEW for <a title="more info" target="_blank" rel="nofollow" href="http://dengine.net/dew/index.php?title=Version_1.9.0-beta5">more info</a>) in an effort to get a new release made as soon as possible.

    So, between now and then I will be focusing mainly on bug fixing and testing. With a bit of luck beta5 will be polished enough to be considered a "stable" release and as such suitable for day-to-day use (though it will NOT be released as 1.9.0-full due to the fact that the planned API changes have not been completed). It has been quite a while since Doomsday 1.8.7 (the last "stable" release)...
  • I think the plan should now be that Beta 5 is released as soon as possible. Bugs or problems that are not absolutely critical should be postponed to later releases. I seem to recall some issues with demo recording and playback. IMHO demos are not critical and should be postponed.

    If you know of any critical bugs or other things that absolutely need polishing, we should apply those now, then open <tt>/branches/1.9.0-beta5</tt> and release 1.9.0-beta5.0 from there. If any bad bugs slip through, we can release 1.9.0-beta5.1 from the release branch. In the trunk we can immediately continue working on Beta 6 stuff, and break the code as necessary.

    I would very much like to see a Beta 6 already sometime around January 10, which is my last vacation day.

    Thinking about the coming spring, I would like to propose that we try to commit to a more regular release cycle. This invariably means that features will have to be sometimes pushed back to later releases, but it's more important to make steady progress in the form of fully functional releases. It helps to keep the code base in a better shape and the user community benefits from having access to more recent releases.

    I was thinking that an 8 week cycle might be suitable. This would include 7 weeks of development work, followed by one week of release preparations, polishing, packaging, some quality assurance, and finally the release on all platforms by the end of the week.

    What do you think?
  • Yeah, that sounds like a good plan to me. I do get the feeling that the community is suffering somewhat due to the infrequency of releases.
  • I'm very happy with an 8 week cycle. I've got to test the recent changes on *NIX shortly before greenlighting beta5.
Sign In or Register to comment.