Spring 0.76: pending stuff
Moderator: Moderators
-
- Spring Developer
- Posts: 1254
- Joined: 24 Jun 2007, 08:34
- Forboding Angel
- Evolution RTS Developer
- Posts: 14673
- Joined: 17 Nov 2005, 02:43
Erm no.
Frequent releases are better because:
As for the other suggestions, thanks for your input and I am sure they would be really useful, but I do not have time to care about those (IMHO minor) points, so I'd suggest you to write a patch yourself or find someone to do it for you.
They also are in no way release blocking. Heck, many of them are even feature requests technically...
Frequent releases are better because:
- The people here who mod and make patches sometimes like to be able to use what they made without waiting a year. IOW, the more frequent releases are, the more people are motivated to contribute.
- Less bugs in releases that follow each other shortly (compare 0.75b1 and 0.75b2 anyone?)
- More feedback. I still barely see any of the people in this thread organising big MP battles on the test server, so the only effective way of testing and getting feedback and bug reports is still doing a real release.
- Related to that, feedback is more useful the sooner you get it, so more frequent releases are better then less frequent releases.
- Not releasing because 'there are not enough major changes' would probably result in feature creep => not releasing at all for a long time.
As for the other suggestions, thanks for your input and I am sure they would be really useful, but I do not have time to care about those (IMHO minor) points, so I'd suggest you to write a patch yourself or find someone to do it for you.
They also are in no way release blocking. Heck, many of them are even feature requests technically...
+1Tobi wrote:Frequent releases are better because:
- The people here who mod and make patches sometimes like to be able to use what they made without waiting a year. IOW, the more frequent releases are, the more people are motivated to contribute.
- Less bugs in releases that follow each other shortly (compare 0.75b1 and 0.75b2 anyone?)
- More feedback. I still barely see any of the people in this thread organising big MP battles on the test server, so the only effective way of testing and getting feedback and bug reports is still doing a real release.
- Related to that, feedback is more useful the sooner you get it, so more frequent releases are better then less frequent releases.
- Not releasing because 'there are not enough major changes' would probably result in feature creep => not releasing at all for a long time.
unitsync seems fixed
Also I added the test trepan made for it to buildbot (and it's being run in valgrind), so hopefully any memory corruption/leaks introduced from now on is catched early on.
Those who look on the buildbot status page sometime may have noticed too that I added some tests of the installer (using silent (un)install in wine ).
One test tests that uninstall truely removes all installed files, and the other test tests that install/upgrade/uninstall truely removes all installed files and also that the installed files are all the same (though that last part is not finished yet, it's still a bit crappy).
Also I added the test trepan made for it to buildbot (and it's being run in valgrind), so hopefully any memory corruption/leaks introduced from now on is catched early on.
Those who look on the buildbot status page sometime may have noticed too that I added some tests of the installer (using silent (un)install in wine ).
One test tests that uninstall truely removes all installed files, and the other test tests that install/upgrade/uninstall truely removes all installed files and also that the installed files are all the same (though that last part is not finished yet, it's still a bit crappy).
props for setting up tests, much needed. sorry for unitsync b0rkage, too; it's nearly impossible to debug on windows with all those different lobbies and gdb printing broken stacktraces after crashes.
PS. test uninstall could also test if it doesn't remove any additional files (one of KP installers had such 'feature' IIRC, and some early Firefox version uninstalled whole Program Files.)
PS. test uninstall could also test if it doesn't remove any additional files (one of KP installers had such 'feature' IIRC, and some early Firefox version uninstalled whole Program Files.)
Yeah, I forgot to add the shaders directory to the scripts to regenerate springcontent.sdz, so some revisions didn't run.KDR_11k wrote:Ah, okay. In 4714 (full install!) I had to copy the folder over because it wouldn't run otherwise.
BTW, a later version worked properly.
Newer builds should run fine again.
Thanks. Yeah, debugging DLLs is kinda hard/impossible without test suite.imbaczek wrote: props for setting up tests, much needed. sorry for unitsync b0rkage, too; it's nearly impossible to debug on windows with all those different lobbies and gdb printing broken stacktraces after crashes.
PS. test uninstall could also test if it doesn't remove any additional files (one of KP installers had such 'feature' IIRC, and some early Firefox version uninstalled whole Program Files.)
I will try to get around to better uninstall test, but first need to fix up the upgrade test a bit more. It doesn't properly check yet that indeed an upgraded install is identical to a fresh install, which was the goal of the test
I also plan to put buildbot stuff into git sometime, so then I could even accept patches for the tests