Page 1 of 1

[engine] known (mentionable) bugs of spring 98.0

Posted: 10 Oct 2014, 16:26
by abma
some users reported issues in the current release of spring 98:

1. Stuck in drawing mode
2. f6 does not sound mute
3. gfx errors with ATI and UsePBO=1
4. some GFX related crashes
5. SPADS: failed to upload demo

1. to exit the "always draw mode" leave the chat mode by pressing enter and press the draw key again (mostly ´ or §)

2. can be easily workarounded by adjusting uikeys.txt or use a widget which fixes the key binding and is already fixed in the current development version of spring

3. can be easily prevented by enabling safemode or setting UsePBO=0.

4. if spring crashes and you have this line in infolog.txt:

Code: Select all

This stack trace indicates a problem with your graphic card driver. Please try upgrading or downgrading it. Specifically recommended is the latest driver, and one that is as old as your graphic card. Make sure to use a driver removal utility, before installing other drivers.
please set "DebugGL = 1" in springsettings.cfg and check if a bug report about this already exists, if not, please create one!

even that 1. is VERY annoying and should be fixed its imo no blocker for setting 98.0 live.

Thanks for the feedback!

Re: [engine] known (mentionable) bugs of spring 98.0

Posted: 10 Oct 2014, 19:49
by abma
when spring 98 crashes for you, please provide:
- infolog.txt
- demo file when relevant

this increases the chance, that we can fix the crash!

thanks!

Re: [engine] known (mentionable) bugs of spring 98.0

Posted: 10 Oct 2014, 20:03
by muckl
i think point 1 workaround will be super annoying
like the problem with the first letter gets lost in chat function

why not wait until it got fixed?
it just makes the crowd mad cause its not fixable - like the crashes in some cases

Re: [engine] known (mentionable) bugs of spring 98.0

Posted: 11 Oct 2014, 01:39
by abma
muckl wrote:why not wait until it got fixed?
it just makes the crowd mad cause its not fixable - like the crashes in some cases
switching back would cause a lot of trouble, too. many problems are because user edited config files (often to fix some other bug/problem i guess) and set some settings which now cause problems. also currently the decision which engine version is used is mostly made by game devs. its not so simple as it sounds.

imo spring 98 has less bugs than 96.0, atm it looks just like the "usual" problems when switching to a new engine.
i'm pretty sure most current known crashes can be workarrounded by disabling some setting. give spring --safemode a try!

also we have way to few testers / very few different machines on which the development versions are tested. without the help of the community (in this case normal players), we can't find and fix all bugs. we really tested a lot before the release, but... we try at least :-)

also switching back is quiet frustrating, pretty bad for motivation. (seen from dev side)

Re: [engine] known (mentionable) bugs of spring 98.0

Posted: 12 Oct 2014, 13:55
by BrainDamage
several windows users cannot reliably get back to fullscreen if they switch window out the game ( doesn't matter if it's alt tab or lose focus due to third party programs popup windows ), the /fullscreen shortcut doesn't reliably work as in they have to try many times until it works

the only workaround so far is not use fullscreen and use borderless windowed, in short:
/set Fullscreen 0
/set WindowBorderless 1
restart

Re: [engine] known (mentionable) bugs of spring 98.0

Posted: 12 Oct 2014, 14:03
by FLOZi
harvestStorage tag and associated lua calls only work with reclamation of units, not features.

Re: [engine] known (mentionable) bugs of spring 98.0

Posted: 21 Oct 2014, 10:48
by Silentwings
Collision checking is unreliable for some weapon types with per-piece colvols. http://springrts.com/mantis/view.php?id=4592

Some (uncommon, but regular enough to be noticeable) circumstances cause labs to create units that cannot move. http://springrts.com/mantis/view.php?id=4587

There are also sometimes non-reproducible crashes while loading for some windows users (including me) but currently I'm not able to produce stacktraces on windows without self-compile, which I don't have set up atm, so either someone else produces a stacktrace or it has to wait.