2021-04-22 20:17 CEST

View Issue Details Jump to Notes ]
IDProjectCategoryView StatusLast Update
0002116Spring engineGeneralpublic2012-05-10 08:51
Reporterdfreeman 
Assigned Toabma 
PrioritynormalSeverityminorReproducibilityhave not tried
StatusresolvedResolutionfixed 
Product Version89.0 
Target VersionFixed in Version 
Summary0002116: Player start pos'ns on minimap are huge
DescriptionAt the start of a game, for no apparent reason, all the player start positions on the minimap were drawn with huge circles. I tried to reconnect, but the same problem appeared. Someone said it was a bug in RedUI's minimap.

When the game started, something (probably a circle) grew to the size of the screen so that the whole screen was green. I couldn't see anything, and the software mouse cursor was soooo slow.
TagsNo tags attached.
Checked infolog.txt for Errors
Attached Files

-Relationships
duplicate of 0002048closed Screen gets sometimes full color of particle(?) effects (nano, dgun,...) from minimap-widget 
has duplicate 0002223closed green screen when construction begins 
has duplicate 0002309closed Minimap bug with lolui 
related to 0003072resolvedabma glPoint used for projectile rendering on minimap breaks on ATI 
+Relationships

-Notes

~0005522

dfreeman (reporter)

I tried again, and RedMiniMap has no effect, on or off the circles are huge.

Another user told me that (s)he has the same bug, and you have to restart the OS for it to work normally again.

Should I admit at this time that I use an ATI card? :)

~0005523

abma (administrator)

maybe a duplicate of http://springrts.com/mantis/view.php?id=2048 ?

it's also an ati-problem, yes...

~0005524

dfreeman (reporter)

I updated my Linux ATI drivers to the latest, Catalyst 10.8.

I found that viewing a replay where this bug was present, will trigger it. Then watching a replay where it wasn't present, causes it to behave normally.

So I've attached a very short replay where I bailed before the game began. You should be able to witness this bug in full effect.. hopefully :)

~0005525

dfreeman (reporter)

Yes, this is a dupe of 2048. Screenshots are similar.

~0005526

abma (administrator)

Last edited: 2010-09-14 19:36

i've added two screenshots. when starting the replay a few times, in about 2/3 of the tries the problem occurs.

edit: 8.png + 9.png are from the same replay!

~0005528

dfreeman (reporter)

I didn't try enough times to notice if it works 1/3 of the time.

Glad you can repeat this :) Are you using ATI?

~0005534

abma (administrator)

yep... :-/

~0005542

dfreeman (reporter)

The "fade to green" thing can happen without giant balls in the minimap, if it's a single-player game. It fades to green when the commander is nanospraying, and fades to normal when it finishes building. Green, normal, green, normal. Ragequit :)

~0005616

dfreeman (reporter)

It turns out that, although RedMinimap has no effect on the giant balls while players are selecting startpoints, it does have an effect on the fade to green (or in one case, fade to red) bug.

While the screen is toggling to solid green during the commander's nanospray, toggling RedMinimap either off or on is enough to clear the problem for the remainder of the game. It will probably happen again on the next game.

I don't think this is a RedUI issue, I think it's just reinitialising the minimap whenever it is toggled off or on.

~0005798

dfreeman (reporter)

For the last few games, I've had giant dots on minimap start pos, but it hasn't led to the whole screen being a single colour during nanospray.

I don't know what I've changed to make this happen. I have RedMiniMap disabled now, but I don't think that explains it.

Could be a recent ATI driver update.

~0006049

abma (administrator)

Last edited: 2010-12-09 16:36

added simple lua-wigdet to allow easy testing.

(taken from here:) http://springrts.com/phpbb/viewtopic.php?p=448645#p448645

~0008643

abma (administrator)

related commit:

https://github.com/spring/spring/commit/2136017c4fa035463dcc21a2c6673e8bbcfdb936

~0008644

abma (administrator)

infolog.txt:

[f=0000000] SDL version: 1.2.14
[f=0000000] GL version: 4.2.11627 Compatibility Profile Context
[f=0000000] GL vendor: ATI Technologies Inc.
[f=0000000] GL renderer: ATI Radeon HD 5700 Series
[f=0000000] GLSL version: 4.20
[f=0000000] GLEW version: 1.6.0
[f=0000000] Video RAM: total 630MB, available 831MB
[f=0000000] GL info:
    haveARB: 1, haveGLSL: 1, ATI hacks: 0
    FBO support: 1, NPOT-texture support: 1, 24bit Z-buffer support: 1
    maximum texture size: 16384, compress MIP-map textures: 1
    maximum SmoothPointSize: 63, maximum vec4 varying/attributes: 32/29
    maximum drawbuffers: 8, maximum recommended indices/vertices: 16777215/2147483647

atifail.lua still triggers the bug for me, the whole screen is yellow when i activate it.

~0008650

jK (developer)

what happens if you wrap the widget's draw code in gl.PointSprite(true) ... gl.PointSprite(false)?

~0008651

abma (administrator)

changed to that: https://gist.github.com/2580827

no change :-/


the bug is like this:

after computer restart it works for some time. then something triggers the bug and the whole screen is yellow. only if i reboot the computer, the problem is gone. Restarting spring doesn't help too.

So: either find the cause, what triggers the bug (i have no idea how gfx / opengl can be debugged) or find a way how the driver/opengl can be reset when spring starts. i saw this bug in an other game, too...

~0008652

jK (developer)

too bad
then try: gl.PointParameter(1,1,1, 1,5,5)

~0008653

abma (administrator)

Last edited: 2012-05-03 11:24

seems to work, the point is ~1px big:
https://gist.github.com/2580827/705ad785e868be49b4a1fc9ba7464b3a12b8d3af

i hope i did the change correct...

~0008654

jK (developer)

no, you should have left the other codelines unchanged -> esp. the PointSize one

~0008655

abma (administrator)

Last edited: 2012-05-03 11:31

this way?

https://gist.github.com/2580827/fcbcaf9bd202475c657d0690f2da1d7275b73192

works, too.

without glPointParameter the whole screen is yellow.

~0008656

abma (administrator)

Last edited: 2012-05-03 12:01

it seems like calling glPointParameter fixes most "whole screen is red/green/..." problems
when the atifail.lua is disabled, the whole screen gets green with middle-click scrolling
when enabled, its fine...

it doesn't fix the "red minimap" problem / "BuildBar" making the whole screen green/red (but maybe calling glPointParameter at some other places could fix that too)

with smoothpoints / lines enabled it seems like i can't get one of the whole screen is red/green/ ... problem.

~0008677

abma (administrator)

the workaround is enabled when ATIHacks is enabled and GLSL is available.

see https://github.com/spring/spring/commit/87e9dbedf1a1a943da2e6d49c2dc9e0a2f7f4237
+Notes

-Issue History
Date Modified Username Field Change
2010-09-14 18:28 dfreeman New Issue
2010-09-14 18:34 dfreeman Note Added: 0005522
2010-09-14 18:36 abma Note Added: 0005523
2010-09-14 19:07 dfreeman Note Added: 0005524
2010-09-14 19:09 dfreeman File Added: 20100915_022901_DeltaSiegeDry_0.82.5.sdf
2010-09-14 19:12 dfreeman Note Added: 0005525
2010-09-14 19:34 abma File Added: screen00008.png
2010-09-14 19:34 abma File Added: screen00009.png
2010-09-14 19:35 abma Note Added: 0005526
2010-09-14 19:36 abma Note Edited: 0005526
2010-09-14 22:36 hoijui Relationship added duplicate of 0002048
2010-09-15 05:31 dfreeman Note Added: 0005528
2010-09-15 10:25 abma Note Added: 0005534
2010-09-16 11:17 dfreeman Note Added: 0005542
2010-09-22 07:37 dfreeman Note Added: 0005616
2010-09-23 21:53 zerver Severity crash => minor
2010-10-28 06:29 dfreeman Note Added: 0005798
2010-12-09 16:04 abma File Added: atifail.lua
2010-12-09 16:05 abma File Deleted: atifail.lua
2010-12-09 16:05 abma File Added: atifail.lua
2010-12-09 16:06 abma Note Added: 0006049
2010-12-09 16:36 abma Note Edited: 0006049
2011-03-18 18:04 abma Relationship added has duplicate 0002223
2011-03-18 18:04 abma Relationship added has duplicate 0002309
2012-04-26 23:22 abma Note Added: 0008643
2012-04-26 23:31 abma Note Added: 0008644
2012-04-27 01:14 abma Product Version 0.82.5 => 89.0
2012-05-02 14:46 abma Relationship added related to 0003072
2012-05-02 19:36 jK Note Added: 0008650
2012-05-03 00:13 abma Note Added: 0008651
2012-05-03 09:57 jK Note Added: 0008652
2012-05-03 11:15 abma Note Added: 0008653
2012-05-03 11:19 abma Note Edited: 0008653
2012-05-03 11:19 abma Note Edited: 0008653
2012-05-03 11:19 abma Note Edited: 0008653
2012-05-03 11:24 abma Note Edited: 0008653
2012-05-03 11:27 jK Note Added: 0008654
2012-05-03 11:28 abma Note Added: 0008655
2012-05-03 11:31 abma Note Edited: 0008655
2012-05-03 11:57 abma Note Added: 0008656
2012-05-03 11:57 abma Note Edited: 0008656
2012-05-03 11:59 abma Note Edited: 0008656
2012-05-03 12:01 abma Note Edited: 0008656
2012-05-10 08:51 abma Note Added: 0008677
2012-05-10 08:51 abma Status new => resolved
2012-05-10 08:51 abma Resolution open => fixed
2012-05-10 08:51 abma Assigned To => abma
+Issue History