| View Issue Details [ Jump to Notes ] | [ Issue History ] [ Print ] | ||||||||
| ID | Project | Category | View Status | Date Submitted | Last Update | ||||
|---|---|---|---|---|---|---|---|---|---|
| 0002671 | Spring engine | General | public | 2011-09-25 15:17 | 2012-05-29 13:08 | ||||
| Reporter | hoijui | ||||||||
| Assigned To | Kloot | ||||||||
| Priority | normal | Severity | minor | Reproducibility | sometimes | ||||
| Status | resolved | Resolution | fixed | ||||||
| Product Version | 0.82.7+git | ||||||||
| Target Version | Fixed in Version | 85.0 | |||||||
| Summary | 0002671: way too long path chosen on TheRockFinal | ||||||||
| Description | best understood by looking at the screenshot: When giving a move command from "start" to "end", the peewees take the long path through the valley. When giving a move command from "turnings point" to "end", the peewees take either the long path through the valley, or the "long way". sometimes though, depending on where exactly they stand (near "turning point"), they may also take the direct route. Same behavior applies when walking form "end" towards "turning point", sometimes they take the "long way", and sometimes the shortest/direct path. | ||||||||
| Steps To Reproduce | map: TheRockFinal mod: XTA 9.66 /cheat /give 10 arm_peewee make them walk! (see screenshot) | ||||||||
| Tags | No tags attached. | ||||||||
| Checked infolog.txt for Errors | |||||||||
| Attached Files |
| ||||||||
 Relationships |
|
 Relationships |
| Notes |
|
|
Kloot (developer) 2011-09-25 16:01 Last edited: 2011-09-25 16:07 |
a short explanation of why this happens: 1) there are three pathfinders in Spring, each operating at a different resolution 2) depending on the distance between start and goal, one of the three is chosen (longer distance ==> lower resolution, saves _a lot_ of CPU) 3) paths that exist at high resolution do not always exist at lower resolution (typically, when there are passages in the high-res path smaller than the resolution of the chosen pathfinder) ==> units go the long way around to the naked eye (to solve this I am considering to limit the number of path requests per frame and executing each query at highest resolution, but not for 0.83 yet) |
|
jK (developer) 2011-09-25 18:37 Last edited: 2011-09-25 20:49 |
how are those low resolution maps computed? shouldn't they be extrapolated from the high res one? and shouldn't it then check if any of the hires-squares it extrapolates from is open and then mark the whole new lowres-square as open in that direction? (similar to a texture that contains only black & white texels and you create a mipmap from it, and if a single texel is white in a 2x2 box it makes the whole mipmap texel white) |
|
Kloot (developer) 2012-05-29 13:08 |
fixed with QTPFS |
| Notes |
| Issue History |
|||
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2011-09-25 15:17 | hoijui | New Issue | |
| 2011-09-25 15:17 | hoijui | File Added: screen00148cut.png | |
| 2011-09-25 16:01 | Kloot | Note Added: 0007408 | |
| 2011-09-25 16:02 | Kloot | Note Edited: 0007408 | |
| 2011-09-25 16:07 | Kloot | Note Edited: 0007408 | |
| 2011-09-25 18:37 | jK | Note Added: 0007410 | |
| 2011-09-25 20:47 | jK | Note Edited: 0007410 | |
| 2011-09-25 20:49 | jK | Note Edited: 0007410 | |
| 2012-05-29 13:08 | Kloot | Note Added: 0008719 | |
| 2012-05-29 13:08 | Kloot | Status | new => resolved |
| 2012-05-29 13:08 | Kloot | Fixed in Version | => 85.0 |
| 2012-05-29 13:08 | Kloot | Resolution | open => fixed |
| 2012-05-29 13:08 | Kloot | Assigned To | => Kloot |
| Issue History |