ID:2115722
 
Resolved
New pixel gliding was broken in several cases.
BYOND Version:511.1348
Operating System:Windows 7 Home Basic
Web Browser:Chrome 51.0.2704.103
Applies to:Dream Seeker
Status: Resolved (511.1349)

This issue has been resolved.
Descriptive Problem Summary:

The picture is self explanatory, downloaded the beta and this is what happens when you move. Every version before works fine.


Nothing new has been added to the game. The client perspective is EDGE_PERSPECTIVE and the map_format is TOPDOWN_MAP.

Workarounds:
None yet.
Try setting atom.animate_movement to SLIDE_STEPS and the mob's appearance_flags to include LONG_GLIDE.
I had the mob's appearance_flags set to LONG_GLIDE beforehand and setting the atom.animate_movement to SLIDE_STEPS didn't do much. From what I read atom.animate_movement doesn't really have any effect with pixel movement.
Was worth a shot, both are known to help with the problem. I'm guessing it has something to do with 511's new pixel movement gliding stuff. Could try without LONG_GLIDE but I doubt that'd help.
Dang. This didn't show up in my own tests on pixel movement. I assume this is using pixel movement, yes?

Any chance you can boil this down to a test case and send it my way?
I am also experiencing this problem.
Hrm, is the client.eye value not the same as the mob? That might explain some of it; I was testing with a client.eye in sync with the mob, except that it did use EDGE_PERSPECTIVE.
for my case, the client.eye is not the same as the mob. It feels like the step_x and step_y values for the client.eye are being treated as zero.
Should I throw together a demo for this?
Sure. A demo that shows this problem in action would be super helpful.

At this stage I think I'm very likely to gather info on bugs for a while before the next release, so we can keep gathering info from 511.1348 for a while.
I had the client.eye set to something other than the mob but even after taking everything to do with the client.eye out, the problem remains. If it means anything, I found that as you walk towards the edge of a map the problem stops.
Lummox JR wrote:
New beta time!

New beta breaks modification of client.pixel_x and client.pixel_y ;__;

Very janky, it seems like sometimes client.pixel_* doesn't update for a few frames even though I haven't touched client.fps/tick_lag at all

I'll make a bug report after I do some more investigating, as there may be a workaround.
Noticed a couple issues, aside from how the camera apparently doesn't move by pixels anymore:

Changing direction (involving NORTHEAST or NORTHWEST, specifically?) causes pixel-glide jumpiness.

Enabling `LONG_GLIDE` seems pretty essential for proper timing for pixel-gliding.

Setting `animate_movement` to `FALSE` doesn't disable pixel-gliding, when I think it probably maybe should?
Ideally the quicker you guys can get me a test case for this, the quicker I can deal with it. I may have to disable pixel glides in the next release altogether till I can get it done more cleanly.
In response to Lummox JR
Lummox JR wrote:
Ideally the quicker you guys can get me a test case for this, the quicker I can deal with it. I may have to disable pixel glides in the next release altogether till I can get it done more cleanly.

Honestly, I just want a way to turn it off. I already have my own lerping system set up.
Actually I think I know what it's doing.


The glide works fine until you move onto a different tile. In this puush, what I think is happening, is it's jumping from step_y = 0 to step_y = 31, naturally, and creating a glide effect for that entire jump. Vice versa as well, from step_y = 31 to step_y = 0.
http://puu.sh/q12S8/6948b68492.zip
Here's a crappy demo I put together. It's really simple; You can make the client.eye follow the mob or follow some arbitrary camera object. You can move the camera and the mob separately, and you can observe the jumpiness of the camera (how it no longer moves pixel by pixel)


I would prefer to have a working pixel glides, so that I can have a higher clientside framerate than the server's tickrate without it looking like garbage.
Thanks. I'll see what I can do with the demo and if I can make the problems go away. (One thing's for sure: IMO, I think I need to treat LONG_GLIDE as on for all pixel-glide scenarios.)

If I can't get this fixed by the next release I'll just turn the pixel glide off until I can pin it down.
Experiencing this as well. client.eye is EDGE_PERSPECTIVE | EYE_PERSPECTIVE in my project.
Lummox JR resolved issue with message:
New pixel gliding was broken in several cases.
Page: 1 2