Page MenuHome

Hair from 2.79 messed up in 2.8 (modifiers prior to PSys and partially broken without 'Use Modifer Stack')
Open, Confirmed, LowPublic

Description

System Information
Operating system:
Graphics card:

Blender Version
Broken: blender-2.80.0-git.dd3f5186260e-windows64
Worked: 2.79b release

Hair particles made in blender 2.79 get messed up, when opend in 2.8. The hairs are pulled in direction of objects origin:

Exact steps for others to reproduce the error
open attached file in 2.80 (and in 2.79, to compare)


It is maybe is related to https://developer.blender.org/T59587 .. i am not sure.

Details

Type
Bug

Event Timeline

Jacques Lucke (JacquesLucke) lowered the priority of this task from Needs Triage by Developer to Confirmed, Medium.Jan 22 2019, 5:36 PM

Sorry this has been lying around for so long, I am looking into some hair bugs atm...

Note: this can be solved by:

  • removing the Edge Split modifier or
  • applying the Edge Split modifier or
  • checking Use Modifier Stack in Particle Settings > Emission > Source

2.8 seems to be stricter about modifiers above the Particle System modifier the usage of Use Modifier Stack here

without looking deeper at code, this sometimes makes more sense tbh:

generative modifiers:

  • bevel, edgesplit (work in 2.8 with Use Modifier Stack, however in 2.8 disabling Use Modifier Stack makes particles disappear/grow from the origin, 2.79 worked without Use Modifier Stack)
  • boolean (works in 2.8 with Use Modifier Stack, wasnt working at all in 2.79)
  • array (works in both 2.79/2.8 with Use Modifier Stack, weird in both 2.79/2.8 without Use Modifier Stack, emits from orig faces, but also (partially) from new faces)
  • remesh (works in both 2.79/2.8 with Use Modifier Stack, makes particles disappear/grow from the origin in both 2.79/2.8 without Use Modifier Stack)
  • solidify (works in both 2.79/2.8 with Use Modifier Stack, weird in both 2.79/2.8 without Use Modifier Stack -- seems to flip normals on the original?)

deforming modifiers:

  • from first glance these are all taken into account in both 2.79/2.8 no matter if Use Modifier Stack is ON or OFF...

So where does this leave us?

  • given the fact that most of it seems fine in 2.8 with the option Use Modifier Stack enabled: maybe this should be ON by default?
  • given the fact that having this disabled in 2.8 will certainly look like bugs in the case of this report (generally bevel, edgesplit, remesh, ...): should we make sure this is really just going after the non-modifier mesh (and nothing else)?
  • the manual is really not of a help here (to me it reads like: OFF="no modifiers whatsoever" -- this is not the case...)
  • @Jacques Lucke (JacquesLucke) : functions to the rescue ;)

Will assign to @Sergey Sharybin (sergey) (for guidance on how to proceed -- feel free to throw back at me and I will check code more thoroughly...)
CC @Brecht Van Lommel (brecht)

Philipp Oeser (lichtwerk) renamed this task from Hair from 2.79 messed up in 2.8 to Hair from 2.79 messed up in 2.8 (modifiers prior to PSys and partially broken without 'Use Modifer Stack').Jul 11 2019, 2:37 PM
Brecht Van Lommel (brecht) lowered the priority of this task from Confirmed, Medium to Confirmed, Low.

Most likely we solve this with a new hair system, and spend no more time on trying to fix this in the current particle system.

Thx for looking into it Philipp... now i feel dumb waiting for so long when a workaround was that easy.