Helps to format :)
I call it a bug, but it’s been around a loooong time.
Maya updates the ikHandles in a separate pass (there’s just one solver for
multiple handles), and I don’t believe that script mode forces an update on
the solver, where moving the root updates the DAG directly.
It’s one of the several places where the maya internals bypass the DAG so
that they can work (IK handles don’t put input connections on the joint
rotations as that *would* be a cycle, so the update needs to live in a
callback somewhere, which isn’t triggered by your code). Another place
where things get weird is with editable animation curves (the ones where
you’ve got an editable curve in the scene that updates the animation
curves).
Long story short, querying an attribute at a given time only works for
direct DAG evaluation. Anything which operates outside of the DAG won’t be
updated when you call getAttr. For those cases, you actually have to set
the time and refresh the evaluation prior to the getAttr call:
from maya import cmds
# Build Scene
# o o
# \ /
# \ /
# o
#
cmds.file(new=True, force=True)
a = cmds.joint(name="Root")
cmds.move(0, 0, 20)
cmds.joint()
cmds.move(0, 0, 0)
b = cmds.joint()
cmds.move(0, 10, -20)
handle, eff = cmds.ikHandle(
solver="ikRPsolver",
startJoint=a,
endEffector=b,
forceSolver=True,
)
assert cmds.getAttr(a + ".rx") == 0.0
cmds.refresh(force=True) # Side note, this means your scene actually
has to play through
assert cmds.getAttr(a + ".rx") == 0.0 # No Cycle warning anymore!
# Move handle
cmds.move(0, 15, -10, "|ikHandle1")
cmds.refresh(force=True)
assert round(cmds.getAttr(a + ".rx"), 0) == -14.0
assert cmds.getAttr(a + ".rx") == 0.0 # Correct value now, assertion is thrown
On Wed, 15 Aug 2018 at 08:13 Joe Weidenbach <scdr...@gmail.com> wrote:
> I call it a bug, but it's been around a loooong time.
>
> Maya updates the ikHandles in a separate pass (there's just one solver for
> multiple handles), and I don't believe that script mode forces an update on
> the solver, where moving the root updates the DAG directly.
>
> It's one of the several places where the maya internals bypass the DAG so
> that they can work (IK handles don't put input connections on the joint
> rotations as that _would_ be a cycle, so the update needs to live in a
> callback somewhere, which isn't triggered by your code). Another place
> where things get weird is with editable animation curves (the ones where
> you've got an editable curve in the scene that updates the animation
> curves).
>
> Long story short, querying an attribute at a given time only works for
> direct DAG evaluation. Anything which operates outside of the DAG won't be
> updated when you call ```getAttr```. For those cases, you actually have to
> set the time and refresh the evaluation prior to the ```getAttr``` call:
>
> ```
> from maya import cmds
>
> # Build Scene
> # o o
> # \ /
> # \ /
> # o
> #
> cmds.file(new=True, force=True)
> a = cmds.joint(name="Root")
> cmds.move(0, 0, 20)
> cmds.joint()
> cmds.move(0, 0, 0)
> b = cmds.joint()
> cmds.move(0, 10, -20)
> handle, eff = cmds.ikHandle(
> solver="ikRPsolver",
> startJoint=a,
> endEffector=b,
> forceSolver=True,
> )
>
> assert cmds.getAttr(a + ".rx") == 0.0
> cmds.refresh(force=True) # Side note, this means your scene actually has
> to play through
> assert cmds.getAttr(a + ".rx") == 0.0 # No Cycle warning anymore!
>
> # Move handle
> cmds.move(0, 15, -10, "|ikHandle1")
> cmds.refresh(force=True)
>
> assert round(cmds.getAttr(a + ".rx"), 0) == -14.0
> assert cmds.getAttr(a + ".rx") == 0.0 # Correct value now, assertion is
> thrown
> ```
>
> On Wed, 15 Aug 2018 at 01:29 Marcus Ottosson <konstrukt...@gmail.com>
> wrote:
>
>> Hi all,
>>
>> TLDR; In Maya 2015-2018, getAttr returns stale value when used in
>> conjunction with an ikHandle
>>
>> I’ve ventured into the depths of querying attributes at an arbitrary
>> time..
>>
>> # Examplefrom maya import cmds
>> cmds.getAttr("myNode.translateX", time=5)
>>
>> And it has worked well so far, except for when IK is involved.
>>
>> from maya import cmds
>> # Build Scene# o o# \ /# \ /# o#
>> cmds.file(new=True, force=True)
>> a = cmds.joint(name="Root")
>> cmds.move(0, 0, 20)
>> cmds.joint()
>> cmds.move(0, 0, 0)
>> b = cmds.joint()
>> cmds.move(0, 10, -20)
>> handle, eff = cmds.ikHandle(
>> solver="ikRPsolver",
>> startJoint=a,
>> endEffector=b,
>> )
>> assert cmds.getAttr(a + ".rx") == 0.0assert cmds.getAttr(a + ".rx", time=1)
>> == 0.0 # Cycle warning!
>>
>> Suddenly, Maya throws a cycle warning. Why is that?
>>
>> Furthermore, the value returned isn’t right anymore.
>>
>> # Move handle
>> cmds.move(0, 15, -10, "|ikHandle1")
>> assert round(cmds.getAttr(a + ".rx"), 0) == -14.0assert cmds.getAttr(a +
>> ".rx", time=1) == 0.0 # Bad value, no warning..
>>
>> Where I should see -14.0 I get the previous value, 0.0. Why is that?
>>
>> And so the story goes, until..
>>
>> # Move root
>> cmds.move(0, 0, 21, a)
>> assert round(cmds.getAttr(a + ".rx"), 0) == -14.0assert round(cmds.getAttr(a
>> + ".rx", time=1), 0) == -14.0 # Correct
>>
>> Viola, all values are correct. Why is that?? :O
>>
>> This happens consistently across all versions of Maya, except 2015
>> doesn’t throw a Cycle Warning.
>>
>> Anyone stumbled upon this before?
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "Python Programming for Autodesk Maya" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to python_inside_maya+unsubscr...@googlegroups.com.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/python_inside_maya/CAFRtmODV_eKp2%3DJ3AdUufuDfBe2qojSWFjNpT9gJO1A6e9OLhg%40mail.gmail.com
>> <https://groups.google.com/d/msgid/python_inside_maya/CAFRtmODV_eKp2%3DJ3AdUufuDfBe2qojSWFjNpT9gJO1A6e9OLhg%40mail.gmail.com?utm_medium=email&utm_source=footer>
>> .
>> For more options, visit https://groups.google.com/d/optout.
>>
>
--
You received this message because you are subscribed to the Google Groups
"Python Programming for Autodesk Maya" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to python_inside_maya+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/python_inside_maya/CAM33%3Da4_ZRZS5VCzh1o8hgcJMLXOEvvgh0PWLv4YcMid9VntSQ%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
