Hi all,

I have used input.width, input.height, and [exists parent.input] in many
Gizmos/Groups without issue.  Recently, I made a Group that detects the
incoming resolution and uses that to enable/disable the desired Write node.
 In simple cases, it works fine.  However, once the comps got a bit more
complex (especially RotoPaint heavy), I saw 2 issues on our render nodes.
 1.) Nuke crashed immediately after loading all plugins.  or  2.) Nuke
completes the job but hangs and doesn't exit properly.  Case 1 doesn't
happen for all frames, but is pretty consistent for the frames that do fail.
  Case 2 usually is for every frame of the job.  If I remove the input.*
expressions, the same .nk file renders without problems.

I am going to send an example to Support, but thought I'd ask here.  Are
there times when it is not safe to use input.width / input.height / [exists
parent.input]?  or cases when you might expect inconsistent results?


A simple example Group is below.  Switching between HD and 2k input changes
which Write is active on the inside.



Thanks,
Paul












set cut_paste_input [stack 0]
version 6.2 v3
push $cut_paste_input
Group {
 name Group1
 selected true
 xpos 539
 ypos 229
}
 Input {
  inputs 0
  name Input1
  xpos 628
  ypos -147
 }
 NoOp {
  name Res
  selected true
  xpos 628
  ypos -83
  addUserKnob {20 Resolution}
  addUserKnob {6 hd +STARTLINE}
  hd {{"\[exists parent.input] && parent.input.width == 1920 &&
parent.input.height == 1080" x1 0}}
  addUserKnob {6 _2k +STARTLINE}
  _2k {{"\[exists parent.input] && parent.input.width == 2048 &&
parent.input.height == 1556"}}
 }
set N994be40 [stack 0]
 Output {
  name Output1
  xpos 628
  ypos 182
 }
push $N994be40
 Write {
  name _2kWrite
  xpos 445
  ypos 63
  disable {{!Res._2k}}
 }
push $N994be40
 Write {
  name HDWrite
  xpos 561
  ypos 62
  disable {{!Res.hd}}
 }
end_group
_______________________________________________
Nuke-users mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users

Reply via email to