I've been chasing down something for several days now, and after carefully
reviewing all the available documentation, I can't make out what the heck Nuke
is doing. It *appears* that it is randomly constructing spurious instances of
Op's.
Here's my setup. I took the simplest node I could think of, the 'Mirror' demo
Iop. I added print statements to the constructor, destructor, and the
_validate call.
I fire up 6.3vx Nuke and clear the buffers and the file cache. I then hook a
Mirror node up to s simple sequence, and I hook the viewer to the Mirror node.
I then stepped one frame to the right and one to the left. here is what I see
in the console:
constructor this: 8584C70
_validate this: 8584C70
Input frame: 309
Input view: 1
Output frame: 309
Output view: 1
_validate this: 8584C70
Input frame: 309
Input view: 1
Output frame: 309
Output view: 1
_validate this: 8584C70
Input frame: 310
Input view: 1
Output frame: 310
Output view: 1
constructor this: 13A05DE0
What the *HECK* is Nuke doing, constructing a *second* Mirror Iop on a
single-view Node?
Note that no _validate calls follow because, of course, the prior frame is
already in the cache. Also note that subsequent movement of the Viewer
actually causes requests to be made to the first instantiated 'Mirror' Iop.
But if you tend to initialize things in your Constructor, as I usually do, this
can really mess you up!
Any clue, anyone? Am I missing something, or is this a bug?
Steve
Oh.. and P.S.? Why is it calling _validate twice on exactly the same data?
Something bizarre with the Viewer I expect.
_______________________________________________
Nuke-dev mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev
