Set Polygon/Edge Position

[Tekano]

ah do I overwrite the original find in array compound with this ? do I make both nodes into a compound or just connect the scalar and export?

ok I put them both in a compound and exported to user compounds 'Find in Array'

I now have two Find in Arrays listed , but when I try again the same thing I get


' WARNING : 3000-EDIT-AddICECompoundNode - Could not set value on [Find in Array 0].Value.Value_x
' WARNING : 3000-EDIT-AddICECompoundNode - Could not set value on [Find in Array 0].Value.Value_y
' WARNING : 3000-EDIT-AddICECompoundNode - Could not set value on [Find in Array 0].Value.Value_z

[iamVFX]

And now I have no idea what causes this.

I made a big mistake once. I've installed SI 2013 and used it.

I shouldn't have.

I hope Steve would explain us. He's the most smartest and experienced guy from the whole AD team now

[Tekano]

And now I have no idea what causes this.

I made a big mistake once. I've installed SI 2013 and used it.

I shouldn't have.

haha user error no doubt! all I wanted to do was to be able to procedurally scale a polygon with ICE Topo . what a journey this has been so far and I havent even come anywhere close yet

anyways to help future 2012 ICE Topo strugglers out I got these latest compounds to work by doing what iamVFX suggested. connecting a scalar with value 0.001 to the Find in Array node, putting this into a compound and exporting. The trick is to change the version number before exporting. so even though the compound list shows 2 find in array nodes it defers to the higher version. (but only after a restart not a refrsh compounds list) So. finally, after a considerable amount of hoops I have had to jump I now have basis to try and work out my stupid procedural polygon size effect both in 2013 & 2012.

Thanks Constantine for your patience!! will definately post any results and compounds made - if I ever get that far

oh and about screensnapr - its not your fault but if its still down tonight I will try and post your example ice tree images again in this thread.

cheers!

[xsisupport]

Hi Stephen, have you been followed this thread? I believe you have to have all of iamvfx's proceduraltopopack installed in user compounds beforehand - which includes a variant of the elusive 'build array from set for 2012'. as well as Guillaume's Polygons Particles addon. I have and do not get same errors as you on 2012 SP1 . Once all these are in place I then get the error described

' WARNING : 3000-EDIT-AddICECompoundNode - Could not find node : Find in Array


and ok if you can open it up in 2013 see where the missing nodes are meant to go then in 2012 re-connect them BECAUSE BACKWARDS COMPATIBILITY IS BROKEN but do we have to do this with every shared compound that uses these nodes? this is a joke right? I see what iamvfx is on about now with sharing compounds no longer working as it should. And for instance, at work we do not have 2013 installed yet. how the heck am I meant to fix a broken compound that works in 2013 (at home) but at work I cannot see how it is meant to work...?

bit of a conundrum eh!

Ok, I created a workgroup and installed all the addons and compounds you mentioned.
Now when I create a Init Islands Extrude node in an ICE tree, there are no errors or warnings.

PS I did make one mistake in my original answer: when I downloaded that rar with the Japanese characters in the file name, I extracted it and didn't change the folder name...that's why I couldn't create the compounds at first.

[Tekano]

ah I see, I just threw all of these compounds into one folder inside my user>data>compounds folder. I do not get error by default by just to bring that node into the ICE tree. do you make a cluster, get is_element from cluster and connect to init islands extrude? if you then see an extrusion then somehow something is different to my setup as I get a correct purple node but no extrusion. bringing in another node 'Islands extrude' generates an error in the logs. doing what iamvfx mentions with the find in array node fixes this error and the extrusion works.

I think the 'Elements to Island Index' node forces this error too. Im on 2012 SP1 btw.

cheers!

[xsisupport]

ah I see, I just threw all of these compounds into one folder inside my user>data>compounds folder. I do not get error by default by just to bring that node into the ICE tree. do you make a cluster, get is_element from cluster and connect to init islands extrude? if you then see an extrusion then somehow something is different to my setup as I get a correct purple node but an error in the logs. doing what iamvfx mentions with the find in array node fixes this error.

I think the 'Elements to Island Index' node forces this error too. Im on 2012 SP1 btw.

cheers!

No, I didn't see those repro steps in the bug you logged. Were they in this thread earlier on? Sorry, I missed them.

[Tekano]

sorry Steve - yes I was still exploring various issues when I logged that bug and was not sure of the cause or the repro steps. it still stands and is only apparent in 2012 SP1 and not 2013 and seems to be related to the 'Find in Array' by default having its epsilon set to 0 and not 0.001 - or something....?

[xsisupport]

sorry Steve - yes I was still exploring various issues when I logged that bug and was not sure of the cause or the repro steps. it still stands and is only apparent in 2012 SP1 and not 2013 and seems to be related to the 'Find in Array' by default having its epsilon set to 0 and not 0.001 - or something....?

Now I'm in 2013 SP1, and I can repro the "Find in Array" warning.
Create a mesh, create an ICE tree on the mesh, and add Elements to Island Index to the ICE tree:

Code: Select all

# WARNING : 3000-EDIT-AddICECompoundNode - Could not find node : Find in Array
Application.AddICECompoundNode("Elements to Island Index", "grid.polymsh.ICETree")

If I look at "Elements to Island Index.xsicompound", I see Find in Array appears to be a compound:

Code: Select all

			<node type="Find in Array" version="1.1" index="0">
				<param name="Value.Value_x" type="4" value="0.000000"></param>
				<param name="Value.Value_y" type="4" value="0.000000"></param>
				<param name="Value.Value_z" type="4" value="0.000000"></param>
				<param name="Array.Array_x" type="4" value="0.000000"></param>
				<param name="Array.Array_y" type="4" value="0.000000"></param>
				<param name="Array.Array_z" type="4" value="0.000000"></param>
				<portdef name="Value" type="16" structure="1" group="-1" instance="-1" port="-1"></portdef>
				<portdef name="Array" type="16" structure="2" group="-1" instance="-1" port="-1"></portdef>
				<datablob category="ui_infos">
					<prop name="posx">-449</prop>
					<prop name="posy">496</prop>
					<prop name="state">Normal</prop>
				</datablob>
			</node>

where is this compound supposed to come from?

The built in Find in Array node looks like this in an xsicompound file:

Code: Select all

			<node type="FindInArrayNode" index="0">
				<param name="epsilon" type="4" value="0.000000"></param>
				<portdef name="value" type="1035263" structure="1" group="0" instance="0" port="0"></portdef>
				<portdef name="array" type="1035263" structure="1" group="0" instance="0" port="1"></portdef>
				<portdef name="epsilon" type="4" structure="1" group="0" instance="0" port="2"></portdef>
				<datablob category="ui_infos">
					<prop name="posx">-658</prop>
					<prop name="posy">-61</prop>
					<prop name="state">Normal</prop>
				</datablob>
			</node>

[iamVFX]

where is this compound supposed to come from?

Here's the original one. Nothing fancy, as I said a scalar and a standard Find in Array node. There is no Value_x/y/z and Array_x/y/z in its code

[xsisupport]

where is this compound supposed to come from?

Here's the original one. Nothing fancy, as I said a scalar and a standard Find in Array node. There is no Value_x/y/z and Array_x/y/z in its code

Thanks

[xsisupport]

where is this compound supposed to come from?

Here's the original one. Nothing fancy, as I said a scalar and a standard Find in Array node. There is no Value_x/y/z and Array_x/y/z in its code

Thanks

Ok, now I'm in 2012 SAP and I don't get any errors or warnings and things seem to work.

[xsisupport]

Well whomever is to monetize the ICE compounds it won't be Autodesk management, they don't appear to know the difference between ICE and 'scripts'. Honestly I don't believe this to be happening Constantine , because:-

a) they can't squeeze anymore money than they already have out of our insignificant Softimage user base .
b) Like I said its easily enough fixed and has been by re-connecting the missing (version) of compound, OK AD have not done that in a public press release but Support has explained the issue and how to deal with it.
c) there has been much bigger bugs and fuckups in the past, they didnt accept responsibility or apologise then nor will they ever. this is the nature of multi million lines of code software dev - shit happens

Also the original author of the newer version of compound is on here and the SI-List to help even though he's been led to a much posher paddock. Personally I would rather he be around and assist than hold a grudge and ignore. I really don't see a trend happening to make ICE compounds version specific. but, who knows, given some of the ludicrous decisions made on behalf of Softimage by AD then perhaps you will be right and it will be thier next step in the saga. cue dramatic music

Build Array from Set not a compound, it's a compiled node.

It was custom node in 2012. In 2013 it was converted to a native node to add functionality [that couldn't be added to a custom node] like multi-threaded writes and support for topology and locator data types.

The original custom node is still in the distribution, so backwards compatibility still works (old stuff can be loaded into the current version).

Forward compatibility (the ability of older versions to load stuff saved by newer versions) is not supported, at least not as far as I know or remember.

If you want to build compounds that can be loaded by older versions, you can use the custom node, either by running this line

Code: Select all

Application.AddICENode("BuildArrayFromSet", "grid.polymsh.ICETree")

or by using a compound that contains the node created by the above line.

[Tekano]

so... is this a bug, a version problem or a user error...? Have just tried this again in 2013 and 2012SAP back home and apart from a refresh issue (have to drag the ice tree up the stack then down again to make the extrusion happen) it works perfectly fine without this extra 'find in array' compound. is this just a 2012SP1 phenomenon then or is it the way the compound folders are set up between different installation paths and workgroups? sure is a strange one.

oh and about the 'Forward compatibility' perhaps it was an assumption by all users, myself included (because ICE compounds dont have version 7 or 2012 whatever added on to their names) that ICE compounds Were meant to be forward compatible. In the past Ive transferred ICE trees many times as compounds from newer to older versions of Softimage.

[xsisupport]

so... is this a bug, a version problem or a user error...? Have just tried this again in 2013 and 2012SAP back home and apart from a refresh issue (have to drag the ice tree up the stack then down again to make the extrusion happen) it works perfectly fine without this extra 'find in array' compound. is this just a 2012SP1 phenomenon then or is it the way the compound folders are set up between different installation paths and workgroups? sure is a strange one.

yeah I had some problems too, at first the extrusion didn't want to work.
I did need that extra Find in Array compound, because the other compound referenced it.

oh and about the 'Forward compatibility' perhaps it was an assumption by all users, myself included (because ICE compounds dont have version 7 or 2012 whatever added on to their names) that ICE compounds Were meant to be forward compatible. In the past Ive transferred ICE trees many times as compounds from newer to older versions of Softimage.

A compiled node is not a compound. A new node introduced in a later version of Softimage won't work in an older version of Softimage. And in the case of Build Array from Set, we have an exceptional case where we created a new node (but kept the older one in the distribution).

In most cases, older versions of Softimage can use compounds from the current version, as long as the compounds don't use new compiled nodes.

[Tekano]

and eventually got the effect working in the end. Had to use repeat nodes only because from the examples I had working it was easier to disect the context jumping. I can now 'procedurally' control the sharpness of each randomly generated extrusion and the size of each is controlled by a random amount multiplied by distance to a null and also are growing over time. If anybody is interested or is struggling in the future I can package this up into a compound.

cheers!

[iamVFX]

Oh no, two repeat nodes...

By looking at the complexity of your tree... my intuition tells me it can be solved easier