[code] ComponentDefinition-delete (another version!)

[TIG]

Alex

I first tried your method to erase the definition.instances and then erase all of the definition's entities, so the empty definition is gone from the browser.
This works manually, as SketchUp removes an empty definition from the component-browser.
However, when I used the ...entities_erase(...) method in a script it emptied the definition BUT it didn't remove it from the browser - if you then picked it in the browser = BugSplat !

I can''t see any difference between your version and my early attempt - only slightly different syntax, but same result - HOWEVER yours works ! Since yours is better than mine - not needing convoluted purging - I'll withdraw mine and adjust my related tools...

.

[thomthom]

I think this method should delete any existing instances belonging to the definition as it doesn't remove the definition when there's instances around.

EDIT: hmm... this is odd. When I run it through the console on a fresh model: Sketchup.active_model.definitions[0].delete it deletes Sang and removes it from the Component Browser.

But when I implement it into my working plugin, the component remains in the browser and will bugsplat if I try to place it... hm...

[AlexMozg]

@thomthom said:

But when I implement it into my working plugin, the component remains in the browser and will bugsplat if I try to place it... hm...

Therefore I foresaw additional possibility of the use of this method.
Use this method with the argument false, if you do not want to begin a new operation and correct forming "undo".
Essence consists in correct determination of start [ start_operation ]
and end [ commit_operation ] of operation!
It most important, bugsplat is otherwise possible!

def my_new_plugin
   model = Sketchup.active_model
   model.start_operation "New Operation"
   definitions = model.definitions
#....................................
#....................................
#....................................
   definitions[0].delete(false) if definitions.to_a[0]
   definitions[1].delete(false) if definitions.to_a[1]
   definitions[2].delete(false) if definitions.to_a[2]
#....................................
#....................................
#....................................
   model.commit_operation
end

If I understood correctly...

P.S
When I tested, bugsplat repeated oneself too often...
I maximally tried to foresee all possibilities of bugsplat appearance and prevent them.

By the way, my example at the end of file "AT-ComponentDefinition-delete.rb" it is possible to delete or comment out if he will not be used.

[thomthom]

I removed the example code. I also didn't use a start_operation on my own code.

[AlexMozg]

Bugsplat also can be at the use of method of "erase!" between "start_operation" and "commit_operation".

[thomthom]

Another thought: Not too sure if it's good that there's two differents methods available with the same name. What if one plugin implements one method and another plugin implements the other. Since the method integrates with the SU native class there's no way to avoid conflict.

Can we decide on one?

[AlexMozg]

You can use a not method (ComponentDefinition.delete), but idea:

definition.entities.erase_entities definition.entities.to_a

between
model.start_operation ("")
and
model.commit_operation

Sometimes so simpler...

[thomthom]

@alexmozg said:

You can use a not method (ComponentDefinition.delete), but idea:

definition.entities.erase_entities definition.entities.to_a

between
model.start_operation ("")
and
model.commit_operation

Sometimes so simpler...

So that's the core of it? The commit operation initiates SU to purge the empty definition? I couldn't really work out how your code worked. So I suppose that something in my code interferes with this?

[TIG]

@thomthom said:

Another thought: Not too sure if it's good that there's two different methods available with the same name. What if one plugin implements one method and another plugin implements the other. Since the method integrates with the SU native class there's no way to avoid conflict.

Can we decide on one?

I'll remove mine from the forum now...
As AlexM points out using the:

model.start_operation("Delete Definition")
  definition.entities.erase_entities(definition.entities.to_a)
model.commit_operation

it has been discovered how to do what we want anyway without a new method at all...

[Frankn]

@tig said:

it has been discovered how to do what we want anyway without a new method at all...

Sorry to bring back a very old post/topic but I'm trying to delete definitions and I tried using the code AlexMozg wrote but I'm having problems with it and to be honest I don't really understand how the code works so I'm not sure what's going wrong.

The error I get is... Error: #<TypeError: reference to deleted ComponentDefinition>

Is there another way to do this if not what should I be looking for in regards to the problem I'm having?

Thanks.

[TIG]

If you want to delete ALL definitions in a model [use with care!]
A one-liner like:

m=Sketchup.active_model;m.start_operation('!');m.definitions.each{|d|d.entities.clear!};m.commit_operation

does it.

BUT if you want to delete just the definitions of Selected component-instances [and of course the instances will vanish too] use:

m=Sketchup.active_model;m.start_operation('!');ds=[];m.selection.grep(Sketchup;;ComponentInstance).each{|i|ds<<i.definition};ds.uniq.each{|d|d.entities.clear!};m.commit_operation

this way you make an array of all affected definitions and process those just once each - otherwise, if you have two instances selected of the same definition, then the second instance processed has an invalid 'deleted definition' so it will fail etc...

[Frankn]

Thanks TIG,

The second code is what I'm looking to do but I'm getting this error...
Error: #<NoMethodError: undefined method `confirm_operation' for #Sketchup::Model:0x8e9a4b4>

Also I tried selecting the definition I want to delete with this but it doesn't seem to work

@model.selection.add(#<Sketchup;;ComponentDefinition;0xbcd77a0>)

I forgot to mention that if I manually select the instance it does delete it and the parent from the components viewer but leaves the subcomponents.

I'm still trying to figure out what your code does, this is a little over my pay grade!

[Frankn]

Ok so after a beer break I figured out why I couldn't select the instance... I was trying to select the definition!

When I comment out m.confirm_operation everything seems to work but obviously there's something wrong with the code or I wouldn't need to comment it out.

[Frankn]

Here's something interesting/weird that's happening...

After I delete the definition I iterate through model.definitions just to make sure everything was getting deleted but I found that the parent instance isn't getting deleted and the names of the subcomponents aren't getting deleted either... which causes the new component being created to be renamed with #1 at the end...

Any idea why?

[TIG]

It's ' commit_operation' NOT ' confirm_operation' - my typo - it's now corrected in my earlier code...
You can manually select objects including instances, and just the instances' definitions will be processed.
You can only add 'entities' to a Selection [like geometry or instances] in code.
NOT a 'definition' - which is NOT an 'entity'...

[Frankn]

@tig said:

It's ' commit_operation' NOT ' confirm_operation' - my typo - it's now corrected in my earlier code...
You can manually select objects including instances, and just the instances' definitions will be processed.
You can only add 'entities' to a Selection [like geometry or instances] in code.
NOT a 'definition' - which is NOT an 'entity'...

Thanks you sir! That works perfectly! I should of caught that but I just figured it was a command I didn't know of yet!