Hi @merkvilson, after more investigation here is the current behavior.
SymbolCache is only monitoring filename, so here is the condition to update the symbolcache file:

• The resource name is new.
• The resource name is removed.

Any modifications of the content of your resource files are not taking into account (aka not updated to the symbols cache)

So another solution for you could be to rename your resource files for each release of your plugin where the content of your file actually changed.

With that's said I've filled a feature request to our development team, to monitor change in files as well.
Cheers,
Maxime.

Hi @mp5gosu, You can react to BFM_INTERACTSTART and detect what key was pressed and in this case, and return True, so next message will not be dispatched and processed.

class MyDialog(c4d.gui.GeDialog):
    

    def CreateLayout(self):
        self.SetTitle("My Python Dialog")

        self.AddEditText(1000, c4d.CUSTOMGUI_DESCRIPTION, c4d.BFH_SCALEFIT | c4d.BFV_SCALEFIT, 100, 500)        
        self.AddEditText(1001, c4d.CUSTOMGUI_DESCRIPTION, c4d.BFH_SCALEFIT | c4d.BFV_SCALEFIT, 100, 500)
        
        return True

    def Message(self, msg, result):         
        if msg.GetId() == c4d.BFM_INTERACTSTART:
            bc = c4d.BaseContainer()
            if c4d.gui.GetInputState(c4d.BFM_INPUT_KEYBOARD, c4d.KEY_LEFT, bc):
                if bc[c4d.BFM_INPUT_VALUE] == 1:
                    if self.IsActive(1000): return True

            if c4d.gui.GetInputState(c4d.BFM_INPUT_KEYBOARD, c4d.KEY_RIGHT, bc):
                if bc[c4d.BFM_INPUT_VALUE] == 1:
                    if self.IsActive(1000): return True           

        return super(MyDialog, self).Message(msg, result)

If you have any question please let me know.
Cheers,
Maxime.

Before R18.048 MCOMMAND_JOIN will join all selected objects, children as well, but have some issue (sometimes when object matrice is shifted, stuff went wrong, or if the object is not a PolygonObject)
After R18.048 MCOMMAND_JOIN will join all children of the passed null.

Last note, we do not support more than year version, so if you have an issue in this particular version, please run your test and try some workaround if possible.

Cheers,
Maxime.

Hi @merkvilson, symbolcache is as the name says, a cache version of the symbols in order to speed up the startup of Cinema 4D.

So delete it each time you launch your plugin is really not recommended and could cause some issues (missing symbols and so on, since python is available after the generation of the symbolcache). While we do understand the issue, we don't have any workaround to it.
Regarding the issue, this is mostly due to utf8 handling as @mikeudin pointed.

The best way is to "educate your customer".

• Lunch Cinema 4D.
• Open the preference.
• Click on the bottom button "Open Preference".
• Go to the pref folder.
• Delete the symbolcache.
• Restart Cinema 4D.

Cheers,
Maxime.

Hi Passion3D as you discovered the behavior of MCOMMAND_JOIN has changed in R18.048.

You can find more information about it in this post. http://www.plugincafe.com/forum/forum_posts.asp?TID=13246&KW=&PID=52423#52423
So before R18.048, you have to pass a list of all objects.
After R18.048, you have to pass the null that holds all the hierarchy.

Cheers,
Maxime.

Hi @klints2, may I ask you why this is not perfect? Moreover here just a few corrections in your code.

• In Cinema 4D (and in general), try to avoid as much as possible global variables. In your case you can call documents.KillDocument(tdoc) just after the SaveDocument.
• Make use of join as the function should behave.

dirpath = "c:/usr/temp/"  # temp directory
full_dirpath = normpath(join(dirpath, (doc.GetActiveObjects(0)[0]).GetName() + ".obj"))

• Always check, you get object selected or you will get some errors in the console.

sobj = doc.GetActiveObjects(0)
if not sobj: return

Cheers,
Maxime.

@knekke said in Pipeline integration:

Has something changed in R20 regarding plugin load order or program_started execution time?
I used a startup script to change the renderer to redshift, but while it works in R19 it doesn't in R20.

@m_adam said in Pipeline integration:

(take care python_init.py is no more loaded in Cinema 4D R20).

C4DPL_PROGRAM_STARTED is working as before, only the python_init.py is no more called in R20 SP1 (This will be fixed in R20 SP2).

Cheers,
Maxime.

Hi @Chubak, first of all, welcome in the plugincafe community.

I would like to point you some rules on this forum regarding how to post questions in order to help us to help you!

• How to Post Questions especially the tagging part and take care to post on the correct category.
• Q&A New Functionality.

Don't worry it's your first post you couldn't know, so I've setup your topic correctly.

In Cinema 4D objects displayed in the ObjectManager are c4d.BaseObject, such a BaseObject are oftenly a PointObject which as this name define store point attribute (Used in Spline and in Polygon Object). This class allow you to reads and modify point position of an object with PointObject.GetAllPoints/PointObject.SetAllPoints.

To generates a noise, there is the c4d.utils.noise.C4DNoise class which provides you all you need to generate different kinds of noise.

If you have any questions, do not hesitate.
Cheers,
Maxime.

Hi Orestis,

Unfortunately, all helpers classes available in C++ such as KDTree, DistanceQueryInterface or either VoxelizationInterface are not available in Python. So the only reliable way is to iterates over all the points like so.

import c4d
import sys

# Main function
def main():
    start = time.time()
    if not op: return
    if not isinstance(op, c4d.PointObject): return

    # WorldPosition to Match
    posToMatch = c4d.Vector(25, -13, -88)

    # Retrieves points position in World Position
    opMat = op.GetMg()
    pts = [pt * opMat for pt in op.GetAllPoints()]
    if not pts: return
  
    # Iterates overs all list to find the closest point
    nearestDist = sys.float_info.max
    ptId = None
    for x, pt in enumerate(pts):
        dist = (pt - posToMatch).GetLength()
        if dist >= nearestDist: continue
        
        nearestDist = dist
        ptId = x
    
    print nearestDist, ptId


# Execute main()
if __name__=='__main__':
    main()

I have also tested to do some threading stuff, but due to the GIL in python, this does not bring any performance gain (even worse).
Cheers,
Maxime.

Hi @mikeudin, due to how our Classic Python API is built deepcopy does not work since our python objects only store pointer to our internal C++ object and not directly to other Python objects, so deepcopy fail in this regards since, the deepcopy operation, free and try to recreate these data, which is not possible since they are not PyObject but real C++ object but copy.copy does works, but since it's a shallow copy operation it only works for one level (e.g. a dict of dict of DescID will fail).

For more information about the difference please read Shallow vs Deep Copying of Python Objects.

However here a couple methods to copy a dict.

import copy
dict1 = {'1':dId_one,'2':dId_two}
dict2 =  copy.copy(dict1)

dict1 = {'1':dId_one,'2':dId_two}
dict2 =  dict1.copy()

dict1 = {'1':dId_one,'2':dId_two}
dict2 =  dict(dict1)

Cheers,
Maxime.