Hi @tdapper thanks for reaching out us.

With regard to your comment, although I confirm that you don't have to put them in a separate framework but it's actually recommended to grant good design and straightforward reusability.

For the sake of clarity, find below the files used to create a project which includes both the declaration and the implementation of the custom error together with the code using it. The example is derived from the code presented on GitHub.

project/projectdefinition.txt // Supported platforms - can be [Win64;OSX] Platform=Win64;OSX // Type of project - can be [Lib;DLL;App] Type=DLL // API dependencies APIS=cinema.framework;mesh_misc.framework;math.framework;crypt.framework;python.framework;core.framework; // C4D component C4D=true stylecheck.level=3 // must be set after c4d=true stylecheck.enum-registration=false stylecheck.enum-class=false ModuleId=net.maxonexample.support.single_plugins.PC12316 source/PC12316_CustomErrorImpl.cpp #include "PC12316_CustomErrorInterface.h" // This example shows the implementation of a custom error type. class PC12316_CustomErrorImpl : public maxon::Component<PC12316_CustomErrorImpl, PC12316_CustomErrorInterface> { // use ErrorObjectClass to implement basic error functionality MAXON_COMPONENT(NORMAL, maxon::ErrorObjectClass); public: maxon::Result<void> CopyFrom(const PC12316_CustomErrorImpl& src) { _errorCode = src._errorCode; return maxon::OK; } public: MAXON_METHOD maxon::String GetMessage() const { return FormatString("Custom error code is @", _errorCode); } // custom methods MAXON_METHOD void SetCustomErrorCode(maxon::Int errorCode) { _errorCode = errorCode; } MAXON_METHOD maxon::Int GetCustomErrorCode() const { return _errorCode; } private: maxon::Int _errorCode; ///< error code value }; // register all the non-static methods in the implementation MAXON_COMPONENT_OBJECT_REGISTER(PC12316_CustomErrorImpl, PC12316_CustomErrorObject); source/PC12316_CustomErrorInterface.h #ifndef PC12316_CUSTOMERRORINTERFACE_H__ #define PC12316_CUSTOMERRORINTERFACE_H__ #include "maxon/object.h" // This example shows the declaration of a custom error type. // The custom error is able to store a custom error code. // --------------------------------------------------------------------- // Custom error class that stores an error code. // --------------------------------------------------------------------- class PC12316_CustomErrorInterface : MAXON_INTERFACE_BASES(maxon::ErrorInterface) { MAXON_INTERFACE(PC12316_CustomErrorInterface, MAXON_REFERENCE_COPY_ON_WRITE, "net.maxonexample.example.errors.PC12316_customerror"); public: MAXON_ADD_TO_COPY_ON_WRITE_REFERENCE_CLASS( void Create(MAXON_SOURCE_LOCATION_DECLARATION, maxon::Int errorCode) { #if API_VERSION >= 20000 && API_VERSION < 21000 * static_cast<typename S::DirectlyReferencedType::ReferenceClassHelper::type*>(this) = S::DirectlyReferencedType::ReferenceClassHelper::object::GetInstance() (); #elif API_VERSION >= 21000 *static_cast<typename S::DirectlyReferencedType::Hxx1::ReferenceClass*>(this) = S::DirectlyReferencedType::Hxx1::ErrObj::GetInstance()(); #endif typename S::DirectlyReferencedType::Ptr e = this->MakeWritable(false).GetPointer(); e.SetLocation(MAXON_SOURCE_LOCATION_FORWARD); e.SetCustomErrorCode(errorCode); } ); // custom methods // --------------------------------------------------------------------- // Stores an custom error code. // --------------------------------------------------------------------- MAXON_METHOD void SetCustomErrorCode(maxon::Int errorCode); // --------------------------------------------------------------------- // Returns the stored custom error code. // --------------------------------------------------------------------- MAXON_METHOD maxon::Int GetCustomErrorCode() const; }; #include "PC12316_CustomErrorInterface1.hxx" #include "PC12316_CustomErrorInterface2.hxx" #endif /* PC12316_CUSTOMERRORINTERFACE_H__ */ source/PC12316.cpp #include "c4d.h" static const Int32 PluginID = 99912316; #include "PC12316_CustomErrorInterface.h" // Dummy test function maxon::Result<void> TestFunction(maxon::Int * val); maxon::Result<void> TestFunction(maxon::Int * val) { iferr_scope; if (!val) return PC12316_CustomError(MAXON_SOURCE_LOCATION, 4242); ApplicationOutput("Value is @", *val); return maxon::OK; } // Command to test the custom error class CustomErrorExample : public CommandData { public: #if API_VERSION >= 20000 && API_VERSION < 21000 Bool Execute(BaseDocument* doc) #elif API_VERSION >= 21000 Bool Execute(BaseDocument* doc, GeDialog* parentManager) #endif { iferr_scope_handler { return false; }; if (!doc) return false; iferr (TestFunction(nullptr)) ApplicationOutput("Error: @", err); maxon::Int a = 100; iferr (TestFunction(&a)) ApplicationOutput("Error: @", err); return true; } static CommandData* Alloc() { return NewObj(CustomErrorExample) iferr_ignore("Unexpected failure on allocating CustomErrorExample plugin."_s); } }; Bool RegisterCustomErrorExample(); Bool RegisterCustomErrorExample() { return RegisterCommandPlugin(PluginID, "PC12316_CustomErrorExample"_s, PLUGINFLAG_COMMAND_OPTION_DIALOG, nullptr, ""_s, CustomErrorExample::Alloc()); } Bool PluginStart() { if (!RegisterCustomErrorExample()) return false; return true; } void PluginEnd() { } Bool PluginMessage(Int32 id, void* data) { return true; }

A final minor note on register=true: since R21 it's not needed anymore because in former releases it was needed to specify when a module was supposed to run in pure legacy-mode but nowadays modules can be either hybrid (classic + maxon API) or maxon API-only.

Hoping this makes using our Error Handling system more clear and easy to integrate in your products, give best and don't hesitate to ask for more.

Best, R

@C4DS

Hi,

no need to be sorry. Actually I am sorry for the (sort of) misinformation. But I have had the same problem and this was my conclusion for other node types (I just quickly reconfirmed it with a user data static text). Maybe DescriptionTooldata behave differently? Anyways, good luck 😉

Cheers,
zipit

@m_magalhaes
Hi Manuel,

Thank you for giving me some time to check this out.
So after a bit of testing and changing the code to work with Cinema I cannot replicate the issue. As you mention it seems to be related to Pyblish.

I did find interesting that when running the code in Cinema without Pyblish you can't see the active frame running through the timeline, while you do see it when running through Pyblish. Perhaps that's the correlation with the crash.

I will close this as solve, for the fact that it does not crash in the Cinema context.

Thanks again!

Andre

@Swinn is None did the trick. Thanks. 🙂

Hi,

thanks a lot. I've opened a bug entry for that one. It will be fixed in a futur release.

Cheers,
Manuel

@m_adam
Dear Maxime thanks for everything!
But, unfortunally now that I almost finished this part I realized that all this scripting doen't give me the desired procedural result. It just changes the values once but if i change the values of the PARENT object the OBJECT_A (child) is not updated...
Anyway I have to approach it in a different way.

I think I should tag this thread as "SOLVED" and in case I need again help i create a new one.
Please advise me Maxime.

And for one more time THANK YOU both , Maxime and Zipit

Hi @codysorgenfrey, thanks for reaching out us.

Glad to hear that the issue has been solved on your end: well done! If there's nothing left please mark the thread as Solved.

Cheers, R

@m_adam Thank you for reply. I would probably send my code to your email to try to figure it out better.
What i see so far is that in renderer, fields returns 0 for each point they sample.
I have a function which samples fields and accepting a result to each vertex. For example imagine extrude with fields. In viewport i see extruded points, but in renderer everything is flat, like there is no field or it is so far to make any extrusion and returns 0. for each point. Second is more likely due to my tests.

As said previously this is not possible in the execute method to add an undo event because this execute method is called during the execute.

But if I understood correctly you have an UserData attached then this UD drive children parameters. If that's the case you can react to this parameter change (coming from the GUI which is a Main Thread thing) and then apply the undo.
You can find an example in https://developers.maxon.net/forum/topic/12479/update-button/3

Cheers,
Maxime.

@zipit

hm yes this sounds too complicated overall, particularly point 3 there could so much happen on the os side.
A gui with more options like renaming the hole texturename or a replace option of certain words could be the way forward. Instead of undo it would be easy then to rename the texture with he old name etc.
for now im happy how this works and let this solved.

Much Thanks!
Cheers,
moko

@SolarPH said in Python NODES - Shouldn't really they can change the scene?:

Edit: One thing though, this code kinda has it's own holes, because I encountered a problem ...

I never said that it is perfect, I just implemented, what I thought you were after. If you want to find all file paths where the file name matches the file name part of a given preset URL, you could use os.walk. Something like this:

def find_preset_url_candidates(url): """Find all OS file paths for which the file name matches the file name of a given MAXON preset url. Args: url (str): The url in the MAXON 'preset' scheme to resolve. Returns: list[str]: The resolved preset paths. Raises: ValueError: When the argument preset_url is not a valid preset url. """ # Half-heartedly sort out some malformed preset urls. url = str(url) if (not url.startswith("preset://") or not url.endswith(".lib4d")): msg = r"Not a preset url: {}." raise ValueError(msg.format(url)) # Split the url into the scheme-path and filename-extension parts _, preset_file = os.path.split(url) candidates = [] # For each preset path walk that path and test if any file matches the # file name of our preset file name. for preset_path in PRESET_PATHS: for root, folders, files in os.walk(preset_path): if preset_file in files: candidates.append((os.path.join(root, preset_file))) return candidates

My example does not deal with any fuzzy path searches / does not sort its output by the edit-distance (you would have to do that yourself). To test successfully against something like testrig0.lib4d you would have to implement a Hamming distance or Levenshtein distance to sort out what is close enough. For your example you would need the latter. Or you could just use a regular expression if you do not care about the edit distance.

Cheers,
zipit

@PluginStudent

You mentioning the BITMAPBUTTON and with the help of ExampleSculptGrabBrush::GetDDescription from the cinemasdk I was able to make it work.
Thanks.

Yes, I use the auto way.
I thought that using ScrollArea() it would only display that part (no need to draw the other parts).
So, fill the virtual UA once and then display parts of it using ScrollArea().

But if "you still need to draw completely the GeUserArea with all the DrawBitmap", there will be no speed benefit.

I will try it under R21 and will read "Profiling Python Plugins".

Thank you, Pim

24fbc034-a80c-42f0-acb3-77deffeadaec-image.png

It's actually due to a lower number of bits used to store the floating value.

@m_adam @zipit @PluginStudent
OK, thank you guys!

@zipit @m_adam

Apologies for the late response. Thank you for the confirmation both.
I will close this for now.

Looking at the doc in R19:

5fdf5fb0-1327-4589-b6b6-d68af8869c11-image.png
821d97fa-e103-4cca-84b1-85f0f50304f0-image.png

So I'm afraid there is no simple solution that will work on each version maxon::String being introduced in R20, you will need to adapt your code.

Cheers,
Maxime.

Hi @JuicyJuggles , I apologize for coming late here but luckly you've been already provided in the meanwhile of some good guidance from @PluginStudent and @zipit .

Rest assured that the API documentation is, by far, the best place where to look for answers although I can see that, due to huge number of covered topics, it could take time to get used to browse it effectively.

Last but not least, if the issue is addressed, don't forget to mark the right post as correct answer or if it's not possible to identify a specific one, to set the whole topic as Solved

Cheers, Riccardo

Thanks @m_adam for the insights!

Works like a charm!!

Cheers,
Lasse

Hi sorry to jump on this topic so late, and I agree that if you could do what you want in a Python Scripting tag its preferred, the mains issue with xpresso is that its a whole system that assumes certain things and was really not designed to be embedded within a generator.

So the issue about the AddNode, not working is because user data things, as an object xpresso node, is something dynamic (they will change according to the linked object) and I found no way except via the UI (by asking a UI Redraw or a whole scene execution) to refresh these data so the GvPort can be created.

But since you are in a Generator you don't have access to the GUI since all GUI operation should be done in the Main thread see Threading Information.

Cheers,
Maxime.