Cleanup / Loading imported models: when "importURI" does have an effect? #380
Labels
Comments
|
@goto40 Thanks for the thorough analysis of the problem. I agree with you. I vote for moving the logic of I agree that this discussion will lead to code changes so let's leave it as an issue instead of a discussion. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Until now (textx 3.0), we have the
ImportURIscope provider base class which allows to use the special attribute "importURI" to load imported models: e.g., see this example.Specializations of
ImportURI(scope providers allowing to import models):How does the base class
ImportURIwork:model.pyby checking if the scope providerìs aModelLoaderinstance: here and here.ImportURIhas a specific way to detect imports. It scans the model for attributes namedimportURIto load models (this could be implemented in a different way, e.g., by using other attributes, but we do not do this any unittest).ImportURIhas a specific way to load models. They are stored in some special `_tx"-atttributes of the model (or the metamodel). This "special way" of loading models is used by our multi model scope providers or tool functions (see, e.g, http://textx.github.io/textX/3.0/scoping/#included-model-retrieval) but could be implemented differently.I would like to start a discussion if we should narrow the possibilities to load models into the model (here and here.):
ModelLoaderscope provider separately (which redundantly tries to load and reload all imported models if more than one such scope provider is registered)importURIattributes (fixed logic) and load all imported models (like we do it now in theImportURIscope provider).+min RREL strings or theFQN-scope provider).This a BIC in any case! We should carefully think about it some time... But I think, if done correctly, this BIC should have no (or nearly no) impact for "normal users".
I am looking forward to discuss this topic with the users of
textx! We can also move it to the discussion section, but since I expect some actions following this discussion I started it as issue... @igordejanovic feel free to move this issue...The text was updated successfully, but these errors were encountered: