BaseModelica prerelease suggestions

[olivleno]

Collection and discussion of suggestions from BaseModelica interested.

[HansOlsson]

From Dassault Systèmes it is mostly Christoff Bürger that will work on it, but one blocking issue I heard (which had been previously discussed) is that the syntax (ignore the semantics at the moment) is close to Modelica without being a strict sub-set.

The closeness makes it tempting to share a parser, but the differences mean that there's then a risk that changes for the Base Modelica parser will negatively impact the Modelica parser. That is a risk that makes prototyping hard to sell internally (even if it "just" impacts the error messages).

[casella]

Do we have a place where all the new syntax constructs are listed?

[henrikt-ma]

It ended up being @qlambert-pro who took care of setting up our Base Modelica parsing, but from what I've heard the syntax overlap was something that our parser generator was excellent at handling. The main extra cost appears to be the need to fine tune parser error messages for Base Modelica, but as far as I understand we already have a Base Modelica parser without any degradation of error messages for full Modelica. Fine tuning the parser errors for Base Modelica is something we can look into once the grammar is settled.

[henrikt-ma]

Generally speaking, I don't think having a single issue to collect prerelease suggestions is going to work. Instead, new issues or pull requests with the MCP0031 label should be opened for each concern. This way, we will be able to have well structured conversations for each suggestion.

Therefore, I suggest that we close this issue before important comments regarding specific suggestions end up here.

[olivleno]

Therefore, I suggest that we close this issue before important comments regarding specific suggestions end up here.

This was intended as a discussion thread to collect the questions of the newcomers.

I'd suggest to leave it open until our next meeting in September.
For relevant topics new issues with an appropriate title should be opened and linked.

[christoff-buerger]

I added a separate ticket about BaseModelica being a subset of Modelica (see #3569). From our perspective, this is an important strategic decision -- we like BaseModelica to be a proper subset of Modelica.