Issue with type-scoped contexts propagation

[kasei]

Test tc015 tests whether type-scoped contexts propagate by default or not. Based on the language in syntax document section 1.7, propagate should be false for type-scoped contexts. However, I can't find anywhere in the API document that implements this default for type-scoped contexts.

In section 4.1.1, there is non-normative text saying:

If context is not to be propagated, a reference to the previous context is retained so that it may be rolled back when a new node object is entered. By default, all contexts are propagated, other than type-scoped contexts.

However, in 4.1.2, propagatge is defaulted to true in the algorithm setup, and type-scoped contexts never change this default.

Have I missed something about how propagation is meant to be handled?

[gkellogg]

I think 11.2 of the Expansion algorithm should pass false for propagate, not true, as is done in step 11 of the Compaction Algorithm.

Note that the WG resolved that corrections to algorithmic text are not considered to be normative changes if they don't result in any changes to the expected test results.

[iherman]

This issue was discussed in a meeting.

• RESOLVED: Add to Expansion 11.2 that it should pass false for the propagate option for type-scoped contexts

View the transcript

Type-scoped context propagation
Rob Sanderson: #246
Gregg Kellogg: when a typed-scoped context is encountered,
… define context is called; the parameters passed to it are wrong
… the ‘propagate’ parameter should be ‘false’, but the default value ‘true’ is currently used
… The good news is that we do have tests for this corner case.
… The intention is clear.
Proposed resolution: Add to Expansion 11.2 that it should pass false for the propagate option for type-scoped contexts (Rob Sanderson)
Gregg Kellogg: +1
Rob Sanderson: +1
Benjamin Young: +1
Harold Solbrig: +1
Pierre-Antoine Champin: +1
Tim Cole: +1
Resolution #8: Add to Expansion 11.2 that it should pass false for the propagate option for type-scoped contexts
Ivan Herman: +1

[gkellogg]

@kasei, please let us know if changes in #252 satisfactorily address your concern.

[kasei]

Looks good, thanks.