<cmsmcq> Good morning, all.
ACTION (2022-06-21): Tom to set up a schematron demo
ACTION (2022-09-13): John to report how many Earley items are reduced
by the changed rewrite rules.
ACTION (2022-09-13): Steve to draft text (in HTML) to resolve the
erratum in issue #145
[Done; please check]
ACTION (2022-09-13): Norm to try out the stylesheet PI trick on the
XML version of the ixml grammar.
Status of implementations
Steven: v4 is up. Mostly code cleanup. Some better error messages. Warns about unreachable rules.
Norm: No change
Steven: I also published my code (using ixml)
… also working on my tutorial.
Status of testing and test suites
Norm: There were commits for a couple of additional ambiguous cases
Issue #145 Character class "LC" doesn't match our grammar
MSM: We agreed to change. We need to confirm the changes.
Norm: Errata is fine
Steven: I'm fine with Norm's version. My action remains to update a future version of the spec.
ACTION: Norm to move the erratum from 'Proposed' to 'Accepted'.
Norm: I would like to propiose spec build process
Spec build process
Norm: I propose moving sources around, and building our docs automatically, so we get to see diff markups.
… Using DeltaXML, for which I have a licence
MSM: Is it possible to inject a paragraph to acknowledge DeltaXML?
Norm: I added a next and previous change button
… all it means is that there will be a source directory
… for this change, edit the spec and make a pull request
… I'll take the action to incorporate the erratum and explain how it works
Bethan: If it's an erratum, it goes into a v1.1?
Norm: Yes, there will be a 1.1 for the change
MSM: 1.0 directory is sacred
Implementation reports (version)
John: I've updated and released a v1
… with drag and drop
John: Ambiguity can be a problem.
Issue #147 Are control characters allowed in an ixml grammar?
Norm: Control characters crash my parser.
… Because it goes through the XML
Bethan: Don't we already cover this?
MSM: Yes and no.
Norm: If so, can we make it more explicit?
MSM: Or otherwise if it doesn't, do we want to say that?
… I have a bias. My parser won't accept such grammars
John: Mine will.
Norm: Some ixml grammars have no XML representation
John: Some implementations can, some can't
MSM: I am reluctant to push hard.
… because it gets to talking about implementation choices.
… on the other hand, different behaviours is problematic
MSM: I don't want to push either way.
Bethan: What is the status of the XML serialisation of the ixml grammar?
MSM: some conforming grammars have no conforming XML representation
Bethan: XML representations are dependent on the ixml syntax?
… so the ixml format is logically prior to the XML format
MSM: I would like to check that claim
… yes you are right
Steven: Couldn't we just say implementations should warn?
… I would worry that future targets would add new restrictions to ixml
Steven: We could warn about grammars that wouldn't have an XML representation
MSM: I could live with that.
Steven: How about demanding that control characters always be encoded, and not allowed in strings?
Norm: Forbidding tabs would be good anyway, since they are not visible.
MSM: What is the set of Unicode characters that are not XML characters?
Norm: Differs between XML versions.
… 5th ed forbids c0 and c1
MSM: No, c1 is allowed
… So proposal: String literals in the ixml grammar should exclude all c0 characters. (Or or c0 and c1)
Norm: I could live with either of those
… I think it would be more consistent to forbid both
Steven: I'd like to have time for this to sink in
ACTION: Norm to write a proposal that covers the c0 c1 problem in strings
Issue #138 Home page links to ixml.xml and ixml.ixml
Norm: I'd like to have two links, one that displays, and one that downloads
MSM: You wanted the XML to behave similarly to the IE or Firefox behaviour
Norm: I will do this.
ACTION: Norm to make XML displayable in all browsers
MSM: Next month's meeting raising the question which is our normative timezone
Next call 15th November
Time: Same clocktime as this call
ACTION: Norm to inform group how build process will work