Meeting minutes
Outstanding Actions
ACTION (2021-08-004): Steven to create list of existing requirements,
so that we can identify when a new one comes.
<trackbot> Sorry, but no Tracker is associated with this channel.
Steven: Mostly done.
ACTION (2021-08-005): Tom and MSM to come with a pragma proposal.
<trackbot> Sorry, but no Tracker is associated with this channel.
Steven: Continues, ongoing
<Tom> Proposal for pragma: https://
Tom: The proposal is being worked up on github. Participation and comments are welcome.
MSM: The draft works nicely, I feel worried about complicated cases, with big pragmas.
… I can imagine a better proposal, but I'm not sure any such thing exists, or how it would work.
ACTION (2021-10-001): Steven to draft a mediatypes proposal (see
https://
<trackbot> Sorry, but no Tracker is associated with this channel.
Status of implementations
Steven: I was v happy that the server stayed up during the tutorial
Steven: I have rewritten the bootstrap parser, and am rewriting the Earley one, as prep for the C version.
Norm: I'm working on one, extending a parser that I have found. On a back burner at the moment.
Tom: The same as usual. Thinking about making progress on JayParser, but nothing new to report
… do I rewrite to optimise the levels of recursion, or do I rethink the approach?
MSM: Aparecium -- Declarative Amsterdam forced me to work on it. It survived the demos.
… the exercises in the tutorial all worked.
… some debugging still needed.
… I'm focussing on testing.
John's experience
John: XPath is an interesting and big testcase.
… I'm converting the EBNF to ixml
… using grammar adjustment to get rid of unwanted intermediate nonterminals
… having a problem with dealing with reserved words.
… is XPath too big of a problem?
MSM: I think the grammar is unambiguous, because the keywords are not reserved.
… but Pascal where you have reserved words, you may get ambiguity.
… but ixml doesn't have that problem.
MSM: Question about the ! in your example. Can you describe it's meaning?
MSM: A nice use case for pragmas.
John: It takes the first alt, and won't serialise it, but the second does.
Bug reports, change requests, etc.
Refactoring of grammar
MSM: not in github, just a report.
Steven: I didn't make the change since your usecase was convinving.
… I made the other changes.
root rule
Steven: I think we decided to keep it as is.
Strings proposal
Steven: Not to allow strings to split over lines.
MSM: in favour
Steven: Anyone against?
[No}
Steven: Consensus.
Renaming proposal
Steven: A future change, but I hear that people have the use case.
… You recognise it with one rule, and serialise with a different name
Tom: Pragmas.
MSM: A comment on the syntax, (and not with a pragma), the mark being moved to the other end of the nonterminal.
Steven: The ^ is followed by the name that will be used for the serialisation.
Removing ^ from tmark
Steven: You were against it.
… It does no harm there, I just don't think it will ever be used ever
Tom: I can't care about it.
… Don't fix what ain't broke.
Consensus: leave as is.
Testing and Test Suite
MSM: A reminder, when we first discussed testsuites, we are all building testcases, it would be nice to exchange cases and build up a suite.
… We agreed to roll our own, and then compare.
… I have some rudimentary documentation.
… I fear that ambiguity complicates the issue.
Steven: I have a similar problem with testcases for syntax rrors
MSM: My results can have a requirement that the output looks like this, or, there is no parse.
… and there should be another case that the grammar is wrong
… and ambiguity should be treated with more machinery
MSM: My documentation: https://
Norm: What worked for XProc, we assign unique codes to the errors, and then can identify the error output in an unambiguous way.
MSM: How Schema 1.0 was a catastrophe in that respect
Norm: Sure, but it may not be so hard with ixml.
John: I think there are fewer cases for us.
Norm: I volunteer to read the spec, and classify the possible erros
ACTION: Norm to classify the error states in the spec
<trackbot> Sorry, but no Tracker is associated with this channel.
Next meeting
Tomos: A half hour later please.
MSM: 15.30 UTC
Tomos: Permanent please
Steven: Tues 14 dec
[ADJOURN]