Meeting minutes
Tom: Prague is happening in June.
… so do we continue to rush, or aim for June?
Steven: The point of doing it for Prague was so there was an audience
MSM: I would prefer to be ready before June
Norm: February seemed like rushing
Tom: Aim for April/May
Previous Actions
ACTION (2021-08-004): Steven to create list of existing requirements,
<trackbot> Sorry, but no Tracker is associated with this channel.
so that we can identify when a new one comes.
Steven: Done
ACTION (2021-08-005): Tom and MSM to come with a pragma proposal.
<trackbot> Sorry, but no Tracker is associated with this channel.
Done
MSM: Let us discuss later in the meeting, to see if we can endorse
ACTION (2021-10-001): Steven to draft a mediatypes proposal (see
https://
<trackbot> Sorry, but no Tracker is associated with this channel.
Steven: Continues
ACTION (2021-11-001): Norm to classify the error states in the spec,
<trackbot> Sorry, but no Tracker is associated with this channel.
as a first step towards standard error codes. (Discharged by
https://
Done
Status of implementations
Norm: I have modified PEP and successfully parsed an ixml grammar
… It parses the ixml grammar!
github.com/xproc/pep
John: I managed to get a modified version of the ixml grammar working on Steven's server
… with pragmas
… I can then transform and serialise back into ixml, and goes through Steven's ixml processor.
… Opens a need for taking the XML form of ixml
MSM: That's why there is prose in the pragma's proposal about XML form of pragmas
Steven: I'm working on serialising to memory, so that you can parse straight from that. Adding XML would be the same process, using a different format, since the XML and the parse tree are essentially isomorphic.
Steven: The new version is running, is written in ABC, as a basis for a future C version
MSM: Stasis. Working on the test suite, but not on the ixml system
Tom: Likewise, no change this month.
Norm: What is the state of the test suite?
… for use
MSM: I have a relaxing schema for a testsuite catalogue, it allows input and output to be inline or at a URL.
Steven: What should I do with my tests?
MSM: Put them on github, I will develop a catalogue for it
… I do not regard what I'm doing as official, but we can use it as a basis
ACTION: Steven to add his tests to github
<trackbot> Sorry, but no Tracker is associated with this channel.
Steven: Are we expecting to have implementations when we release?
[Yes]
MSM: a list of ongoing implementations and their characteristics
MSM: There's something on CPAN that might count
Tom: Honorable mention
MSM: as long as we can distinguish things that are trying to pass the testsuite
… separately
Bug reports / change requests
Bugs - Strings proposal
https://
Steven: It doesn't reduce functionality, it does improve diagnostics
RESOLVED
ACTION: Steven to add strings proposal to spec
<trackbot> Sorry, but no Tracker is associated with this channel.
Bugs - Character ranges
Bugs - Conformance issue: rule names and XML names
https://
Steven: This is a static vs dynamic issue
a: @b; -"!", c.
b: c.
c: "d".
MSM: I could live with it either way.
… "all names that are serialized must conform to an XML name"
… plus implementations may warn statically if a name doesn't conform
Steven: Agreed?
[Yes]
[RESOLVED]
Bugs - Conformance issue: one rule per nonterminal
https://
MSM: It doesn't cover the start rule, because it doesn't occur on any RHS
Steven: Ah,
MSM: "For every nonterminal occurring in the grammar, there must be exactly one rule in the grammar defining that nonterminal."
Dave: "Every nonterminal should be defined by a single rule".
MSM: replace should with must.
John: Modularity, libraries, may have unreachable rules.
Tom: Not a 1.0 feature
Tom: Not sure we cover the start rule
MSM: The LHS name is a nonterminal; if we don't say that, we should.
Tom: I propose keeping both rules in this issue
MSM: I just looked at "nonterminal" in the spec. I thought "of course the name of a rule is a nonterminal". But the spec doesn't say that. So the redundancy of both rules may help.
Tom: Or us "nonterminal" instead of "name".
Tom: Leave both in.
… use new wording though for the 1st
ACTION: Steven implement the resolution for issue 19.
<trackbot> Sorry, but no Tracker is associated with this channel.
AOB
Tom: Move pragmas and release schedule to top of next meeting
… next meeting 11 Jan at 15:30 UTC
ACTION: Steven implement resolution for issue 18 on XML names
<trackbot> Sorry, but no Tracker is associated with this channel.