<kaz> scribenick: ege
McCool: review where we are
... (shows implementation report)
... changes to the TD repo added new assertions
... most of them have to do with the context
... some of them can be tested with JSON Schema but some other
ones like URL should be dereferencable, we cannot test with
JSON Schema
... there are duplicates
... we need to create context at form and security level
... I think we have to assume to keep these context
assertions
Kaz: this is great. thank you!
... on the other hand, please remember that for CR transition what is required is clarifying all the assertions based on the spec
... fullfilling all the assertions with concrete results is not required
... getting concrete results earlier is of course nice, though
Ege: can we leave all of them not tested for CR
McCool: we can test all of them
manually
... at the worst case, we can ,ake them "features at risk"
Kaz: note that having 20 features as "features at risk" would not be good
McCool: these are all related to one
issue, context
... some are, like scopes, are very easy to fix
... panasonic added some examples
... so if some other companies do that too, it will be
fine
... ege do you know why this td-data-schema-objects is not
tested
Ege: I don't know, I will write an issue for it
McCool: I should mark some as
resolved
... this is related to some subitems, td-event-names
... there is also multi languages missing
... we also need uriVariables
Ege: urivariables in events are complicated
McCool: each time I see 0s it looks
weird
... td-forms is not tested?
... i think we miss a test here
... ege you should go through this list here
Ege: yes I can do that
<Zakim> kaz, you wanted to ask about (1) whether this report.html is generated based on the CSV files manually generated (or using the assertion tester) and (2) if this result includes all the results from Princeton TestFest
Kaz: I just want to make
sure
... are these manual or automatic
McCool: both
<McCool> https://github.com/w3c/wot/tree/master/testfest/2019-03-online/templates
McCool: (shows the directory)
... panasonic had some updates as well
Kaz: so you merged automatic and manual
McCool: yes
... (explains how the merging works)
Kaz: I wasn't so sure whether these were equal
McCool: they don't overlap
Kaz: it would be nice to explain the fact that the inputs directory includes both the manually tested CSVs and the automatically tested CSVs, and the covered areas of assertions are different between them (manually tested CSV and automatically tested CSV)
Ege: in the readme maybe?
McCool: documentation is there if you look at the update.sh script but it would be clearer to split the "inputs" area into 2 pieces, one for manually tested CSVs and another for automatically tested CSVs. let me think about that.
Ege: the links are not valid
anymore in the descriptions
... we have /inputs instead of /TDs anymore
Lagally: something has changed
<McCool> https://github.com/w3c/wot-thing-description/tree/master/testing/inputs/implementations
McCool: the readme is not that
important but the html descriptions are
... I will edit them
... I have highlight for all the assertions, gray by default
and yellow if it is at risk
Lagally: why is that name assertion yellow
Ege: this must because of the name being similar
McCool: (explains the bug)
Lagally: there is something wrong with description as well?
McCool: it seems so
... this happens because of the render
... and I can't change it easily
... maybe it is because the data is not provided but still
marked as risk
Lagally: can we go through the spec and do a sanity check?
McCool: there is the content-type as
well
... we have two contentType at the output data
... but only one here
... somehow handling multiple assertions is not good
... highlighting is not being generated right
... I can go over them manually and change the ids
... it is kind of tricky
Lagally: it is important to know whether some specs will go out of the spec