SPARQL String. Unicode escapes exclude surrogates. #190
Conversation
afs
force-pushed
the
strings
branch
2 times, most recently
from
9b68e44 to
cda46be
Compare
| @@ -10551,15 +10570,19 @@ <h3>Codepoint Escape Sequences</h3> | |||
| <a href="#HEX">HEX</a> <a href="#HEX">HEX</a> | |||
| </td> | |||
| <td>A Unicode code point in the range U+0 to U+FFFF inclusive corresponding to the | |||
| encoded hexadecimal value.</td> | |||
| encoded hexadecimal value, excluding U+D800 to U+DFFF, the | |||
There was a problem hiding this comment.
We don't make this restriction in the Turtle (or related) grammars. It's arguably not necessary, as the value space already restricts RDF/SPARQL strings from including bare surrogates. There should probably be some tests that attempt to create strings using such escape sequences and result in syntax errors.
There was a problem hiding this comment.
syn-invalid-codepoint-escaped-bad-01.rq
There was a problem hiding this comment.
The input is an RDF string but the output isn't.
By adding the restriction to unicode escapes, the outcome is an RDF string and nothing more needs to be said. Saying it at the point where bad things happen ™️ is IMO clearer. Could be a note.
Otherwise, somewhere should say that EBNF parsing is on an RDF string again, which is really just moving the unicode escape text about.
Slightly different issue in Turtle because of the different way Unicode escapes are handled which is during parsing. But the text there says the outcome is in the range U+0000 to U+FFFF which includes surrogates.
There was a problem hiding this comment.
sorry for missing this earlier, this prevents code points that can be encoded as paired surrogates to be encoded as as two consecutive escape sequences (one for the high surrogate and one for the low one). I am not sure we allowed that explicitely before so I am not sure it's a big deal. See w3c/rdf-turtle#84
| @@ -5364,7 +5364,7 @@ <h4>Operator Extensibility</h4> | |||
| </section> | |||
| <section id="SparqlOps"> | |||
| <h3>Function Definitions</h3> | |||
| <p>This section defines the operators and functions introduced by the SPARQL Query language. | |||
| <p>This section defines the operators and functions introduced by the SPARQL query language. | |||
There was a problem hiding this comment.
I do not understand the variance in capitalization. SPARQL Query language here (line 5367) is being changed to SPARQL query language, while earlier in the document (line 393), SPARQL query language is being changed to SPARQL Query Language. Note that neither line contains (part of) a title; they're both body prose.
There are other variances elsewhere in the document. Why do these vary? What is being communicated by the difference in casing, besides confusion?
This closes #188
This closes #189
"Request" is too oriented to protocol action
Used more reSpec
Preview | Diff