Unicode function seems to break title on Wikidata Query Service
Closed, DeclinedPublic
Actions

Assigned To

None

Authored By

	VIGNERON_en_residence
	Dec 13 2021, 10:54 AM

Description

Context:
Since the #title: function to display to title on the on Wikidata Query Service (see T225883) was undocumented and since it's a simple functionality, I decide to document it myself. So I did some tests on what is acceptable and works or not in the title when I stumble on this issue.

Problem:

When using the Unicode function \u{} in the title on Wikidata Query Service, I've got a Server error (I'm guessing some string not sanitize somewhere?).

Example:
See this query https://w.wiki/4YRh (or any query with a title containing \u)

Acceptance criteria:
No more error
(whether the function should be interpreted or not is not important here, the point is not to be blocked by the title)

Related Objects

Mentioned Here: T225883: Wikidata Query Service results should be capable of displaying a report title (new feature request)

Event Timeline

VIGNERON_en_residence created this task.Dec 13 2021, 10:54 AM

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptDec 13 2021, 10:54 AM

VIGNERON_en_residence triaged this task as Low priority.Dec 13 2021, 10:55 AM

Maintenance_bot added a project: Wikidata.Dec 13 2021, 11:45 AM

Hm, I guess that must be a Blazegraph bug, it shouldn’t try to interpret comments like that. Checking that it also happens without the UI:

$ curl --data-urlencode query='ASK{}#\u' https://query.wikidata.org/sparql
SPARQL-QUERY: queryStr=ASK{}#\u
java.util.concurrent.ExecutionException: java.lang.Error: Invalid escape character at line 1 column 8.
        at java.util.concurrent.FutureTask.report(FutureTask.java:122)
        at java.util.concurrent.FutureTask.get(FutureTask.java:206)
        at com.bigdata.rdf.sail.webapp.BigdataServlet.submitApiTask(BigdataServlet.java:292)
        at com.bigdata.rdf.sail.webapp.QueryServlet.doSparqlQuery(QueryServlet.java:678)
        at com.bigdata.rdf.sail.webapp.QueryServlet.doPost(QueryServlet.java:275)
…
Caused by: java.lang.Error: Invalid escape character at line 1 column 8.
        at com.bigdata.rdf.sail.sparql.ast.JavaCharStream.readChar(JavaCharStream.java:331)
        at com.bigdata.rdf.sail.sparql.ast.SyntaxTreeBuilderTokenManager.jjMoveNfa_0(SyntaxTreeBuilderTokenManager.java:3716)
        at com.bigdata.rdf.sail.sparql.ast.SyntaxTreeBuilderTokenManager.jjMoveStringLiteralDfa0_0(SyntaxTreeBuilderTokenManager.java:182)
        at com.bigdata.rdf.sail.sparql.ast.SyntaxTreeBuilderTokenManager.getNextToken(SyntaxTreeBuilderTokenManager.java:3948)
        at com.bigdata.rdf.sail.sparql.ast.SyntaxTreeBuilder.jj_ntk(SyntaxTreeBuilder.java:9637)
        at com.bigdata.rdf.sail.sparql.ast.SyntaxTreeBuilder.AskQuery(SyntaxTreeBuilder.java:871)
        at com.bigdata.rdf.sail.sparql.ast.SyntaxTreeBuilder.Query(SyntaxTreeBuilder.java:337)
        at com.bigdata.rdf.sail.sparql.ast.SyntaxTreeBuilder.QueryContainer(SyntaxTreeBuilder.java:216)
        at com.bigdata.rdf.sail.sparql.ast.SyntaxTreeBuilder.parseQuery(SyntaxTreeBuilder.java:32)
 …

In DBPedia (Virtuoso) it’s not an issue: https://dbpedia.org/sparql?query=ASK{}%23\u

Thanks for bringing up this bug. As we are planning on migrating off of Blazegraph, Search is not currently addressing bugs in Blazegraph at the moment -- luckily it also seems like this issue can be worked around by removing the commented code.

MPhamWMF moved this task from Incoming to Blazegraph on the Wikidata-Query-Service board.Dec 13 2021, 4:36 PM

I can confirm that it indeed comes from Blazegraph (at least it happens on LinguaLibre SPARQL endpoint https://lingualibre.org/bigdata/#query which is also Blazegraph) and no matter where/what the comment is, so unrelated to title:.

I guess we can close this bug as declined (or stalled ?).
I can't wait to see the replacement for Blazegraph!

Let’s just decline this, I don’t think there’s any other realistic outcome of this task (and keeping it open until we’ve migrated off Blazegraph doesn’t seem useful).

Unicode function seems to break title on Wikidata Query ServiceClosed, DeclinedPublicActions

Description

Related Objects

Event Timeline

Unicode function seems to break title on Wikidata Query Service
Closed, DeclinedPublic
Actions