SPARUL updates: HTTP response encoding expectations too restrictive

[u-u-h]

When using an %endpoint to submit SparQL Update (SPARUL) update statements like LOAD, the W3C spec does not specify much about the response format. Some implementations like CGE return no useful HTTP body, with a "200 OK" or "204 No Content" code and a content body format of text/html. That seems permitted by the spec.

The current implementation sets the acceptable mime types (connection.py:36 and following) to application/sparql-results+json , text/rdf or similar, but cannot deal with text/* in general.

It would be good to either introduce an '%update-endpoint' and use that for SPARUL commands, or at least have a %format specifier that can be set near SPARUL commands to handle the no-result situation appropriately.

--uuh

[paulovn]

You are right. I wasn't thinking on SPARQL Update when developing the code, so it is not considered.

https://github.com/paulovn/sparql-kernel/pull/17 should fix this. Specifically:

• Now a specific format is requested only for SELECT, CONSTRUCT or DESCRIBE queries
• In any case, if the automatic format selection fails, it is always possible to use a %format any magic to suppress format checking

Tell me if it works. Thanks for detecting this.

[paulovn]

Should be fixed in version 1.0.5