Prevent RDFa parser from failing on time elements with child nodes #577
Conversation
The _get_literal() function in the RDFa parser referred to a self instance that does not exist. This led to a NameError exception if the HTML to be parsed included a TIME element with one or more child nodes, such as '<time><em>2016-01-01</em></time>'. Removing the reference to 'self' fixes the problem and fixes RDFLib#576. Signed-off-by: Dan Scott <dan@coffeecode.net>
|
hmm, actually same problem as with pyMicrodata: maybe such things should go to https://github.com/RDFLib/PyRDFa or https://github.com/RDFLib/pyrdfa3 so they could be included in upstream test cases? @iherman ? |
|
Thanks @dbs and @joernhees. I have made the change in the https://github.com/RDFLib/pyrdfa3 repo (see RDFLib/pyrdfa3#13). The https://github.com/RDFLib/PyRDFa is explicitly marked as 'outdated', so I did not touch it. |
|
I do not do merges of PR myself, I leave that to @joernhees ... |
Prevent RDFa parser from failing on time elements with child nodes, backported in RDFLib/pyrdfa3@f1030e7
|
I keep hitting this bug over and over. Is v4.2.2 going to be released any time soon please so I can just depend on that version in all my builds? |
|
@avengerpenguin sorry for that... my plan was to fix a couple of SPARQL bugs (Milestone 4.2.2) before the release, but i couldn't find enough time to work through them yet... maybe we should just release 4.2.2 and care about the SPARQL issues in a future release? |
|
That's up to you if you feel 4.2.2 is in a fit state to be released. I would normally advocate a bug fix release for each bug just to get fixes out there for people, just as with my situation where I would benefit from this issue being fixed in a proper package and I haven't encountered any of the SPARQL bugs you're holding off the release for. Nothing's stopping you having a 4.2.3 release with those bugs fixed when you have time :-) If you think 4.2.2 has is no worse than 4.2.1 and it has this issue fixed in it, then please do release it if you can. |
|
I've got my application building ok using the master branch of this repo, so that takes some pressure off me. I guess it'd still be nice to see a release with this fix and any other fixes if 4.2.2-dev has fewer bugs than 4.2.1. |
The _get_literal() function in the RDFa parser referred to a self instance that
does not exist. This led to a NameError exception if the HTML to be parsed
included a TIME element with one or more child nodes, such as
'2016-01-01'. Removing the reference to 'self' fixes
the problem and fixes #576.
Signed-off-by: Dan Scott dan@coffeecode.net