szajbus / saxy Goto Github PK

I noticed in writing a small sample script to try Saxy out that ARGF wasn't handled. Threw an error about trying to get the class name of the object. ARGF is IO-like, but not actually an ancestor of IO. Seems like Nokogiri gracefully duck-types this IO object in either parse or parse_io.

Could https://github.com/monterail/saxy/blob/master/lib/saxy/parser.rb#L71 check for an IO-like interface instead of class ancestry?

I'd be more than happy to supply a PR, just let me know if you're interested.

Greetings, @szajbus

We recently encountered a case in which we needed to constrain the element matches to those within a parent element.

We implemented this functionality in this commit and will use our fork to solve the problem in our application. We had to write this feature against 0.5.2 because we're still on 0.5.2—some later releases like 0.6 and 0.7 had breaking API changes that we are unable to upgrade to.

I'm wondering how you feel about the functionality implemented before I try to pull request it against this repo. If you were interested, I would request that we create a 0.5.x branch, implement the feature, tag a 0.5.3 release, and publish it. I could then port the feature to the 0.7.0 tip.

Given the element

<products>
    <product>
      <name>This &amp; That</name>

The resulting parsed string for name reads 'This&That' without the pre-existing spacing.