On 01/25/19 00:45, Pavel Stehule wrote: > pá 25. 1. 2019 v 5:46 odesílatel Chapman Flack <chap@anastigmatix.net> > napsal: >> I am: >> - re-attaching xmltable-xpath-result-processing-bugfix-5.patch unchanged >> (just so CF app does not lose track) >> - re-attaching xmltable-xmlexists-passing-mechanisms-1.patch unchanged >> - attaching for the first time xml-functions-type-docfix-1.patch >> >> The doc patch is made to go on top of the passing-mechanisms patch
Realized xmltable-xmlexists-passing-mechanisms-1.patch didn't add a regression test. Here attaching (or re-attaching):
- xmltable-xpath-result-processing-bugfix-5.patch - unchanged - xmltable-xmlexists-passing-mechanisms-2.patch - now with test - xml-functions-type-docfix-1.patch - unchanged
I'll venture a review opinion that all of this applies, builds, and passes check-world on top of 18c0da8, and that, of the issues I had identified at the start of this thread, these changes resolve the ones they set out to resolve.
But the second two patches are my own work, so another reviewer is needed. The passing-mechanisms patch is tiny while the docfix patch is not, so there's an opening for a reviewer with an interest in documentation. :)
There is still nothing in this patch set to address [1], though that also seems worth doing, perhaps in another patch, and probably not difficult, perhaps needing only a regex.
And of course we're still saddled with all the unfixable limits of XPath 1.0; this patch set is fixing a few peripheral fixable things around that.
I am sending a review of these patches
xmltable-xpath-result-processing-bugfix-5.patch - I'll skip it - just all tests passed
xmltable-xmlexists-passing-mechanisms-2.patch - this patch introduce new PASSING mechanism BY VALUE - it is just syntactic sugar due compatibility with standard. It is unhappy so previous implementation was broken and introduced "BY REF" instead "BY VALUE", but this bug should be fixed 10 years ago. It change nothing, all tests passed and the documentation looks ok.
Last patch is documentation only patch - I am thinking so the difference and limits our implementation of XPath based functions are described well and correctly.
