XProc PRs are published by continuous integration. The following pull requests are open:
Fix #41
This name seems shorter and simpler than the names proposed in the issue. Rather than raise a dynamic error for invalid media types, I opted to simply return the empty sequence. I don’t feel strongly about it.
Fix #123
Fix #4
Fix #122
The intent of this PR is to specify, in section 3.3, that a document containing a single text node has the content type text/plain rather than application/xml. In the course of making that change, I decided that there was scope for editorial improvement throughout the section.
Fix #5
Fix #15
Fix #19
Fix #21
Clarify by removing a paragraph that introduced ambiguity.
Fix #22
Fix #104
Once the section titles had been changed, it seemed sensible to put them in alphabetical order. That’s going to lead to the appearance of much more significant changes than are actually present. This is a purely editorial PR; the only change is to the section titles.
Fix #10
This change mirrors the existing text that restricts where [p:]depends can appear.
Fix #17
Fix #12
This change only effects the XProc grammar, it doesn’t change any of the spec prose.
The definition of a “QName attribute”, that is, an attribute that contains a QName, is updated to allow an EQName as well.
This PR also includes an archival copy of the XProc 3.1 grammar as originally published and a version with a few small, obvious errors corrected:
p:encode, p:invisible-xml, p:message, and p:send-mail steps are includedattributes attribute is included on p:pack, p:wrap, and p:wrap-sequence