Separate pattern for every instance of <anyElement> — fixes #631 (I hope) #632
Conversation
Generate a different RELAX NG pattern (and reference to it) for every occurence of an anyElement element, and update tests to match. (Previously we were trying to generate one for every content model in which an anyElement occurred, but were generation one for every content model in which an anyElement element occurred only once.)
sydb
added
type: bug
sydb
removed
the
priority: high
| <xsl:variable name="anyElement_disambiguator_format" as="xs:string"> | ||
| <!-- A string used as the xsl:number/@format to append to pattern names for <anyElement>s (thus to disambiguate them) --> | ||
| <xsl:variable name="zeroes" as="xs:string+"> | ||
| <xsl:for-each select="1 to ( count( //tei:anyElement ) => xs:string() => string-length() )">0</xsl:for-each> |
There was a problem hiding this comment.
I think you may be able to simplify this a bit and just use format-integer and set the count as the $picture to format 0:
<xsl:variable name="zeroes" as="xs:string"
select="format-integer(0, xs:string( count( //tei:anyElement ) ) )"/>There was a problem hiding this comment.
Good point. But format-integer() is an XPath 3 function, and this is an XSLT 2 program. I will not be at all surprised if Saxon will process it correctly even in this XSLT 2 context. But I am not sure if we should rely in that.
On the other hand, why are we still using XSLT 2 programs? Any reason not to just change them to XSLT 3? (What, if anything, would break?)
…RELAX NG pattern reduction code (which I wrote > 2 weeks ago).
|
Despite the GitHub build error at |
|
OK, can we merge this puppy, then? |
|
In my local machine, I was getting the same ID-generated error, so I updated the expected results. |
Generate a different RELAX NG pattern (and reference to it) for every occurence of an anyElement element, and update tests to match. (Previously we were trying to generate one for every content model in which an anyElement occurred, but were generation one for every content model in which an anyElement element occurred only once.)
This is high priority because it is making hard to work on #627.