sharplispers / cxml Goto Github PK

@ruricolist made his test suite available in a separate repository:

https://github.com/ruricolist/cxml-tests

I've made a separate repository for it because it's quite large (>30 Mb). You might decide (as I did) that keeping the test suite with the rest of the code is worth the extra size.

(The other advantage of a separate repository, of course, is that you can easily test against older versions of CXML.)

Originally posted by @ruricolist in #3 (comment)

I'd like to propose switching CXML to use named-readtables. It would allow removing the component classes from the .asd file, moving it in a much more declarative direction.

This isn't technically challenging (I already did it for FXML) but it would require a coordinated change across closure-common and CXML so I wanted to raise it for discussion first.

You can see the contemplated changes in these commits:

ruricolist/FXML@fd1ffd9
ruricolist/FXML@fd1ffd9

This codeblock contributes to (IMO unnecessarily) verbose Quicklisp loads of cxml, for instance:

To load "foo-system":
  Load 1 ASDF system:
    foo-system

; Loading "foo-system"
.
;;; Checking for wide character support... WARNING: Lisp implementation doesn't use UTF-16, but accepts surrogate code points.
 yes, using code points.
..
;;; Checking for wide character support... WARNING: Lisp implementation doesn't use UTF-16, but accepts surrogate code points.
 yes, using code points.
;;; Building Closure with CHARACTER RUNES
...............................................
...................

Is it possible/desired to remove the FORMAT calls altogether?

Is it possible to call DOM's getBoundingClientRect() on an Element which i get by:

(aref (dom:get-elements-by-tag-name
       (cxml:parse-file "./test.svg" (cxml-dom:make-dom-builder))
       "path") 0)

This returns a #<RUNE-DOM::ELEMENT path {10042A87A3}> from the SVG file test.svg.

I have noticed that cxml is looping for a certain combination of xml/dtd. It loops in the recurse function inside compile-content-model in xml-parse.lisp. I have put a lot of time on trying to understand how the code is supposed to work but without success.
It would be great if someone knowing the code could have a look at this!

A minimal reproducible case is available here https://doremir.com/files/cxml/xml-test.zip.

The xml file from the case looks like this:

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE a PUBLIC "" "">
<a>
  <c/>
</a>

And the dtd file looks like this:

<!ELEMENT b EMPTY>
<!ELEMENT c EMPTY>
<!ELEMENT a ((b?)*, c?)>

The real world case is from parsing MusicXML and the problematic xml excerpt looks like this:

<sound tempo="90">
  <offset sound="yes">8</offset>
</sound>

And the corresponding dtd excerpt looks like this:
<!ELEMENT sound ((midi-device?, midi-instrument?, play?)*, offset?)>