PCDATA invalid Char value

[billdenney]

I'm not sure if this is a bug in commonmark or xml2 I will report it here because I think that the issue is with the original encoding here. This happened when trying to use usethis::use_spell_check() on the janitor package.

When converting the attached NEWS.md file, I think that commonmark doesn't correctly parse and escape values. (The issue may be that xml2 doesn't correctly handle these values, though.)

text <- readLines("https://github.com/user-attachments/files/18189322/NEWS.md")
md <- commonmark::markdown_xml(text, sourcepos = TRUE)
doc <- xml2::xml_ns_strip(xml2::read_xml(md))
#> Error in read_xml.raw(charToRaw(enc2utf8(x)), "UTF-8", ..., as_html = as_html, : PCDATA invalid Char value 19 [9]

^{Created on 2024-12-18 with reprex v2.1.1}

[jeroen]

Thanks, I can reproduce this. Strangely I don't see any PCDATA in the generated xml text. Do you want to help narrow down which line of the NEWS.md is triggering the bug?

[billdenney]

The error relates to line 295 (I just looped over each line to see where it happened).

text <- readLines("https://github.com/user-attachments/files/18189322/NEWS.md")
md <- commonmark::markdown_xml(text[[295]], sourcepos = TRUE)
doc <- xml2::xml_ns_strip(xml2::read_xml(md))