Match \pagebreak in commonmark_x+sourcepos

[Duncan Murdoch <murdoch.duncan-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>]

R Markdown normally uses Pandoc Markdown with the raw_tex extension, so 
text like

   \pagebreak

is recognized as Latex, and translated to a different format by the 
pagebreak.lua filter here:

 
https://github.com/rstudio/rmarkdown/blob/main/inst/rmarkdown/lua/pagebreak.lua

That filter relies on the fact that the \pagebreak has already become a 
single RawBlock.

This filter doesn't work with commonmark_x+sourcepos, because that 
reader outputs this from \pagebreak in the source:

  Div
       ( "" , [] , [ ( "data-pos" , "test.knit.md@27:3-28:1" ) ] )
       [ Para
           [ Span
               ( ""
               , []
               , [ ( "data-pos" , "test.knit.md@27:3-27:4" ) ]
               )
               [ Str "\\" ]
           , Span
               ( ""
               , []
               , [ ( "data-pos" , "test.knit.md@27:4-27:13" ) ]
               )
               [ Str "pagebreak" ]
           ]
       ]

So it appears the test I need to use is to look for a Para containing 
two Spans whose content is "\\" and "pagebreak", and replace that with 
whatever is appropriate for the output.  This function works, but I 
wonder if it could be simplified:

   function Para (el)
     if #el.content == 2 and
      el.content[1].content == pandoc.Span('\\').content and
      (el.content[2].content == pandoc.Span('pagebreak').content or
       el.content[2].content == pandoc.Span('newpage').content) then
       return newpage(FORMAT)
     end
   end

Do I really need to construct those Spans to do the test?  I'm happy it 
works, but it seems like there should be simpler right hand sides in the 
comparisons.

Duncan Murdoch