DocBook reader ignores the `id` attribute of `informalequation`

[snwnde]

Explain the problem.

DocBook reader ignores the id attribute of informalequation. This attribute is useful for crossref, and the equivalent attribute of section is very well preserved. It is thus expected that for informalequation this attribute also passes to AST.

Here is an example DocBook input,

<?xml version="1.0" encoding="UTF-8"?>
<?asciidoc-toc?>
<?asciidoc-numbered?>
<article xmlns="http://docbook.org/ns/docbook" xmlns:xl="http://www.w3.org/1999/xlink" version="5.0" xml:lang="en">
<info>
<title>Example</title>
<date>2023-03-03</date>
</info>
<informalequation xml:id="eq-massenergy">
<alt><![CDATA[E = mc^2]]></alt>
<mathphrase><![CDATA[E = mc^2]]></mathphrase>
</informalequation>
<simpara>Let&#8217;s refer to <link linkend="eq-massenergy">the equation</link>.</simpara>
</article>

and the output AST,

Pandoc
  Meta
    { unMeta =
        fromList
          [ ( "date" , MetaInlines [ Str "2023-03-03" ] )
          , ( "title" , MetaInlines [ Str "Example" ] )
          ]
    }
  [ Para [ Math DisplayMath "E = mc^2" ]
  , Para
      [ Str "Let\8217s"
      , Space
      , Str "refer"
      , Space
      , Str "to"
      , Space
      , Link
          ( "" , [] , [] )
          [ Str "the" , Space , Str "equation" ]
          ( "#eq-massenergy" , "" )
      , Str "."
      ]
  ]

returned by pandoc -s -t native -f docbook example.xml.

Pandoc version?
v3.1 (https://pandoc.org/try/)

[snwnde]

Related to #8666.

[tombolano]

I am the user that reported #8666, it's an interesting coincidence that I encountered an very similar problem at the same time 😃.

Looking at the DocBook reader code it seems that the id attribute is ignored for some of the block DocBook elements. Other attributes such as role are also ignored as reported in #3657, but the id attribute is essential for cross-references to work so it should not be ignored.

For the case of informalequation, the DocBook reader creates a simple Para element containing the equation. I think that this can be fixed by creating a Div with the correct id value. I tried the following change in the code that seems to fix this:

diff --git a/src/Text/Pandoc/Readers/DocBook.hs b/src/Text/Pandoc/Readers/DocBook.hs
index e11da4253..d5d468f92 100644
--- a/src/Text/Pandoc/Readers/DocBook.hs
+++ b/src/Text/Pandoc/Readers/DocBook.hs
@@ -878,7 +878,8 @@ parseBlock (Elem e) =
         "bibliomisc" -> parseMixed para (elContent e)
         "bibliomixed" -> parseMixed para (elContent e)
         "equation"         -> para <$> equation e displayMath
-        "informalequation" -> para <$> equation e displayMath
+        "informalequation" -> divWith (attrValue "id" e,["informalequation"],[]) .
+                              para <$> equation e displayMath
         "glosssee" -> para . (\ils -> text "See " <> ils <> str ".")
                          <$> getInlines e
         "glossseealso" -> para . (\ils -> text "See also " <> ils <> str ".")

With this change the resulting AST for the DocBook code provided by @snwnde in the first message when executing pandoc -t native -f docbook example.xml (I omit the -s for simplicity) is:

[ Div
    ( "eq-massenergy" , [ "informalequation" ] , [] )
    [ Para [ Math DisplayMath "E = mc^2" ] ]
, Para
    [ Str "Let\8217s"
    , Space
    , Str "refer"
    , Space
    , Str "to"
    , Space
    , Link
        ( "" , [] , [] )
        [ Str "the" , Space , Str "equation" ]
        ( "#eq-massenergy" , "" )
    , Str "."
    ]
]