HTML reader does not interpret type attribute on <ol>

[adunning]

The HTML reader does not interpret the type attribute on <ol> (output by the HTML5 writer), as in the following case (version 1.15.0.6):

pandoc -f html -t markdown << EOT
<ol type="1">
<li>One</li>
<li>Two</li>
<li>Three</li>
</ol>
<ol type="I">
<li>One</li>
<li>Two</li>
<li>Three</li>
</ol>
<ol type="i">
<li>One</li>
<li>Two</li>
<li>Three</li>
</ol>
<ol type="A">
<li>One</li>
<li>Two</li>
<li>Three</li>
</ol>
<ol type="a">
<li>One</li>
<li>Two</li>
<li>Three</li>
</ol>
EOT

This is the result:

1.  One
2.  Two
3.  Three

<!-- -->

1.  One
2.  Two
3.  Three

<!-- -->

1.  One
2.  Two
3.  Three

<!-- -->

1.  One
2.  Two
3.  Three

<!-- -->

1.  One
2.  Two
3.  Three

[adunning]

Am I missing something here? Given either of these commands:

pandoc -f markdown -t html5 << EOT | pandoc -f html -t markdown

pandoc -f markdown -t html << EOT | pandoc -f html -t markdown

I would expect the following to return identical results:

1.  One
2.  Two

<!-- -->

I.  One
II. Two

<!-- -->

i.  One
ii. Two

<!-- -->

A.  One
B.  Two

<!-- -->

a.  One
b.  Two

EOT

Instead, I receive the following in both cases:

1.  One
2.  Two

<!-- -->

1.  One
2.  Two

<!-- -->

1.  One
2.  Two

<!-- -->

1.  One
2.  Two

<!-- -->

1.  One
2.  Two

[hftf]

Oops, should have said I started fixing this bug. FYI, I think the nearby lookup "style" code (added in e814a3f from 8 years ago) doesn't work as intended:

$ cat <<EOF > foo
<ol class="lower-roman"></ol>
<ol style="lower-roman"></ol>
<ol style="list-style: lower-roman;"></ol>
<ol style="list-style-type: lower-roman;"></ol>
EOF
$ pandoc -f html -t native foo
[OrderedList (1,LowerRoman,DefaultDelim)
 []
,OrderedList (1,LowerRoman,DefaultDelim)
 []
,OrderedList (1,DefaultStyle,DefaultDelim)
 []
,OrderedList (1,DefaultStyle,DefaultDelim)
 []]

I also had trouble running the HTML tests using only cabal test. It seems I needed to completely install pandoc before my changes were used.

[jgm]

Yes, the style thing is messed up. We need to look for
list-style-type. This may even be wrong in the tests
(tests/html-reader.html).

cabal test sometimes doesn't work for me either,
for reasons I don't understand. But you can always run
the test suite manually,
dist/build/test-pandoc/test-pandoc.

+++ Ophir Lifshitz [Jul 21 15 13:20 ]:

Oops, should have said I started fixing this bug. FYI, I think the
nearby lookup "style" code (added in [1]e814a3f from 8 years ago)
doesn't work as intended:
$ cat < foo

EOF
$ pandoc -f html -t native foo
[OrderedList (1,LowerRoman,DefaultDelim)
[]
,OrderedList (1,LowerRoman,DefaultDelim)
[]
,OrderedList (1,DefaultStyle,DefaultDelim)
[]
,OrderedList (1,DefaultStyle,DefaultDelim)
[]]

I also had trouble running the HTML tests using only cabal test. It
seems I needed to completely install pandoc before my changes were
used.

—
Reply to this email directly or [2]view it on GitHub.

References

1. e814a3f#diff-b1c35a3f694e124b0b4ea092ccf36515
2. HTML reader does not interpret type attribute on <ol> #2313 (comment)

[jgm]

PS. I'll wait for you to submit a PR fixing these things, if you still want to work on it.