Unclear error message when used with pandoc --to=jats

[dhimmel]

In a build for manubot/rootstock#82, I'm getting the following error:

  File "/home/appveyor/miniconda/envs/manubot/lib/python3.7/site-packages/pandoc_eqnos.py", line 213, in _add_markup
    return ret
UnboundLocalError: local variable 'ret' referenced before assignment

The corresponding pandoc-eqnos function is:

pandoc-eqnos/pandoc_eqnos.py

Lines 178 to 213 in 6341171

	def _add_markup(fmt, eq, value):
	"""Adds markup to the output."""
	attrs = eq['attrs']
	# Context-dependent output
	if eq['is_unnumbered']: # Unnumbered is also unreferenceable
	ret = None
	elif fmt in ['latex', 'beamer']:
	ret = RawInline('tex',
	r'\begin{equation}%s\end{equation}'%value[-1])
	elif fmt in ('html', 'html5', 'epub', 'epub2', 'epub3') and \
	LABEL_PATTERN.match(attrs.id):
	# Present equation and its number in a span
	num = str(references[attrs.id].num)
	outer = RawInline('html',
	'<span%sclass="eqnos">' % \
	(' ' if eq['is_unreferenceable'] else
	' id="%s" '%attrs.id))
	inner = RawInline('html', '<span class="eqnos-number">')
	eqno = Math({"t":"InlineMath"}, '(%s)' % num[1:-1]) \
	if num.startswith('$') and num.endswith('$') \
	else Str('(%s)' % num)
	endtags = RawInline('html', '</span></span>')
	ret = [outer, AttrMath(*value), inner, eqno, endtags]
	elif fmt == 'docx':
	# As per http://officeopenxml.com/WPhyperlink.php
	bookmarkstart = \
	RawInline('openxml',
	'<w:bookmarkStart w:id="0" w:name="%s"/><w:r><w:t>'
	%attrs.id)
	bookmarkend = \
	RawInline('openxml',
	'</w:t></w:r><w:bookmarkEnd w:id="0"/>')
	ret = [bookmarkstart, AttrMath(*value), bookmarkend]
	return ret

We're specifying pandoc's --to=jats, which does not appear to be supported by this function. Should there be a final else to capture unsupported formats and either:

1. provide a more clear error message
2. issue a warning that pandoc-eqnos does not support JATS, but proceed such that the pandoc command does not fail

[tomduck]

Thank you for the report and sorry for the long wait. This is definitely a bug. I have fixed it in the nextrelease branch. If the problem should reappear following the next release, please feel welcome to re-open this issue.

Tom

[dhimmel]

Thanks! Reference the commit at d479d14

[tomduck]

Sorry, Daniel, but I am not sure if you are asking me to do something here?

[dhimmel]

Ah no. You are all good. Just wanted to link to the commit that fixed the issue in case that is ever helpful in the future.

…

On Thu, Jul 30, 2020, 10:06 AM Thomas J. Duck ***@***.***> wrote: Sorry, Daniel, but I am not sure if you are asking me to do something here? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#43 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIQ4B57IGQWFLEYKMAOBD3R6F45HANCNFSM4KILDBWA> .