Skip to content

Commit d2cd923

Browse files
waylanwaylan
committed
Round out tests of valid markup.
There are some design desisions to make as noted in comments.
1 parent 194b7d3 commit d2cd923

File tree

1 file changed

+232
-2
lines changed

1 file changed

+232
-2
lines changed

‎tests/test_syntax/blocks/test_html_blocks.py‎

Lines changed: 232 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -486,7 +486,18 @@ def test_raw_comment_one_line(self):
486486
'<!-- *foo* -->'
487487
)
488488

489-
# TODO: Confirm this is correct
489+
# TODO: Decide behavior here. Python-Markdown current outputs:
490+
#
491+
# <!-- *foo* -->
492+
# <p><em>bar</em></p>
493+
#
494+
# But the reference implementation outputs:
495+
#
496+
# <p><!-- *foo* --><em>bar</em></p>
497+
#
498+
# As the raw HTML is not alone on the line, the reference implementation
499+
# considers it inline rather than block level. The behavior defined in
500+
# the test below is from the CommonMark spec, which we don't follow.
490501
def test_raw_comment_one_line_followed_by_text(self):
491502
self.assertMarkdownRenders(
492503
'<!-- *foo* -->*bar*',
@@ -533,4 +544,223 @@ def test_raw_comment_with_blank_lines(self):
533544
)
534545
)
535546

536-
# TODO: processing instruction, declaration, CDATA...
547+
def test_raw_comment_indented(self):
548+
self.assertMarkdownRenders(
549+
self.dedent(
550+
"""
551+
<!--
552+
553+
*foo*
554+
555+
-->
556+
"""
557+
),
558+
self.dedent(
559+
"""
560+
<!--
561+
562+
*foo*
563+
564+
-->
565+
"""
566+
)
567+
)
568+
569+
def test_raw_processing_instruction_one_line(self):
570+
self.assertMarkdownRenders(
571+
"<?php echo '>';' ?>",
572+
"<?php echo '>';' ?>"
573+
)
574+
575+
# This is inline as it is not on a line by itself.
576+
def test_raw_processing_instruction_one_line_followed_by_text(self):
577+
self.assertMarkdownRenders(
578+
"<?php echo '>';' ?>*bar*",
579+
"<p><?php echo '>'; ' ?><em>bar</em></p>"
580+
)
581+
582+
def test_raw_multiline_processing_instruction(self):
583+
self.assertMarkdownRenders(
584+
self.dedent(
585+
"""
586+
<?php
587+
echo '>';'
588+
?>
589+
"""
590+
),
591+
self.dedent(
592+
"""
593+
<?php
594+
echo '>';'
595+
?>
596+
"""
597+
)
598+
)
599+
600+
def test_raw_processing_instruction_with_blank_lines(self):
601+
self.assertMarkdownRenders(
602+
self.dedent(
603+
"""
604+
<?php
605+
606+
echo '>';'
607+
608+
?>
609+
"""
610+
),
611+
self.dedent(
612+
"""
613+
<?php
614+
615+
echo '>';'
616+
617+
?>
618+
"""
619+
)
620+
)
621+
622+
def test_raw_processing_instruction_indented(self):
623+
self.assertMarkdownRenders(
624+
self.dedent(
625+
"""
626+
<?php
627+
628+
echo '>';'
629+
630+
?>
631+
"""
632+
),
633+
self.dedent(
634+
"""
635+
<?php
636+
637+
echo '>';'
638+
639+
?>
640+
"""
641+
)
642+
)
643+
644+
def test_raw_declaration_one_line(self):
645+
self.assertMarkdownRenders(
646+
'<!DOCTYPE html>',
647+
'<!DOCTYPE html>'
648+
)
649+
650+
# TODO: Decide correct behavior. This matches current behavior and Commonmark.
651+
# The reference implementation considers this inline not block level:
652+
#
653+
# <p><!DOCTYPE html><em>bar</em></p>
654+
#
655+
# But most implementations do this instead:
656+
#
657+
# <p>&lt;!DOCTYPE html&gt;<em>bar</em></p>
658+
#
659+
# Either makes sense, but the later seems more correct to me.
660+
def test_raw_declaration_one_line_followed_by_text(self):
661+
self.assertMarkdownRenders(
662+
'<!DOCTYPE html>*bar*',
663+
'<!DOCTYPE html>*bar*'
664+
)
665+
666+
def test_raw_multiline_declaration(self):
667+
self.assertMarkdownRenders(
668+
self.dedent(
669+
"""
670+
<!DOCTYPE html PUBLIC
671+
"-//W3C//DTD XHTML 1.1//EN"
672+
"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
673+
"""
674+
),
675+
self.dedent(
676+
"""
677+
<!DOCTYPE html PUBLIC
678+
"-//W3C//DTD XHTML 1.1//EN"
679+
"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
680+
"""
681+
)
682+
)
683+
684+
def test_raw_cdata_one_line(self):
685+
self.assertMarkdownRenders(
686+
'<![CDATA[ document.write(">"); ]]>',
687+
'<![CDATA[ document.write(">"); ]]>'
688+
)
689+
690+
# TODO: Decide correct behavior. This matches current behavior and Commonmark.
691+
# The reference implementation considers this inline not block level:
692+
#
693+
# <p><![CDATA[ document.write(">"); ]]><em>bar</em></p>
694+
#
695+
# But most implementations do this instead:
696+
#
697+
# <p>&lt;[CDATA[ document.write(“&gt;”); ]]&gt;<em>bar</em></p>
698+
#
699+
# Either makes sense, but the later seems more correct to me.
700+
def test_raw_cdata_one_line_followed_by_text(self):
701+
self.assertMarkdownRenders(
702+
'<![CDATA[ document.write(">"); ]]>*bar*',
703+
'<![CDATA[ document.write(">"); ]]>*bar*'
704+
)
705+
706+
def test_raw_multiline_cdata(self):
707+
self.assertMarkdownRenders(
708+
self.dedent(
709+
"""
710+
<![CDATA[
711+
document.write(">");
712+
]]>
713+
"""
714+
),
715+
self.dedent(
716+
"""
717+
<![CDATA[
718+
document.write(">");
719+
]]>
720+
"""
721+
)
722+
)
723+
724+
def test_raw_cdata_with_blank_lines(self):
725+
self.assertMarkdownRenders(
726+
self.dedent(
727+
"""
728+
<![CDATA[
729+
730+
document.write(">");
731+
732+
]]>
733+
"""
734+
),
735+
self.dedent(
736+
"""
737+
<![CDATA[
738+
739+
document.write(">");
740+
741+
]]>
742+
"""
743+
)
744+
)
745+
746+
def test_raw_cdata_indented(self):
747+
self.assertMarkdownRenders(
748+
self.dedent(
749+
"""
750+
<![CDATA[
751+
752+
document.write(">");
753+
754+
]]>
755+
"""
756+
),
757+
self.dedent(
758+
"""
759+
<![CDATA[
760+
761+
document.write(">");
762+
763+
]]>
764+
"""
765+
)
766+
)

0 commit comments

Comments
 (0)