ADDED: Add pathological test suite from CommonMark

diff --git a/tools/pathological_tests.sh b/tools/pathological_tests.sh
new file mode 100755 (executable)
index 0000000..2eeecb7
--- /dev/null
+++ b/tools/pathological_tests.sh
@@ -0,0 +1,108 @@
+#!/bin/bash
+
+X=65535
+#X=100
+
+Y=$((X-1))
+
+# Pathological tests from CommonMark
+
+# Nested strong/emph
+seq -f "*a **a" $X >                           path1.text
+echo -n "b " >>                                                path1.text
+seq -f "a** a*" $X  >>                         path1.text
+
+echo -n "<p>" >                                                path1.html
+seq -f "<em>a <strong>a" $X >>         path1.html
+echo -n "b " >>                                                path1.html
+seq -f "a</strong> a</em>" $Y >>       path1.html
+echo "a</strong> a</em></p>" >>                path1.html
+
+
+# Close unopened emph
+seq -f "a_" $X >       path2.text
+
+echo -n "<p>" >                path2.html
+seq -f "a_" $Y >>      path2.html
+echo "a_</p>" >>       path2.html
+
+
+# Open unclosed emph
+seq -f "_a" $X >       path3.text
+
+echo -n "<p>" >                path3.html
+seq -f "_a" $Y >>      path3.html
+echo "_a</p>" >>       path3.html
+
+
+# Close unopened links
+seq -f "a]" $X >       path4.text
+
+echo -n "<p>" >                path4.html
+seq -f "a]" $Y >>      path4.html
+echo "a]</p>" >>       path4.html
+
+
+# Open unclosed links
+seq -f "[a" $X >       path5.text
+
+echo -n "<p>" >                path5.html
+seq -f "[a" $Y >>      path5.html
+echo "[a</p>" >>       path5.html
+
+
+# Mismatched * and _
+# TODO: Starts getting slow > 10k iterations
+# Walking entire stack to find that there is no match.
+# Could keep count of different token types on stack
+# To ensure there's even something to find.  Which does
+# mean more overhead CPU time
+seq -f "*a_" $X >      path6.text
+
+echo -n "<p>" >                path6.html
+seq -f "*a_" $Y >>     path6.html
+echo "*a_</p>" >>      path6.html
+
+
+# Unclosed links and unopened emph
+seq -f "[ a_" $X >     path7.text
+
+echo -n "<p>" >                path7.html
+seq -f "[ a_" $Y >>    path7.html
+echo "[ a_</p>" >>     path7.html
+
+
+# Complex case
+seq -f "**x [*b**c*](d)" $X >                                                          path8.text
+
+echo -n "<p>" >                                                                                                path8.html
+seq -f "**x <a href=\"d\">a<em>b</em><em>c</em><a>" $Y >>      path8.html
+echo "**x <a href=\"d\">a<em>b</em><em>c</em><a></p>" >>       path8.html
+
+
+# Nested brackets
+# TODO: Performance hit between 1000 and 10,000 iterations
+# Checking each level to see if it's a valid link on the stack
+seq -s " " -f "[" $X > path9.text
+echo -n "a" >>                         path9.text
+seq -s " " -f "]" $X >>        path9.text
+
+echo -n "<p>" >                        path9.html
+seq -s " " -f "[" $X >>        path9.html
+echo -n "a" >>                         path9.html
+seq -s " " -f "]" $Y >>        path9.html
+echo "]</p>" >>                        path9.html
+
+
+# Nested block quotes
+# TODO: Segfaults -- recursive parsing after ~~10k nests
+seq -s " " -f ">" $X >                 path10.text
+echo -n "a" >>                                         path10.text
+
+echo -n "<p>" >                                        path10.html
+seq -f "<blockquote>" $X >>            path10.html
+echo -n "a" >>                                         path10.html
+seq -f "</blockquote>" $X >>   path10.html
+echo -n "</p>" >>                              path10.html
+
+