Improvements to smart-quote regexs. Now we can better handle

cases where latex commands or HTML entity references appear
after quotes.


git-svn-id: https://pandoc.googlecode.com/svn/trunk@202 788f1e2b-df1e-0410-8736-df70ead52e1b

3 files changed, 21 insertions, 12 deletions

diff --git a/TODO b/TODO
index ee710a070..5264b8a67 100644
--- a/TODO
+++ b/TODO
@@ -1,7 +1,5 @@
 # TODO
 
-*   Fix bug in latex reader?  "... gets converted the wrong way.
-
 *   Use XHTML library for HTML writer?
 
 *   Revisions for building with windows under cygwin:
diff --git a/src/Text/Pandoc/Writers/HTML.hs b/src/Text/Pandoc/Writers/HTML.hs
index f0283dd48..1f1258753 100644
--- a/src/Text/Pandoc/Writers/HTML.hs
+++ b/src/Text/Pandoc/Writers/HTML.hs
@@ -55,19 +55,26 @@ stringToHtml str = escapePreservingRegex stringToHtmlString (mkRegex "\"|(&[[:al
 stringToSmartHtml :: String -> String
 stringToSmartHtml = 
     let escapeDoubleQuotes = 
-            gsub "(\"|"|'')" "”" . -- rest are right quotes
-            gsub "([[:space:]])(\"|")" "\\1“" . -- never right quo after space 
-            gsub "(\"|"|``)('|`|‘)([^[:punct:][:space:]])" "“‘\\3" . -- "'word left
-            gsub "(\"|"|``)([^[:punct:][:space:]])" "“\\2"  -- "word left
+            gsub "(\"|")" "”" . -- rest are right quotes
+            gsub "(\"|")(&r[sd]quo;)" "”\\2" . -- never left quo before right quo
+            gsub "(&l[sd]quo;)(\"|")" "\\2“" . -- never right quo after left quo
+            gsub "([ \t])(\"|")" "\\1“" . -- never right quo after space 
+            gsub "(\"|")([^,.;:!?^) \t-])" "“\\2" . -- "word left
+            gsub "(\"|")('|`|‘)" "”’" . -- right if it got through last filter
+            gsub "(\"|")('|`|‘)([^,.;:!?^) \t-])" "“‘\\3" . -- "'word left
+            gsub "``" "“" .
+            gsub "''" "”"
         escapeSingleQuotes =
             gsub "'" "’"  . -- otherwise right
-            gsub "([[:space:]])'" "\\1‘" . -- never right quo after space 
+            gsub "'(&r[sd]quo;)" "’\\1" . -- never left quo before right quo
+            gsub "(&l[sd]quo;)'" "\\1‘" . -- never right quo after left quo
+            gsub "([ \t])'" "\\1‘" . -- never right quo after space 
             gsub "`" "‘"  . -- ` is left
-            gsub "([^[:punct:][:space:]])'" "\\1’" .  -- word' right
+            gsub "([^,.;:!?^) \t-])'" "\\1’" .  -- word' right
+            gsub "^('|`)([^,.;:!?^) \t-])" "‘\\2" . -- 'word left 
             gsub "('|`)(\"|"|“|``)" "‘“" .  -- '"word left
-            gsub "^('|`)([^[:punct:][:space:]])" "‘\\2" . -- 'word left 
-            gsub "([^[:punct:][:space:]])'(s|S)" "\\1’\\2" . -- possessive
-            gsub "([[:space:]])'([^[:punct:][:space:]])" "\\1‘\\2" . -- 'word left
+            gsub "([^,.;:!?^) \t-])'(s|S)" "\\1’\\2" . -- possessive
+            gsub "([[:space:]])'([^,.;:!?^) \t-])" "\\1‘\\2" . -- 'word left
             gsub "'([0-9][0-9](s|S))" "’\\1"  -- '80s - decade abbrevs.
         escapeDashes = gsub " ?-- ?" "—" .
                        gsub " ?--- ?" "—" .
diff --git a/src/Text/Pandoc/Writers/LaTeX.hs b/src/Text/Pandoc/Writers/LaTeX.hs
index 13eac31c2..22a96a423 100644
--- a/src/Text/Pandoc/Writers/LaTeX.hs
+++ b/src/Text/Pandoc/Writers/LaTeX.hs
@@ -63,12 +63,16 @@ escapeLt        = gsub "<" "\\\\textless{}"
 escapeGt        = gsub ">" "\\\\textgreater{}"
 
 escapeDoubleQuotes = 
-    gsub "\"" "''" . -- rest are right quotes
+    gsub "\"" "''" . -- rest are right quotes .
+    gsub "``\\\\footnote" "''\\\\footnote" . -- except \footnote
+    gsub "\"\\\\" "``\\\\" . -- left quote before latex command
     gsub "([[:space:]])\"" "\\1``" . -- never right quote after space 
     gsub "\"('|`)([^[:punct:][:space:]])" "``{}`\\2" . -- "'word left
     gsub "\"([^[:punct:][:space:]])" "``\\1"  -- "word left
     
 escapeSingleQuotes =
+    gsub "`\\\\footnote" "'\\\\footnote" . -- except \footnote
+    gsub "'\\\\" "`\\\\" . -- left quote before latex command
     gsub "('|`)(\"|``)" "`{}``" .  -- '"word left
     gsub "([^[:punct:][:space:]])`(s|S)" "\\1'\\2" . -- catch possessives 
     gsub "^'([^[:punct:][:space:]])" "`\\1" .  -- 'word left