checkpoint

[sbp.git] / tests / tibdoc.g

diff --git a/tests/tibdoc.g b/tests/tibdoc.g
index df52b10..9cff556 100644 (file)
--- a/tests/tibdoc.g
+++ b/tests/tibdoc.g
@@ -1,10 +1,14 @@
+// indentation styling...
+// literal blocks [[need to ignore bracing]] double-colon style?
+// definition -- by prior line indentation, like headings in the original structured text
+// tables
+// dropcap
+// output formats: latex, contex, ps, pdf, html, man, txt, rfc
 
 // URGENT: why does swapping [a-zA-Z0-9] for alphanum in "item" cause severe breakage?
 // URGENT: why does enabling "url" cause severe breakage? (probably same cause)
 //   ... something to do with unit productions
 
-
-
 // output formats: latex, contex, ps, pdf, html, man, txt, rfc
 // bullet list
 // numbered list
@@ -26,7 +30,7 @@
 // FIXME: these have to go at the top so they have their dropAll bit set before PreSequence.build...
 ws      !::= w**
 w       !::=  [\r\n\ ]
-nw       ::= [~\r\n\ ]
+nw       ::= ~[\r\n\ ]
 
 //////////////////////////////////////////////////////////////////////////////
 
@@ -35,21 +39,29 @@ s                 ::= {Doc} => top
 Doc               ::= Header   Body                       /ws => doc
 Header            ::= "header" { kv */ ws }               /ws => header
 Body              ::= Section*/ws                             => body
-Section           ::= SectionHeader ({text}*/ws)          /ws => section
+Section           ::= SectionHeader Paragraph*/ws         /ws => section
 SectionHeader     ::= "==" SectionHeaderBody "=="
 SectionHeaderBody ::=  "=" SectionHeaderBody "="
                     >      ws text ws
 
 kv         ::= word "=" text /ws => kv1
 
+num !::= [0-9]++ => "stringify"
+Paragraph  ::= { "\"\"" ws text }        => "blockquote"
+             > { "*" " " ws text }       => "ul"
+             > { "#" " " ws text }       => "ol"
+             > { num " " ws text }       => "ol"
+             > { "---" "-"* }            => "hr"
+             > { text }                  => "p"
+
 text       ::= item */ ws
-pre        ::= "[verbatim]" { [~]+ } /ws => "verbatim"   // FIXME doesn't work
+pre        ::= "[verbatim]" { ~[]+ } /ws => "verbatim"   // FIXME doesn't work
 item       ::= pre
              > email
              > structured
              > styled
              > "\"" text "\""    => quoted
-             > [a-zA-Z0-9]++
+             > alphanum++        => "stringify"
              > symbol
 
 symbol     ::= symbolx & sym++
@@ -76,8 +88,8 @@ glyph      ::= "(r)" | "(c)" | "(tm)"  // euro symbol?
 // only gets parsed once
 
 urlpath    ::= urlchar*
-username   ::= [a-zA-Z0-9;/?:&=$\-_.+]++
-password   ::= [a-zA-Z0-9;/?:&=$\-_.+]++
+username   ::= [a-zA-Z0-9;/?:&=$\-_.+]++ => "stringify"
+password   ::= [a-zA-Z0-9;/?:&=$\-_.+]++ => "stringify"
 urlchar    ::= [a-zA-Z0-9;/?:&=$\-_.+@]
              | "%" [0-9] [0-9]       => "%"
 url        ::= "mailto" ":"   email
@@ -87,8 +99,8 @@ method     ::= [+\-.a-z0-9]+
 port       ::= [0-9]+
 
 domain     ::= part +/ "."
-part       ::= [a-zA-Z0-9\-]++    // interesting use of boolean grammars
-//            &~ ([\-0-9] [~]* | [~]* [\-0-9])
+part       ::= [a-zA-Z0-9\-]++ => "stringify"   // interesting use of boolean grammars
+//            &~ ([\-0-9] ~[]* | ~[]* [\-0-9])
 
 email      ::= username "@" host      => email
 host       ::= [0-9]+ "." [0-9]+ "." [0-9]+ "." [0-9]+ => "ip"
@@ -98,14 +110,14 @@ host       ::= [0-9]+ "." [0-9]+ "." [0-9]+ "." [0-9]+ => "ip"
 
 // Tokens ///////////////////////////////////////////////////////////////////
 
-word     ::= alphanum++
+word     ::= alphanum++ => "stringify"
            | quoted
 
-quoted   ::= "\"" (([~\"\\] | escaped)+) "\""
+quoted   ::= "\"" ((~[\"\\] | escaped)+) "\""
            | "\"\"" => ""
 escaped  ::= "\\n" => "\n"
            | "\\r" => "\r"
-           | "\\" [~nr]
+           | "\\" ~[nr]
 
 
 // Chars ///////////////////////////////////////////////////////////////
@@ -113,6 +125,6 @@ escaped  ::= "\\n" => "\n"
 alpha    ::= [a-zA-Z]
 num      ::= [0-9]
 alphanum ::= [a-zA-Z0-9]
-sym      ::= [~a-zA-Z0-9\ \r\n]
+sym      ::= ~[a-zA-Z0-9\ \r\n]