PEG: tweak commas and symbols

src/preserves/pegs.nim

@@ -9,7 +9,8 @@ when defined(nimHasUsed): {.used.}
 
 grammar "Preserves":
 
-  ws <- *(' ' | '\t' | '\r' | '\n' | ',')
+  ws <- *(' ' | '\t' | '\r' | '\n' )
+  commas <- *(ws * ',') * ws
 
   Document <- Value * ws * !1
 
@@ -22,13 +23,13 @@ grammar "Preserves":
 
   Atom <- Boolean | Float | Double | FloatRaw | DoubleRaw | SignedInteger | String | ByteString | Symbol
 
-  Record <- '<' * Value * *Value * ws * '>'
+  Record <- '<' * +Value * ws * '>'
 
-  Sequence <- '[' * *Value * ws * ']'
+  Sequence <- '[' * *(commas * Value) * commas * ']'
 
-  Dictionary <- '{' * *(Value * ':' * Value) * ws * '}'
+  Dictionary <- '{' * *(commas * Value * ws * ':' * Value) * commas * '}'
 
-  Set <- "#{" * *Value * ws * '}'
+  Set <- "#{" * *(commas * Value) * commas * '}'
 
   Boolean <- "#f" | "#t"
 
@@ -55,14 +56,12 @@ grammar "Preserves":
   binchar <- binunescaped | (escape * (escaped | '"' | ('x' * Xdigit[2])))
   binunescaped <- {' '..'!', '#'..'[', ']'..'~'}
 
-  symstart <- Alpha | sympunct | symustart
-  symcont <- Alpha | sympunct | symustart | symucont | Digit | '-'
-  sympunct <- {'~', '!', '$', '%', '^', '&', '*', '?', '_', '=', '+', '/', '.'}
-  symchar <- unescaped | '"' | (escape * (escaped | '|' | ('u' * Xdigit)))
-  symustart <- utf8.any - {0..127}
-  symucont <- utf8.any - {0..127}
-    # TODO: exclude some unicode ranges
-  Symbol <- >(symstart * *symcont) | ('|' * >(*symchar) * '|')
+  symchar <- (utf8.any - { 0..127, '\\', '|' }) | (escape * (escaped | ('u' * Xdigit[4]))) | "\\|"
+  QuotedSymbol <- '|' * >(*symchar) * '|'
+  sympunct <- {'~', '!', '$', '%', '^', '&', '*', '?', '_', '=', '+', '-', '/', '.'}
+  symuchar <- utf8.any - { 0..127 }
+  SymbolOrNumber <- >(+(Alpha | Digit | sympunct | symuchar))
+  Symbol <- QuotedSymbol | SymbolOrNumber
 
   Embedded <- "#!" * Value