lib/dom/parser.nit

   1 # This file is part of NIT ( http://www.nitlanguage.org ).
   2 #
   3 # This file is free software, which comes along with NIT.  This software is
   4 # distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
   5 # without  even  the implied warranty of  MERCHANTABILITY or  FITNESS FOR A
   6 # PARTICULAR PURPOSE.  You can modify it is you want,  provided this header
   7 # is kept unaltered, and a notification of the changes is added.
   8 # You  are  allowed  to  redistribute it and sell it, alone or is a part of
   9 # another product.
  10
  11 # XML DOM-parsing facilities
  12 module parser
  13
  14 intrude import parser_base
  15 intrude import xml_entities
  16
  17 # Provides XML parsing facilities
  18 class XMLProcessor
  19         super StringProcessor
  20
  21         # Parses a full XML document
  22         fun parse_document: XMLEntity do
  23                 var stack = new Array[XMLStartTag]
  24                 var doc = new XMLDocument
  25                 loop
  26                         ignore_whitespaces
  27                         if pos >= src.length then break
  28                         if src[pos] == '<' then
  29                                 var tag = read_tag
  30                                 if tag isa XMLStartTag then
  31                                         if stack.is_empty then
  32                                                 tag.parent = doc
  33                                         else
  34                                                 var st_last = stack.last
  35                                                 tag.parent = st_last
  36                                         end
  37                                         stack.push tag
  38                                 else if tag isa XMLEndTag then
  39                                         if stack.is_empty then
  40                                                 return new XMLError(tag.location, "Missing matching tag for `{tag.tag_name}`")
  41                                         end
  42                                         var st_last = stack.last
  43                                         if tag.tag_name == st_last.tag_name then
  44                                                 st_last.matching = tag
  45                                                 tag.matching = st_last
  46                                                 stack.pop
  47                                         else
  48                                                 var miss = stack.pop
  49                                                 return new XMLError(miss.location, "Missing matching tag for `{miss.tag_name}`")
  50                                         end
  51                                 else if tag isa XMLError then
  52                                         return tag
  53                                 else
  54                                         if stack.is_empty then
  55                                                 tag.parent = doc
  56                                         else
  57                                                 tag.parent = stack.last
  58                                         end
  59                                 end
  60                         else
  61                                 var st = pos
  62                                 var end_pc = ignore_until("<") - 1
  63                                 var loc = new Location(line, line_offset)
  64                                 var pc = new PCDATA(loc, src.substring(st, end_pc - st + 1).trim)
  65                                 if stack.is_empty then
  66                                         pc.parent = doc
  67                                 else
  68                                         pc.parent = stack.last
  69                                 end
  70                         end
  71                 end
  72                 if not stack.is_empty then
  73                         var miss = stack.pop
  74                         return new XMLError(miss.location, "Missing matching tag for `{miss.tag_name}`")
  75                 end
  76                 return doc
  77         end
  78
  79         # Reads the tag starting in `src` at current position
  80         private fun read_tag: XMLEntity do
  81                 var st_loc = new Location(line, line_offset)
  82                 var c = src[pos]
  83                 if not c == '<' then return new XMLError(st_loc, "Expected start of tag, got `{c}`")
  84                 pos += 1
  85                 if pos >= src.length then return new XMLError(st_loc, "Malformed tag")
  86                 c = src[pos]
  87                 if c == '!' then
  88                         # Special tag
  89                         return read_special_tag(st_loc)
  90                 else if c == '?' then
  91                         # Prolog tag
  92                         return read_prolog_tag(st_loc)
  93                 else if c == '/' then
  94                         # End tag
  95                         return read_end_tag(st_loc)
  96                 else
  97                         # Start tag
  98                         return read_start_tag(st_loc)
  99                 end
 100         end
 101
 102         # Reads a Special tag (starting with <!)
 103         #
 104         # In case of error, returns a `XMLError`
 105         private fun read_special_tag(st_loc: Location): XMLEntity do
 106                 var srclen = src.length
 107                 pos += 1
 108                 if (pos + 2) >= srclen then return new XMLError(st_loc, "Unexpected EOF on start of Special tag")
 109                 if src[pos] == '-' and src[pos + 1] == '-' then
 110                         pos += 2
 111                         var comst = pos
 112                         var endcom = ignore_until("-->")
 113                         if endcom == -1 then return new XMLError(st_loc, "Malformed comment")
 114                         pos += 3
 115                         return new XMLCommentTag(st_loc ,src.substring(comst, endcom - comst + 1))
 116                 end
 117                 var st = pos
 118                 if srclen - pos >= 7 then
 119                         var spe_type = src.substring(pos, 7)
 120                         if spe_type == "[CDATA[" then
 121                                 pos += 7
 122                                 var cdst = pos
 123                                 var cdend = ignore_until("]]>")
 124                                 pos += 3
 125                                 if pos >= srclen then return new XMLError(st_loc, "Unfinished CDATA block")
 126                                 return new CDATA(st_loc, src.substring(cdst, cdend - cdst))
 127                         else if spe_type == "DOCTYPE" then
 128                                 pos += 7
 129                                 return parse_doctype(st_loc)
 130                         end
 131                 end
 132                 var end_spec = ignore_until(">")
 133                 pos += 1
 134                 return new XMLSpecialTag(st_loc, src.substring(st, end_spec - st))
 135         end
 136
 137         # Parse a Doctype declaration tag
 138         private fun parse_doctype(st_loc: Location): XMLEntity do
 139                 var elemts = new Array[String]
 140                 var srclen = src.length
 141                 loop
 142                         ignore_whitespaces
 143                         if pos >= srclen then return new XMLError(st_loc, "Malformed doctype")
 144                         var c = src[pos]
 145                         # TODO: Properly support intern DOCTYPE definitions
 146                         if c == '[' then
 147                                 var intern_st = pos
 148                                 var intern_end = ignore_until("]")
 149                                 if intern_end == -1 then return new XMLError(st_loc, "Unfinished internal doctype declaration")
 150                                 pos += 1
 151                                 elemts.push src.substring(intern_st, intern_end - intern_st + 1)
 152                                 continue
 153                         end
 154                         var elm_st = pos
 155                         while pos < srclen and not src[pos].is_whitespace and src[pos] != '>' do pos += 1
 156                         if pos >= srclen then return new XMLError(st_loc, "Malformed doctype")
 157                         if pos - elm_st > 1 then
 158                                 var str = src.substring(elm_st, pos - elm_st)
 159                                 elemts.push str
 160                         end
 161                         if src[pos] == '>' then
 162                                 pos += 1
 163                                 return new XMLDoctypeTag(st_loc, "DOCTYPE", elemts.join(" "))
 164                         end
 165                 end
 166         end
 167
 168         # Reads a Prolog or Processing Instruction tag (starting with <?)
 169         #
 170         # In case of error, returns a `XMLError`
 171         private fun read_prolog_tag(st_loc: Location): XMLEntity do
 172                 var srclen = src.length
 173                 pos += 1
 174                 if pos >= srclen then return new XMLError(st_loc, "Invalid start of prolog")
 175                 var tag_name = parse_tag_name(['<', '>'])
 176                 var c = src[pos]
 177                 if c == '<' or c == '>' then return new XMLError(st_loc ,"Unexpected character `{c}` in prolog declaration")
 178                 if tag_name == "xml" then
 179                         var args = parse_args(['?'])
 180                         for i in args do
 181                                 if i isa BadXMLAttribute then return new XMLError(i.location, i.name)
 182                         end
 183                         if src[pos] == '?' then
 184                                 if src[pos + 1] == '>' then
 185                                         pos += 2
 186                                         return new XMLPrologTag(st_loc, tag_name, args)
 187                                 end
 188                         end
 189                 else
 190                         if tag_name.has("xml") then return new XMLError(st_loc, "Forbidden keyword xml in Processing Instruction")
 191                         var cont_st = pos
 192                         var cont_end = ignore_until("?>")
 193                         if cont_end == -1 then
 194                                 pos += 2
 195                                 return new XMLError(st_loc, "Malformed Processing Instruction tag")
 196                         end
 197                         pos += 2
 198                         return new XMLProcessingInstructionTag(st_loc, tag_name, src.substring(cont_st, cont_end - cont_st))
 199                 end
 200                 pos += 1
 201                 return new XMLError(st_loc, "Malformed prolog tag")
 202         end
 203
 204         # Reads an End tag (starting with </)
 205         #
 206         # In case of error, returns a `XMLError`
 207         private fun read_end_tag(st_loc: Location): XMLEntity do
 208                 pos += 1
 209                 var tag_name = parse_tag_name(['<', '>'])
 210                 ignore_whitespaces
 211                 if src[pos] == '>' then
 212                         pos += 1
 213                         return new XMLEndTag(st_loc, tag_name)
 214                 end
 215                 return new XMLError(st_loc, "Bad end tag `{tag_name}`")
 216         end
 217
 218         # Reads a Start tag (starting with <)
 219         #
 220         # In case of error, returns a `XMLError`
 221         private fun read_start_tag(st_loc: Location): XMLEntity do
 222                 var tag_name = parse_tag_name(['/', '>'])
 223                 var args = parse_args(['/', '>'])
 224                 for i in args do
 225                         if i isa BadXMLAttribute then return new XMLError(i.location, i.name)
 226                 end
 227                 if src[pos] == '/' then
 228                         if src[pos + 1] == '>' then
 229                                 pos += 2
 230                                 return new XMLOnelinerTag(st_loc, tag_name, args)
 231                         end
 232                 end
 233                 pos += 1
 234                 return new XMLStartTag(st_loc, tag_name, args)
 235         end
 236
 237         # Parses an xml tag name
 238         private fun parse_tag_name(delims: Array[Char]): String do
 239                 var idst = pos
 240                 var srclen = src.length
 241                 while pos < srclen do
 242                         var c = src[pos]
 243                         if c.is_whitespace or delims.has(c) then break
 244                         pos += 1
 245                 end
 246                 return src.substring(idst, pos - idst).trim
 247         end
 248
 249         # Parse the arguments of a tag
 250         private fun parse_args(endtags: Array[Char]): Array[XMLAttribute] do
 251                 var attrs = new Array[XMLAttribute]
 252                 loop
 253                         var arg = parse_arg(endtags)
 254                         if arg isa XMLAttributeEnd then return attrs
 255                         attrs.add arg
 256                         if arg isa BadXMLAttribute then return attrs
 257                 end
 258         end
 259
 260         # Parses the next argument in `src`
 261         private fun parse_arg(endtags: Array[Char]): XMLAttribute do
 262                 var srclen = src.length
 263                 ignore_whitespaces
 264                 var st_loc = new Location(line, line_offset)
 265                 if pos >= srclen then return new BadXMLAttribute(st_loc, "Unfinished attribute name")
 266                 # FIXME: Ugly, but as long as it remains private, it is OK I guess
 267                 if endtags.has(src[pos]) then return new XMLAttributeEnd(st_loc, "")
 268                 var attrname_st = pos
 269                 while pos < srclen and src[pos] != '=' and not endtags.has(src[pos]) do pos += 1
 270                 if pos >= srclen then return new BadXMLAttribute(st_loc, "Unfinished attribute name")
 271                 if src[pos] != '=' then return new BadXMLAttribute(st_loc, "Malformed attribute")
 272                 var attrname_end = pos - 1
 273                 var name = src.substring(attrname_st, attrname_end - attrname_st + 1).trim
 274                 pos += 1
 275                 ignore_whitespaces
 276                 var attrval_st = pos
 277                 if pos >= srclen then return new BadXMLAttribute(st_loc, "Unfinished attribute `{name}`")
 278                 var match = src[pos]
 279                 if match != '\'' and match != '"' then return new BadXMLAttribute(st_loc, "Invalid string delimiter `{match}` for attribute `{name}`")
 280                 pos += 1
 281                 while pos < srclen and src[pos] != match do pos += 1
 282                 if pos >= srclen then return new BadXMLAttribute(st_loc, "Unfinished attribute `{name}`")
 283                 var attrval_end = pos
 284                 var val = src.substring(attrval_st, attrval_end - attrval_st + 1).trim
 285                 pos += 1
 286                 return new XMLStringAttr(st_loc, name, val.substring(1, val.length - 2), match)
 287         end
 288 end
 289
 290 redef class Text
 291         # Tries to parse the current string to XML
 292         #
 293         # Returns an `XMLDocument` if successful, or an `XMLError` if not
 294         fun to_xml: XMLEntity do return (new XMLProcessor(self.to_s)).parse_document
 295 end