src/gallium/state_trackers/python/retrace/parser.py

   1 #!/usr/bin/env python
   2 ##########################################################################
   3 #
   4 # Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas.
   5 # All Rights Reserved.
   6 #
   7 # Permission is hereby granted, free of charge, to any person obtaining a
   8 # copy of this software and associated documentation files (the
   9 # "Software"), to deal in the Software without restriction, including
  10 # without limitation the rights to use, copy, modify, merge, publish,
  11 # distribute, sub license, and/or sell copies of the Software, and to
  12 # permit persons to whom the Software is furnished to do so, subject to
  13 # the following conditions:
  14 #
  15 # The above copyright notice and this permission notice (including the
  16 # next paragraph) shall be included in all copies or substantial portions
  17 # of the Software.
  18 #
  19 # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  20 # OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  21 # MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
  22 # IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
  23 # ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
  24 # TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
  25 # SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  26 #
  27 ##########################################################################
  28
  29
  30 import sys
  31 import xml.parsers.expat
  32 import binascii
  33
  34 from model import *
  35
  36
  37 ELEMENT_START, ELEMENT_END, CHARACTER_DATA, EOF = range(4)
  38
  39
  40 class XmlToken:
  41
  42     def __init__(self, type, name_or_data, attrs = None, line = None, column = None):
  43         assert type in (ELEMENT_START, ELEMENT_END, CHARACTER_DATA, EOF)
  44         self.type = type
  45         self.name_or_data = name_or_data
  46         self.attrs = attrs
  47         self.line = line
  48         self.column = column
  49
  50     def __str__(self):
  51         if self.type == ELEMENT_START:
  52             return '<' + self.name_or_data + ' ...>'
  53         if self.type == ELEMENT_END:
  54             return '</' + self.name_or_data + '>'
  55         if self.type == CHARACTER_DATA:
  56             return self.name_or_data
  57         if self.type == EOF:
  58             return 'end of file'
  59         assert 0
  60
  61
  62 class XmlTokenizer:
  63     """Expat based XML tokenizer."""
  64
  65     def __init__(self, fp, skip_ws = True):
  66         self.fp = fp
  67         self.tokens = []
  68         self.index = 0
  69         self.final = False
  70         self.skip_ws = skip_ws
  71
  72         self.character_pos = 0, 0
  73         self.character_data = ''
  74
  75         self.parser = xml.parsers.expat.ParserCreate()
  76         self.parser.StartElementHandler  = self.handle_element_start
  77         self.parser.EndElementHandler    = self.handle_element_end
  78         self.parser.CharacterDataHandler = self.handle_character_data
  79
  80     def handle_element_start(self, name, attributes):
  81         self.finish_character_data()
  82         line, column = self.pos()
  83         token = XmlToken(ELEMENT_START, name, attributes, line, column)
  84         self.tokens.append(token)
  85
  86     def handle_element_end(self, name):
  87         self.finish_character_data()
  88         line, column = self.pos()
  89         token = XmlToken(ELEMENT_END, name, None, line, column)
  90         self.tokens.append(token)
  91
  92     def handle_character_data(self, data):
  93         if not self.character_data:
  94             self.character_pos = self.pos()
  95         self.character_data += data
  96
  97     def finish_character_data(self):
  98         if self.character_data:
  99             if not self.skip_ws or not self.character_data.isspace():
 100                 line, column = self.character_pos
 101                 token = XmlToken(CHARACTER_DATA, self.character_data, None, line, column)
 102                 self.tokens.append(token)
 103             self.character_data = ''
 104
 105     def next(self):
 106         size = 16*1024
 107         while self.index >= len(self.tokens) and not self.final:
 108             self.tokens = []
 109             self.index = 0
 110             data = self.fp.read(size)
 111             self.final = len(data) < size
 112             data = data.rstrip('\0')
 113             try:
 114                 self.parser.Parse(data, self.final)
 115             except xml.parsers.expat.ExpatError, e:
 116                 #if e.code == xml.parsers.expat.errors.XML_ERROR_NO_ELEMENTS:
 117                 if e.code == 3:
 118                     pass
 119                 else:
 120                     raise e
 121         if self.index >= len(self.tokens):
 122             line, column = self.pos()
 123             token = XmlToken(EOF, None, None, line, column)
 124         else:
 125             token = self.tokens[self.index]
 126             self.index += 1
 127         return token
 128
 129     def pos(self):
 130         return self.parser.CurrentLineNumber, self.parser.CurrentColumnNumber
 131
 132
 133 class TokenMismatch(Exception):
 134
 135     def __init__(self, expected, found):
 136         self.expected = expected
 137         self.found = found
 138
 139     def __str__(self):
 140         return '%u:%u: %s expected, %s found' % (self.found.line, self.found.column, str(self.expected), str(self.found))
 141
 142
 143
 144 class XmlParser:
 145     """Base XML document parser."""
 146
 147     def __init__(self, fp):
 148         self.tokenizer = XmlTokenizer(fp)
 149         self.consume()
 150
 151     def consume(self):
 152         self.token = self.tokenizer.next()
 153
 154     def match_element_start(self, name):
 155         return self.token.type == ELEMENT_START and self.token.name_or_data == name
 156
 157     def match_element_end(self, name):
 158         return self.token.type == ELEMENT_END and self.token.name_or_data == name
 159
 160     def element_start(self, name):
 161         while self.token.type == CHARACTER_DATA:
 162             self.consume()
 163         if self.token.type != ELEMENT_START:
 164             raise TokenMismatch(XmlToken(ELEMENT_START, name), self.token)
 165         if self.token.name_or_data != name:
 166             raise TokenMismatch(XmlToken(ELEMENT_START, name), self.token)
 167         attrs = self.token.attrs
 168         self.consume()
 169         return attrs
 170
 171     def element_end(self, name):
 172         while self.token.type == CHARACTER_DATA:
 173             self.consume()
 174         if self.token.type != ELEMENT_END:
 175             raise TokenMismatch(XmlToken(ELEMENT_END, name), self.token)
 176         if self.token.name_or_data != name:
 177             raise TokenMismatch(XmlToken(ELEMENT_END, name), self.token)
 178         self.consume()
 179
 180     def character_data(self, strip = True):
 181         data = ''
 182         while self.token.type == CHARACTER_DATA:
 183             data += self.token.name_or_data
 184             self.consume()
 185         if strip:
 186             data = data.strip()
 187         return data
 188
 189
 190 class TraceParser(XmlParser):
 191
 192     def parse(self):
 193         self.element_start('trace')
 194         while self.token.type not in (ELEMENT_END, EOF):
 195             call = self.parse_call()
 196             self.handle_call(call)
 197         if self.token.type != EOF:
 198             self.element_end('trace')
 199
 200     def parse_call(self):
 201         attrs = self.element_start('call')
 202         klass = attrs['class']
 203         method = attrs['method']
 204         args = []
 205         ret = None
 206         while self.token.type == ELEMENT_START:
 207             if self.token.name_or_data == 'arg':
 208                 arg = self.parse_arg()
 209                 args.append(arg)
 210             elif self.token.name_or_data == 'ret':
 211                 ret = self.parse_ret()
 212             elif self.token.name_or_data == 'call':
 213                 # ignore nested function calls
 214                 self.parse_call()
 215             else:
 216                 raise TokenMismatch("<arg ...> or <ret ...>", self.token)
 217         self.element_end('call')
 218
 219         return Call(klass, method, args, ret)
 220
 221     def parse_arg(self):
 222         attrs = self.element_start('arg')
 223         name = attrs['name']
 224         value = self.parse_value()
 225         self.element_end('arg')
 226
 227         return name, value
 228
 229     def parse_ret(self):
 230         attrs = self.element_start('ret')
 231         value = self.parse_value()
 232         self.element_end('ret')
 233
 234         return value
 235
 236     def parse_value(self):
 237         expected_tokens = ('null', 'bool', 'int', 'uint', 'float', 'string', 'enum', 'array', 'struct', 'ptr', 'bytes')
 238         if self.token.type == ELEMENT_START:
 239             if self.token.name_or_data in expected_tokens:
 240                 method = getattr(self, 'parse_' +  self.token.name_or_data)
 241                 return method()
 242         raise TokenMismatch(" or " .join(expected_tokens), self.token)
 243
 244     def parse_null(self):
 245         self.element_start('null')
 246         self.element_end('null')
 247         return Literal(None)
 248
 249     def parse_bool(self):
 250         self.element_start('bool')
 251         value = int(self.character_data())
 252         self.element_end('bool')
 253         return Literal(value)
 254
 255     def parse_int(self):
 256         self.element_start('int')
 257         value = int(self.character_data())
 258         self.element_end('int')
 259         return Literal(value)
 260
 261     def parse_uint(self):
 262         self.element_start('uint')
 263         value = int(self.character_data())
 264         self.element_end('uint')
 265         return Literal(value)
 266
 267     def parse_float(self):
 268         self.element_start('float')
 269         value = float(self.character_data())
 270         self.element_end('float')
 271         return Literal(value)
 272
 273     def parse_enum(self):
 274         self.element_start('enum')
 275         name = self.character_data()
 276         self.element_end('enum')
 277         return NamedConstant(name)
 278
 279     def parse_string(self):
 280         self.element_start('string')
 281         value = self.character_data()
 282         self.element_end('string')
 283         return Literal(value)
 284
 285     def parse_bytes(self):
 286         self.element_start('bytes')
 287         value = binascii.a2b_hex(self.character_data())
 288         self.element_end('bytes')
 289         return Literal(value)
 290
 291     def parse_array(self):
 292         self.element_start('array')
 293         elems = []
 294         while self.token.type != ELEMENT_END:
 295             elems.append(self.parse_elem())
 296         self.element_end('array')
 297         return Array(elems)
 298
 299     def parse_elem(self):
 300         self.element_start('elem')
 301         value = self.parse_value()
 302         self.element_end('elem')
 303         return value
 304
 305     def parse_struct(self):
 306         attrs = self.element_start('struct')
 307         name = attrs['name']
 308         members = []
 309         while self.token.type != ELEMENT_END:
 310             members.append(self.parse_member())
 311         self.element_end('struct')
 312         return Struct(name, members)
 313
 314     def parse_member(self):
 315         attrs = self.element_start('member')
 316         name = attrs['name']
 317         value = self.parse_value()
 318         self.element_end('member')
 319
 320         return name, value
 321
 322     def parse_ptr(self):
 323         self.element_start('ptr')
 324         address = self.character_data()
 325         self.element_end('ptr')
 326
 327         return Pointer(address)
 328
 329     def handle_call(self, call):
 330         pass
 331
 332
 333 class TraceDumper(TraceParser):
 334
 335     def __init__(self, fp):
 336         TraceParser.__init__(self, fp)
 337         self.formatter = format.DefaultFormatter(sys.stdout)
 338         self.pretty_printer = PrettyPrinter(self.formatter)
 339
 340     def handle_call(self, call):
 341         call.visit(self.pretty_printer)
 342         self.formatter.newline()
 343
 344
 345 def main(ParserFactory):
 346     for arg in sys.argv[1:]:
 347         if arg.endswith('.gz'):
 348             import gzip
 349             stream = gzip.GzipFile(arg, 'rt')
 350         else:
 351             stream = open(arg, 'rt')
 352         parser = ParserFactory(stream)
 353         parser.parse()
 354
 355
 356 if __name__ == '__main__':
 357     main(TraceDumper)