def __prepare_buffer_element_specification(setup):
global global_character_type_db
if Setup.buffer_lexatom_size_in_byte == "wchar_t":
error.log("Since Quex version 0.53.5, 'wchar_t' can no longer be specified\n"
"with option '--buffer-element-size' or '-bes'. Please, specify\n"
"'--buffer-element-type wchar_t' or '--bet'.")
if Setup.buffer_lexatom_type == "wchar_t":
Setup.converter_ucs_coding_name = "WCHAR_T"
# (*) Determine buffer element type and size (in bytes)
lexatom_size_in_byte = Setup.buffer_lexatom_size_in_byte
if lexatom_size_in_byte == -1:
if global_character_type_db.has_key(Setup.buffer_lexatom_type):
lexatom_size_in_byte = global_character_type_db[Setup.buffer_lexatom_type][3]
elif Setup.buffer_lexatom_type == "":
lexatom_size_in_byte = 1
else:
# Buffer element type is not identified in 'global_character_type_db'.
# => here Quex cannot know its size on its own.
lexatom_size_in_byte = -1
if Setup.buffer_lexatom_type == "":
if lexatom_size_in_byte in [1, 2, 4]:
Setup.buffer_lexatom_type = {
1: "uint8_t", 2: "uint16_t", 4: "uint32_t",
}[lexatom_size_in_byte]
elif lexatom_size_in_byte == -1:
pass
else:
error.log("Buffer element type cannot be determined for size '%i' which\n" \
% lexatom_size_in_byte +
"has been specified by '-b' or '--buffer-element-size'.")
return lexatom_size_in_byte
def do(setup, command_line, argv):
"""Does a consistency check for setup and the command line.
"""
setup.output_directory = os.path.normpath(setup.output_directory)
if setup.output_directory:
# Check, if the output directory exists
if os.access(setup.output_directory, os.F_OK) == False:
error.log(
"The directory %s was specified for output, but does not exists."
% setup.output_directory)
if os.access(setup.output_directory, os.W_OK) == False:
error.log(
"The directory %s was specified for output, but is not writeable."
% setup.output_directory)
# if the mode is '--language dot' => check character display options.
if setup.character_display not in ["hex", "utf8"]:
error.log(
"Character display must be either 'hex' or 'utf8'.\nFound: '%s'" %
setup.character_display)
# ensure that options are not specified twice
for parameter, info in SETUP_INFO.items():
if type(info) != list: continue
occurence_n = 0
for option in info[0]:
occurence_n += argv.count(option)
if occurence_n > 1 and info[1] not in (SetupParTypes.LIST,
SetupParTypes.INT_LIST):
error.log("Received more than one of the following options:\n" + \
"%s" % repr(info[0])[1:-1])
# (*) Check for 'Depraceted' Options ___________________________________________________
for name, info in DEPRECATED.items():
command_line_options = SETUP_INFO[name][0]
comment = info[0]
depreciated_since_version = info[1]
for option in command_line_options:
if command_line.search(option):
error.log("Command line option '%s' is ignored.\n" % option + \
comment + "\n" + \
"Last version of Quex supporting this option is version %s. Please, visit\n" % \
depreciated_since_version + \
"http://quex.org for further information.")
# (*) Check for 'Straying' Options ___________________________________________________
options = []
for key, info in SETUP_INFO.items():
if type(info) != list: continue
if key in DEPRECATED: continue
if info[1] is not None: options.extend(info[0])
options.sort(lambda a, b: cmp(a.replace("-", ""), b.replace("-", "")))
ufos = command_line.unidentified_options(options)
if len(ufos) != 0:
error.log("Unidentified option(s) = " + repr(ufos) + "\n" + \
__get_supported_command_line_option_description(options))
if setup.analyzer_derived_class_name != "" and \
setup.analyzer_derived_class_file == "":
error.log("Specified derived class '%s' on command line, but it was not\n" % \
setup.analyzer_derived_class_name + \
"specified which file contains the definition of it.\n" + \
"use command line option '--derived-class-file'.\n")
if setup.buffer_element_size not in [-1, 1, 2, 4]:
error.log(
"The setting of '--buffer-element-size' (or '-b') can only be\n"
"1, 2, or 4 (found %s)." % repr(setup.buffer_element_size))
if setup.buffer_byte_order not in ["<system>", "little", "big"]:
error.log("Byte order (option --endian) must be 'little', 'big', or '<system>'.\n" + \
"Note, that this option is only interesting for cross plattform development.\n" + \
"By default, quex automatically chooses the endian type of your system.")
# Manually written token class requires token class name to be specified
if setup.token_class_file != "" and command_line.search(
"--token-class", "--tc") == False:
error.log(
"The use of a manually written token class requires that the name of the class\n"
"is specified on the command line via the '--token-class' option.")
# Token queue
if setup.token_policy != "queue" and command_line.search(
"--token-queue-size"):
error.log("Option --token-queue-size determines a fixed token queue size. This makes\n" + \
"only sense in conjunction with '--token-policy queue'.\n")
if setup.token_queue_size <= setup.token_queue_safety_border + 1:
if setup.token_queue_size == setup.token_queue_safety_border:
cmp_str = "equal to"
else:
cmp_str = "less than"
error.log("Token queue size is %i is %s token queue safety border %i + 1.\n" % \
(setup.token_queue_size, cmp_str, setup.token_queue_safety_border) +
"Set appropriate values with --token-queue-size and --token-queue-safety-border.")
# Check that names are valid identifiers
if len(setup.token_id_prefix_plain) != 0:
__check_identifier(setup, "token_id_prefix_plain", "Token prefix")
__check_identifier(setup, "analyzer_class_name", "Engine name")
if setup.analyzer_derived_class_name != "":
__check_identifier(setup, "analyzer_derived_class_name",
"Derived class name")
__check_file_name(setup, "token_class_file",
"file containing token class definition")
__check_file_name(setup, "analyzer_derived_class_file",
"file containing user derived lexer class")
__check_file_name(
setup,
"token_id_foreign_definition_file",
"file containing user token ids",
0,
CommandLineOption=SETUP_INFO["token_id_foreign_definition"][0])
__check_file_name(setup, "input_mode_files", "quex source file")
# Check that not more than one converter is specified
converter_n = 0
if setup.converter_iconv_f: converter_n += 1
if setup.converter_icu_f: converter_n += 1
if len(setup.converter_user_new_func) != 0: converter_n += 1
if converter_n > 1:
error.log("More than one character converter has been specified. Note, that the\n" + \
"options '--icu', '--iconv', and '--converter-new' (or '--cn') are\n" + \
"to be used mutually exclusively.")
if converter_n == 1 and setup.buffer_codec.name != "unicode":
# If the buffer codec is other than unicode, then no converter shall
# be used to fill the buffer. Instead, the engine is transformed, so
# that it works directly on the codec.
error.log("An engine that is to be generated for a specific codec cannot rely\n" + \
"on converters. Do no use '--codec' together with '--icu', '--iconv', or\n" + \
"`--converter-new`.")
# If a converter has been specified and no bytes-element-size has been specified,
# it defaults to '1 byte' which is most likely not what is desired for unicode.
if converter_n == 1 \
and setup.buffer_element_size == 1 \
and not command_line_args_defined(command_line, "buffer_element_size") \
and not command_line_args_defined(command_line, "buffer_element_type"):
error.log("A converter has been specified, but the default buffer element size\n" + \
"is left to 1 byte. Consider %s or %s." \
% (command_line_args_string("buffer_element_size"),
command_line_args_string("buffer_element_type")))
# If a user defined type is specified for 'engine character type' and
# a converter, then the name of the target type must be specified explicitly.
if setup.buffer_element_type != "" \
and not global_character_type_db.has_key(setup.buffer_element_type) \
and setup.converter_ucs_coding_name == "" \
and converter_n != 0:
tc = setup.buffer_element_type
error.log("A character code converter has been specified. It is supposed to convert\n" + \
"incoming data into an internal buffer of unicode characters. The size of\n" + \
"each character is determined by '%s' which is a user defined type.\n" % tc + \
"\n" + \
"Quex cannot determine automatically the name that the converter requires\n" + \
"to produce unicode characters for type '%s'. It must be specified by the\n" % tc + \
"command line option %s." \
% command_line_args_string("converter_ucs_coding_name"))
# Token transmission policy
token_policy_list = ["queue", "single", "users_token", "users_queue"]
if setup.token_policy not in token_policy_list:
error.log("Token policy '%s' not supported. Use one of the following:\n" % setup.token_policy + \
repr(token_policy_list)[1:-1])
elif setup.token_policy == "users_token":
error.log(
"Token policy 'users_queue' has be deprecated since 0.49.1. Use\n"
"equivalent policy 'single'.")
elif setup.token_policy == "users_queue":
error.log(
"Token policy 'users_queue' has be deprecated since 0.49.1\n")
# Internal engine character encoding
def __codec_vs_buffer_element_size(CodecName, RequiredBufferElementSize):
if setup.buffer_codec.name != CodecName: return
elif setup.buffer_element_size == RequiredBufferElementSize: return
if setup.buffer_element_size == -1:
msg_str = "undetermined (found type '%s')" % setup.buffer_element_type
else:
msg_str = "is not %i (found %i)" % (RequiredBufferElementSize,
setup.buffer_element_size)
error.log("Using codec '%s' while buffer element size %s.\n" % (CodecName, msg_str) +
"Consult command line argument %s" \
% command_line_args_string("buffer_element_size"))
if setup.buffer_codec.name != "unicode":
if not setup.buffer_codec_file:
error.verify_word_in_list(
setup.buffer_codec_name,
codec_db.get_supported_codec_list() + ["utf8", "utf16"],
"Codec '%s' is not supported." % setup.buffer_codec.name)
__codec_vs_buffer_element_size("utf8", 1)
__codec_vs_buffer_element_size("utf16", 2)
if setup.external_lexeme_null_object and setup.token_class_only_f:
error.log(
"Specifying an external lexeme null object signalizes an\n"
"external token class implementation. The 'token class only\n"
"flag' generates a token class considered to be externally\n"
"shared. Both flags are mutually exclusive.")
if setup.string_accumulator_f:
error_n = NotificationDB.warning_on_no_token_class_take_text
if error_n in setup.suppressed_notification_list:
error.warning(
"The warning upon missing 'take_text' in token type definition is de-\n"
+
"activated by '--suppress %i'. This is dangerous, if there is a string\n"
% error_n +
"accumulator. May be, use '--no-string-accumulator'.",
-1,
SuppressCode=NotificationDB.
warning_on_no_warning_on_missing_take_text)
def do(setup, command_line, argv):
"""Does a consistency check for setup and the command line.
"""
setup.output_directory = os.path.normpath(setup.output_directory)
if setup.output_directory != "":
# Check, if the output directory exists
if os.access(setup.output_directory, os.F_OK) == False:
error_msg("The directory %s was specified for output, but does not exists." % setup.output_directory)
if os.access(setup.output_directory, os.W_OK) == False:
error_msg("The directory %s was specified for output, but is not writeable." % setup.output_directory)
# if the mode is 'plotting', then check wether a graphic format is speicified
for plot_option in SETUP_INFO["plot_graphic_format"][0]:
if plot_option in argv and setup.plot_graphic_format == "":
error_msg("Option '%s' must be followed by a graphic format specifier (bmp, svg, jpg, ...)" % \
plot_option)
if setup.plot_character_display not in ["hex", "utf8"]:
error_msg("Plot character display must be either 'hex' or 'utf8'.\nFound: '%s'" %
setup.plot_character_display)
# ensure that options are not specified twice
for parameter, info in SETUP_INFO.items():
if type(info) != list: continue
occurence_n = 0
for option in info[0]:
occurence_n += argv.count(option)
if occurence_n > 1:
error_msg("Received more than one of the following options:\n" + \
"%s" % repr(info[0])[1:-1])
# (*) Check for 'Depraceted' Options ___________________________________________________
for name, info in DEPRECATED.items():
command_line_options = SETUP_INFO[name][0]
comment = info[0]
depreciated_since_version = info[1]
for option in command_line_options:
if command_line.search(option):
error_msg("Command line option '%s' is ignored.\n" % option + \
comment + "\n" + \
"Last version of Quex supporting this option is version %s. Please, visit\n" % \
depreciated_since_version + \
"http://quex.org for further information.")
# (*) Check for 'Straying' Options ___________________________________________________
options = []
for key, info in SETUP_INFO.items():
if type(info) != list: continue
if key in DEPRECATED: continue
if info[1] != None: options.extend(info[0])
options.sort(lambda a,b: cmp(a.replace("-",""), b.replace("-","")))
ufos = command_line.unidentified_options(options)
if ufos != []:
error_msg("Unidentified option(s) = " + repr(ufos) + "\n" + \
__get_supported_command_line_option_description(options))
if setup.analyzer_derived_class_name != "" and \
setup.analyzer_derived_class_file == "":
error_msg("Specified derived class '%s' on command line, but it was not\n" % \
setup.analyzer_derived_class_name + \
"specified which file contains the definition of it.\n" + \
"use command line option '--derived-class-file'.\n")
if setup.buffer_element_size not in [-1, 1, 2, 4]:
error_msg("The setting of '--buffer-element-size' (or '-b') can only be\n"
"1, 2, or 4 (found %s)." % repr(setup.buffer_element_size))
if setup.buffer_byte_order not in ["<system>", "little", "big"]:
error_msg("Byte order (option --endian) must be 'little', 'big', or '<system>'.\n" + \
"Note, that this option is only interesting for cross plattform development.\n" + \
"By default, quex automatically chooses the endian type of your system.")
# Manually written token class requires token class name to be specified
if setup.token_class_file != "" and command_line.search("--token-class", "--tc") == False:
error_msg("The use of a manually written token class requires that the name of the class\n"
"is specified on the command line via the '--token-class' option.")
# Token queue
if setup.token_policy != "queue" and command_line.search("--token-queue-size"):
error_msg("Option --token-queue-size determines a fixed token queue size. This makes\n" + \
"only sense in conjunction with '--token-policy queue'.\n")
if setup.token_queue_size <= setup.token_queue_safety_border + 1:
if setup.token_queue_size == setup.token_queue_safety_border: cmp_str = "equal to"
else: cmp_str = "less than"
error_msg("Token queue size is %i is %s token queue safety border %i + 1.\n" % \
(setup.token_queue_size, cmp_str, setup.token_queue_safety_border) +
"Set appropriate values with --token-queue-size and --token-queue-safety-border.")
# Check that names are valid identifiers
__check_identifier(setup, "token_id_prefix_plain", "Token prefix")
__check_identifier(setup, "analyzer_class_name", "Engine name")
if setup.analyzer_derived_class_name != "":
__check_identifier(setup, "analyzer_derived_class_name", "Derived class name")
__check_file_name(setup, "token_class_file", "file containing token class definition")
__check_file_name(setup, "analyzer_derived_class_file", "file containing user derived lexer class")
__check_file_name(setup, "token_id_foreign_definition_file", "file containing user token ids")
__check_file_name(setup, "input_mode_files", "quex source file")
# Check that not more than one converter is specified
converter_n = 0
if setup.converter_iconv_f: converter_n += 1
if setup.converter_icu_f: converter_n += 1
if setup.converter_user_new_func != "": converter_n += 1
if converter_n > 1:
error_msg("More than one character converter has been specified. Note, that the\n" + \
"options '--icu', '--iconv', and '--converter-new' (or '--cn') are\n" + \
"to be used mutually exclusively.")
if converter_n == 1 and setup.buffer_codec != "":
error_msg("An engine that is to be generated for a specific codec cannot rely\n" + \
"on converters. Do no use '--codec' together with '--icu', '--iconv', or\n" + \
"`--converter-new`.")
# If a user defined type is specified for 'engine character type' and
# a converter, then the name of the target type must be specified explicitly.
if setup.buffer_element_type != "" \
and not global_character_type_db.has_key(setup.buffer_element_type) \
and setup.converter_ucs_coding_name == "" \
and converter_n != 0:
tc = setup.buffer_element_type
error_msg("A character code converter has been specified. It is supposed to convert\n" + \
"incoming data into an internal buffer of unicode characters. The size of\n" + \
"each character is determined by '%s' which is a user defined type.\n" % tc + \
"\n" + \
"Quex cannot determine automatically the name that the converter requires\n" + \
"to produce unicode characters for type '%s'. It must be specified by the\n" % tc + \
"command line option '--converter-ucs-coding-name' or '--cucn'.")
# Token transmission policy
token_policy_list = ["queue", "single", "users_token", "users_queue"]
if setup.token_policy not in token_policy_list:
error_msg("Token policy '%s' not supported. Use one of the following:\n" % setup.token_policy + \
repr(token_policy_list)[1:-1])
elif setup.token_policy == "users_token":
error_msg("Token policy 'users_queue' has be deprecated since 0.49.1. Use\n"
"equivalent policy 'single'.")
elif setup.token_policy == "users_queue":
error_msg("Token policy 'users_queue' has be deprecated since 0.49.1\n")
# Internal engine character encoding
def __codec_vs_buffer_element_size(CodecName, RequiredBufferElementSize):
if setup.buffer_codec != CodecName: return
elif setup.buffer_element_size == RequiredBufferElementSize: return
if setup.buffer_element_size == -1:
msg_str = "undetermined (found type '%s')" % setup.buffer_element_type
else:
msg_str = "is not %i (found %i)" % (RequiredBufferElementSize, setup.buffer_element_size)
error_msg("Using codec '%s' while buffer element size %s.\n" % (CodecName, msg_str) +
"Consult command line argument '--buffer-element-size'.")
if setup.buffer_codec != "":
verify_word_in_list(setup.buffer_codec,
codec_db.get_supported_codec_list() + ["utf8", "utf16"],
"Codec '%s' is not supported." % setup.buffer_codec)
__codec_vs_buffer_element_size("utf8", 1)
__codec_vs_buffer_element_size("utf16", 2)
# Path Compression
if setup.compression_path_uniform_f and setup.compression_path_f:
error_msg("Both flags for path compression were set: '--path-compression' and\n"
"'--path-compression-uniform'. Please, choose only one!")
def prepare(command_line, argv):
"""RETURN: True, if process needs to be started.
False, if job is done.
"""
global Setup
# (*) Classes and their namespace
__setup_analyzer_class(Setup)
__setup_token_class(Setup)
__setup_token_id_prefix(Setup)
__setup_lexeme_null(Setup) # Requires 'token_class_name_space'
# (*) Output programming language
Setup.language = Setup.language.upper()
error.verify_word_in_list(Setup.language, output_language_db.keys(),
"Programming language '%s' is not supported." % Setup.language)
Setup.language_db = output_language_db[Setup.language]
Setup.extension_db = global_extension_db[Setup.language]
# Is the output file naming scheme provided by the extension database
# (Validation must happen immediately)
if Setup.extension_db.has_key(Setup.output_file_naming_scheme) == False:
error.log("File extension scheme '%s' is not provided for language '%s'.\n" \
% (Setup.output_file_naming_scheme, Setup.language) + \
"Available schemes are: %s." % repr(Setup.extension_db.keys())[1:-1])
if Setup.buffer_byte_order == "<system>":
Setup.buffer_byte_order = sys.byteorder
Setup.byte_order_is_that_of_current_system_f = True
else:
Setup.byte_order_is_that_of_current_system_f = False
Setup.buffer_element_specification_prepare()
if Setup.buffer_codec_name == "utf8": module = utf8_state_split
elif Setup.buffer_codec_name == "utf16": module = utf16_state_split
else: module = None
Setup.buffer_codec_prepare(Setup.buffer_codec_name,
Setup.buffer_codec_file, module)
# AFTER: Setup.buffer_codec_prepare() !!!
if Setup.language not in ["DOT"]:
prepare_file_names(Setup)
type_info = global_character_type_db.get(Setup.buffer_element_type)
if type_info is not None and len(type_info) >= 4 \
and type_info[3] != -1 and Setup.buffer_element_size != -1 \
and type_info[3] != Setup.buffer_element_size:
error.log("\nBuffer element type ('--bet' or '--buffer-element-type') was set to '%s'.\n" \
% Setup.buffer_element_type \
+ "It is well known to be of size %s[byte]. However, the buffer element size\n" \
% type_info[3] \
+ "('-b' or '--buffer-element-type') was specified as '%s'.\n\n" \
% Setup.buffer_element_size \
+ "Quex can continue, but the result is questionable.\n", \
DontExitF=True)
Setup.converter_f = False
if Setup.converter_iconv_f or Setup.converter_icu_f or len(Setup.converter_user_new_func) != 0:
Setup.converter_f = True
# The only case where no converter helper is required is where ASCII
# (Unicode restricted to [0, FF] is used.
Setup.converter_helper_required_f = True
if Setup.converter_f == False and Setup.buffer_element_size == 1 and Setup.buffer_codec.name == "unicode":
Setup.converter_helper_required_f = False
validation.do(Setup, command_line, argv)
if Setup.converter_ucs_coding_name == "":
if global_character_type_db.has_key(Setup.buffer_element_type):
if Setup.buffer_byte_order == "little": index = 1
else: index = 2
Setup.converter_ucs_coding_name = global_character_type_db[Setup.buffer_element_type][index]
if len(Setup.token_id_foreign_definition) != 0:
if len(Setup.token_id_foreign_definition) > 3:
error.log("Option '--foreign-token-id-file' received > 3 followers.\n"
"Found: %s" % str(Setup.token_id_foreign_definition)[1:-1])
if len(Setup.token_id_foreign_definition) > 1:
Setup.token_id_foreign_definition_file_region_begin_re = \
__compile_regular_expression(Setup.token_id_foreign_definition[1], "token id region begin")
if len(Setup.token_id_foreign_definition) > 2:
Setup.token_id_foreign_definition_file_region_end_re = \
__compile_regular_expression(Setup.token_id_foreign_definition[2], "token id region end")
Setup.token_id_foreign_definition_file = \
Setup.token_id_foreign_definition[0]
CommentDelimiterList = [["//", "\n"], ["/*", "*/"]]
token_id_file_parse(Setup.token_id_foreign_definition_file,
CommentDelimiterList)
# (*) Compression Types
compression_type_list = []
for name, ctype in [("compression_template_f", E_Compression.TEMPLATE),
("compression_template_uniform_f", E_Compression.TEMPLATE_UNIFORM),
("compression_path_f", E_Compression.PATH),
("compression_path_uniform_f", E_Compression.PATH_UNIFORM)]:
if command_line_args_defined(command_line, name):
compression_type_list.append((command_line_arg_position(name), ctype))
compression_type_list.sort(key=itemgetter(0))
Setup.compression_type_list = map(lambda x: x[1], compression_type_list)
# (*) return Setup ___________________________________________________________________
return True
def do(setup, command_line, argv):
"""Does a consistency check for setup and the command line.
"""
setup.output_directory = os.path.normpath(setup.output_directory)
if setup.output_directory:
# Check, if the output directory exists
if os.access(setup.output_directory, os.F_OK) == False:
error.log("The directory %s was specified for output, but does not exists." % setup.output_directory)
if os.access(setup.output_directory, os.W_OK) == False:
error.log("The directory %s was specified for output, but is not writeable." % setup.output_directory)
# if the mode is '--language dot' => check character display options.
if setup.character_display not in ["hex", "utf8"]:
error.log("Character display must be either 'hex' or 'utf8'.\nFound: '%s'" %
setup.character_display)
# ensure that options are not specified twice
for parameter, info in SETUP_INFO.items():
if type(info) != list: continue
occurence_n = 0
for option in info[0]:
occurence_n += argv.count(option)
if occurence_n > 1 and info[1] not in (SetupParTypes.LIST, SetupParTypes.INT_LIST):
error.log("Received more than one of the following options:\n" + \
"%s" % repr(info[0])[1:-1])
# (*) Check for 'Depraceted' Options ___________________________________________________
for name, info in DEPRECATED.items():
command_line_options = SETUP_INFO[name][0]
comment = info[0]
depreciated_since_version = info[1]
for option in command_line_options:
if command_line.search(option):
error.log("Command line option '%s' is ignored.\n" % option + \
comment + "\n" + \
"Last version of Quex supporting this option is version %s. Please, visit\n" % \
depreciated_since_version + \
"http://quex.org for further information.")
# (*) Check for 'Straying' Options ___________________________________________________
options = []
for key, info in SETUP_INFO.items():
if type(info) != list: continue
if key in DEPRECATED: continue
if info[1] is not None: options.extend(info[0])
options.sort(lambda a,b: cmp(a.replace("-",""), b.replace("-","")))
ufos = command_line.unidentified_options(options)
if len(ufos) != 0:
error.log("Unidentified option(s) = " + repr(ufos) + "\n" + \
__get_supported_command_line_option_description(options))
if setup.analyzer_derived_class_name != "" and \
setup.analyzer_derived_class_file == "":
error.log("Specified derived class '%s' on command line, but it was not\n" % \
setup.analyzer_derived_class_name + \
"specified which file contains the definition of it.\n" + \
"use command line option '--derived-class-file'.\n")
if setup.buffer_element_size not in [-1, 1, 2, 4]:
error.log("The setting of '--buffer-element-size' (or '-b') can only be\n"
"1, 2, or 4 (found %s)." % repr(setup.buffer_element_size))
if setup.buffer_byte_order not in ["<system>", "little", "big"]:
error.log("Byte order (option --endian) must be 'little', 'big', or '<system>'.\n" + \
"Note, that this option is only interesting for cross plattform development.\n" + \
"By default, quex automatically chooses the endian type of your system.")
# Manually written token class requires token class name to be specified
if setup.token_class_file != "" and command_line.search("--token-class", "--tc") == False:
error.log("The use of a manually written token class requires that the name of the class\n"
"is specified on the command line via the '--token-class' option.")
# Token queue
if setup.token_policy != "queue" and command_line.search("--token-queue-size"):
error.log("Option --token-queue-size determines a fixed token queue size. This makes\n" + \
"only sense in conjunction with '--token-policy queue'.\n")
if setup.token_queue_size <= setup.token_queue_safety_border + 1:
if setup.token_queue_size == setup.token_queue_safety_border: cmp_str = "equal to"
else: cmp_str = "less than"
error.log("Token queue size is %i is %s token queue safety border %i + 1.\n" % \
(setup.token_queue_size, cmp_str, setup.token_queue_safety_border) +
"Set appropriate values with --token-queue-size and --token-queue-safety-border.")
# Check that names are valid identifiers
if len(setup.token_id_prefix_plain) != 0:
__check_identifier(setup, "token_id_prefix_plain", "Token prefix")
__check_identifier(setup, "analyzer_class_name", "Engine name")
if setup.analyzer_derived_class_name != "":
__check_identifier(setup, "analyzer_derived_class_name", "Derived class name")
__check_file_name(setup, "token_class_file", "file containing token class definition")
__check_file_name(setup, "analyzer_derived_class_file", "file containing user derived lexer class")
__check_file_name(setup, "token_id_foreign_definition_file", "file containing user token ids", 0,
CommandLineOption=SETUP_INFO["token_id_foreign_definition"][0])
__check_file_name(setup, "input_mode_files", "quex source file")
# Check that not more than one converter is specified
converter_n = 0
if setup.converter_iconv_f: converter_n += 1
if setup.converter_icu_f: converter_n += 1
if len(setup.converter_user_new_func) != 0: converter_n += 1
if converter_n > 1:
error.log("More than one character converter has been specified. Note, that the\n" + \
"options '--icu', '--iconv', and '--converter-new' (or '--cn') are\n" + \
"to be used mutually exclusively.")
if converter_n == 1 and setup.buffer_codec.name != "unicode":
# If the buffer codec is other than unicode, then no converter shall
# be used to fill the buffer. Instead, the engine is transformed, so
# that it works directly on the codec.
error.log("An engine that is to be generated for a specific codec cannot rely\n" + \
"on converters. Do no use '--codec' together with '--icu', '--iconv', or\n" + \
"`--converter-new`.")
# If a converter has been specified and no bytes-element-size has been specified,
# it defaults to '1 byte' which is most likely not what is desired for unicode.
if converter_n == 1 \
and setup.buffer_element_size == 1 \
and not command_line_args_defined(command_line, "buffer_element_size") \
and not command_line_args_defined(command_line, "buffer_element_type"):
error.log("A converter has been specified, but the default buffer element size\n" + \
"is left to 1 byte. Consider %s or %s." \
% (command_line_args_string("buffer_element_size"),
command_line_args_string("buffer_element_type")))
# If a user defined type is specified for 'engine character type' and
# a converter, then the name of the target type must be specified explicitly.
if setup.buffer_element_type != "" \
and not global_character_type_db.has_key(setup.buffer_element_type) \
and setup.converter_ucs_coding_name == "" \
and converter_n != 0:
tc = setup.buffer_element_type
error.log("A character code converter has been specified. It is supposed to convert\n" + \
"incoming data into an internal buffer of unicode characters. The size of\n" + \
"each character is determined by '%s' which is a user defined type.\n" % tc + \
"\n" + \
"Quex cannot determine automatically the name that the converter requires\n" + \
"to produce unicode characters for type '%s'. It must be specified by the\n" % tc + \
"command line option %s." \
% command_line_args_string("converter_ucs_coding_name"))
# Token transmission policy
token_policy_list = ["queue", "single", "users_token", "users_queue"]
if setup.token_policy not in token_policy_list:
error.log("Token policy '%s' not supported. Use one of the following:\n" % setup.token_policy + \
repr(token_policy_list)[1:-1])
elif setup.token_policy == "users_token":
error.log("Token policy 'users_queue' has be deprecated since 0.49.1. Use\n"
"equivalent policy 'single'.")
elif setup.token_policy == "users_queue":
error.log("Token policy 'users_queue' has be deprecated since 0.49.1\n")
# Internal engine character encoding
def __codec_vs_buffer_element_size(CodecName, RequiredBufferElementSize):
if setup.buffer_codec.name != CodecName: return
elif setup.buffer_element_size == RequiredBufferElementSize: return
if setup.buffer_element_size == -1:
msg_str = "undetermined (found type '%s')" % setup.buffer_element_type
else:
msg_str = "is not %i (found %i)" % (RequiredBufferElementSize, setup.buffer_element_size)
error.log("Using codec '%s' while buffer element size %s.\n" % (CodecName, msg_str) +
"Consult command line argument %s" \
% command_line_args_string("buffer_element_size"))
if setup.buffer_codec.name != "unicode":
if not setup.buffer_codec_file:
error.verify_word_in_list(setup.buffer_codec_name,
codec_db.get_supported_codec_list() + ["utf8", "utf16"],
"Codec '%s' is not supported." % setup.buffer_codec.name)
__codec_vs_buffer_element_size("utf8", 1)
__codec_vs_buffer_element_size("utf16", 2)
if setup.external_lexeme_null_object and setup.token_class_only_f:
error.log("Specifying an external lexeme null object signalizes an\n"
"external token class implementation. The 'token class only\n"
"flag' generates a token class considered to be externally\n"
"shared. Both flags are mutually exclusive.")
if setup.string_accumulator_f:
error_n = NotificationDB.warning_on_no_token_class_take_text
if error_n in setup.suppressed_notification_list:
error.warning("The warning upon missing 'take_text' in token type definition is de-\n"
+ "activated by '--suppress %i'. This is dangerous, if there is a string\n" % error_n
+ "accumulator. May be, use '--no-string-accumulator'.", -1,
SuppressCode=NotificationDB.warning_on_no_warning_on_missing_take_text)
def do(setup, command_line, argv):
"""Does a consistency check for setup and the command line.
"""
setup.output_directory = os.path.normpath(setup.output_directory)
if setup.output_directory != "":
# Check, if the output directory exists
if os.access(setup.output_directory, os.F_OK) == False:
error_msg(
"The directory %s was specified for output, but does not exists."
% setup.output_directory)
if os.access(setup.output_directory, os.W_OK) == False:
error_msg(
"The directory %s was specified for output, but is not writeable."
% setup.output_directory)
# if the mode is 'plotting', then check wether a graphic format is speicified
for plot_option in SETUP_INFO["plot_graphic_format"][0]:
if plot_option in argv and setup.plot_graphic_format == "":
error_msg("Option '%s' must be followed by a graphic format specifier (bmp, svg, jpg, ...)" % \
plot_option)
if setup.plot_character_display not in ["hex", "utf8"]:
error_msg(
"Plot character display must be either 'hex' or 'utf8'.\nFound: '%s'"
% setup.plot_character_display)
# ensure that options are not specified twice
for parameter, info in SETUP_INFO.items():
if type(info) != list: continue
occurence_n = 0
for option in info[0]:
occurence_n += argv.count(option)
if occurence_n > 1:
error_msg("Received more than one of the following options:\n" + \
"%s" % repr(info[0])[1:-1])
# (*) Check for 'Depraceted' Options ___________________________________________________
for name, info in DEPRECATED.items():
command_line_options = SETUP_INFO[name][0]
comment = info[0]
depreciated_since_version = info[1]
for option in command_line_options:
if command_line.search(option):
error_msg("Command line option '%s' is ignored.\n" % option + \
comment + "\n" + \
"Last version of Quex supporting this option is version %s. Please, visit\n" % \
depreciated_since_version + \
"http://quex.org for further information.")
# (*) Check for 'Straying' Options ___________________________________________________
options = []
for key, info in SETUP_INFO.items():
if type(info) != list: continue
if key in DEPRECATED: continue
if info[1] != None: options.extend(info[0])
options.sort(lambda a, b: cmp(a.replace("-", ""), b.replace("-", "")))
ufos = command_line.unidentified_options(options)
if ufos != []:
error_msg("Unidentified option(s) = " + repr(ufos) + "\n" + \
__get_supported_command_line_option_description(options))
if setup.analyzer_derived_class_name != "" and \
setup.analyzer_derived_class_file == "":
error_msg("Specified derived class '%s' on command line, but it was not\n" % \
setup.analyzer_derived_class_name + \
"specified which file contains the definition of it.\n" + \
"use command line option '--derived-class-file'.\n")
if setup.buffer_element_size not in [-1, 1, 2, 4]:
error_msg(
"The setting of '--buffer-element-size' (or '-b') can only be\n"
"1, 2, or 4 (found %s)." % repr(setup.buffer_element_size))
if setup.buffer_byte_order not in ["<system>", "little", "big"]:
error_msg("Byte order (option --endian) must be 'little', 'big', or '<system>'.\n" + \
"Note, that this option is only interesting for cross plattform development.\n" + \
"By default, quex automatically chooses the endian type of your system.")
# Manually written token class requires token class name to be specified
if setup.token_class_file != "" and command_line.search(
"--token-class", "--tc") == False:
error_msg(
"The use of a manually written token class requires that the name of the class\n"
"is specified on the command line via the '--token-class' option.")
# Token queue
if setup.token_policy != "queue" and command_line.search(
"--token-queue-size"):
error_msg("Option --token-queue-size determines a fixed token queue size. This makes\n" + \
"only sense in conjunction with '--token-policy queue'.\n")
if setup.token_queue_size <= setup.token_queue_safety_border + 1:
if setup.token_queue_size == setup.token_queue_safety_border:
cmp_str = "equal to"
else:
cmp_str = "less than"
error_msg("Token queue size is %i is %s token queue safety border %i + 1.\n" % \
(setup.token_queue_size, cmp_str, setup.token_queue_safety_border) +
"Set appropriate values with --token-queue-size and --token-queue-safety-border.")
# Check that names are valid identifiers
__check_identifier(setup, "token_id_prefix_plain", "Token prefix")
__check_identifier(setup, "analyzer_class_name", "Engine name")
if setup.analyzer_derived_class_name != "":
__check_identifier(setup, "analyzer_derived_class_name",
"Derived class name")
__check_file_name(setup, "token_class_file",
"file containing token class definition")
__check_file_name(setup, "analyzer_derived_class_file",
"file containing user derived lexer class")
__check_file_name(setup, "token_id_foreign_definition_file",
"file containing user token ids")
__check_file_name(setup, "input_mode_files", "quex source file")
# Check that not more than one converter is specified
converter_n = 0
if setup.converter_iconv_f: converter_n += 1
if setup.converter_icu_f: converter_n += 1
if setup.converter_user_new_func != "": converter_n += 1
if converter_n > 1:
error_msg("More than one character converter has been specified. Note, that the\n" + \
"options '--icu', '--iconv', and '--converter-new' (or '--cn') are\n" + \
"to be used mutually exclusively.")
if converter_n == 1 and setup.buffer_codec != "":
error_msg("An engine that is to be generated for a specific codec cannot rely\n" + \
"on converters. Do no use '--codec' together with '--icu', '--iconv', or\n" + \
"`--converter-new`.")
# If a user defined type is specified for 'engine character type' and
# a converter, then the name of the target type must be specified explicitly.
if setup.buffer_element_type != "" \
and not global_character_type_db.has_key(setup.buffer_element_type) \
and setup.converter_ucs_coding_name == "" \
and converter_n != 0:
tc = setup.buffer_element_type
error_msg("A character code converter has been specified. It is supposed to convert\n" + \
"incoming data into an internal buffer of unicode characters. The size of\n" + \
"each character is determined by '%s' which is a user defined type.\n" % tc + \
"\n" + \
"Quex cannot determine automatically the name that the converter requires\n" + \
"to produce unicode characters for type '%s'. It must be specified by the\n" % tc + \
"command line option '--converter-ucs-coding-name' or '--cucn'.")
# Token transmission policy
token_policy_list = ["queue", "single", "users_token", "users_queue"]
if setup.token_policy not in token_policy_list:
error_msg("Token policy '%s' not supported. Use one of the following:\n" % setup.token_policy + \
repr(token_policy_list)[1:-1])
elif setup.token_policy == "users_token":
error_msg(
"Token policy 'users_queue' has be deprecated since 0.49.1. Use\n"
"equivalent policy 'single'.")
elif setup.token_policy == "users_queue":
error_msg(
"Token policy 'users_queue' has be deprecated since 0.49.1\n")
# Internal engine character encoding
def __codec_vs_buffer_element_size(CodecName, RequiredBufferElementSize):
if setup.buffer_codec != CodecName: return
elif setup.buffer_element_size == RequiredBufferElementSize: return
if setup.buffer_element_size == -1:
msg_str = "undetermined (found type '%s')" % setup.buffer_element_type
else:
msg_str = "is not %i (found %i)" % (RequiredBufferElementSize,
setup.buffer_element_size)
error_msg("Using codec '%s' while buffer element size %s.\n" %
(CodecName, msg_str) +
"Consult command line argument '--buffer-element-size'.")
if setup.buffer_codec != "":
verify_word_in_list(
setup.buffer_codec,
codec_db.get_supported_codec_list() + ["utf8", "utf16"],
"Codec '%s' is not supported." % setup.buffer_codec)
__codec_vs_buffer_element_size("utf8", 1)
__codec_vs_buffer_element_size("utf16", 2)
# Path Compression
if setup.compression_path_uniform_f and setup.compression_path_f:
error_msg(
"Both flags for path compression were set: '--path-compression' and\n"
"'--path-compression-uniform'. Please, choose only one!")
def do(argv):
"""RETURN: True, if process needs to be started.
False, if job is done.
"""
global setup
# (*) Interpret Command Line (A) _____________________________________________________
command_line = GetPot(argv)
if command_line.search("--version", "-v"):
print "Quex - Fast Universal Lexical Analyzer Generator"
print "Version " + QUEX_VERSION
print "(C) 2006-2010 Frank-Rene Schaefer"
print "ABSOLUTELY NO WARRANTY"
return False
if command_line.search("--help", "-h"):
print "Quex - Fast Universal Lexical Analyzer Generator"
print "Please, consult the quex documentation for further help, or"
print "visit http://quex.org"
print "(C) 2006-2010 Frank-Rene Schaefer"
print "ABSOLUTELY NO WARRANTY"
return False
if command_line.search("--plot-format-list"):
print quex.output.graphviz.interface.get_supported_graphic_format_description()
return False
for variable_name, info in SETUP_INFO.items():
# Some parameters are not set on the command line. Their entry is not associated
# with a description list.
if type(info) != list: continue
if info[1] == FLAG:
setup.__dict__[variable_name] = command_line.search(info[0])
elif info[1] == NEGATED_FLAG:
setup.__dict__[variable_name] = not command_line.search(info[0])
elif info[1] == LIST:
if not command_line.search(info[0]):
setup.__dict__[variable_name] = []
else:
the_list = command_line.nominus_followers(info[0])
if the_list == []:
error_msg("Option %s\nnot followed by anything." % repr(info[0])[1:-1])
if setup.__dict__.has_key(variable_name):
setup.__dict__[variable_name].extend(the_list)
else:
setup.__dict__[variable_name] = the_list
elif command_line.search(info[0]):
if not command_line.search(info[0]):
setup.__dict__[variable_name] = info[1]
else:
value = command_line.follow("--EMPTY--", info[0])
if value == "--EMPTY--":
error_msg("Option %s\nnot followed by anything." % repr(info[0])[1:-1])
setup.__dict__[variable_name] = value
# (*) Classes and their namespace
setup.analyzer_class_name, \
setup.analyzer_name_space, \
setup.analyzer_name_safe = \
read_namespaced_name(setup.analyzer_class_name,
"analyzer engine (options -o, --engine, --analyzer-class)")
setup.analyzer_derived_class_name, \
setup.analyzer_derived_class_name_space, \
setup.analyzer_derived_class_name_safe = \
read_namespaced_name(setup.analyzer_derived_class_name,
"derived analyzer class (options --derived-class, --dc)",
AllowEmptyF=True)
if setup.analyzer_name_space == []:
setup.analyzer_name_space = ["quex"]
if setup.token_class_name == "":
setup.token_class_name = "%s::Token" % reduce(lambda a, b: a + "::" + b, setup.analyzer_name_space)
# Token classes and derived classes have the freedom not to open a namespace,
# thus no check 'if namespace == []'.
setup.token_class_name, \
setup.token_class_name_space, \
setup.token_class_name_safe = \
read_namespaced_name(setup.token_class_name,
"token class (options --token-class, --tc)")
if setup.token_class_file != "":
lexer_mode.token_type_definition = \
ManualTokenClassSetup(setup.token_class_file,
setup.token_class_name,
setup.token_class_name_space,
setup.token_class_name_safe,
setup.token_id_type)
if setup.token_class_name_space == []:
setup.token_class_name_space = deepcopy(setup.analyzer_name_space)
setup.token_id_prefix_plain, \
setup.token_id_prefix_name_space, \
dummy = \
read_namespaced_name(setup.token_id_prefix,
"token prefix (options --token-prefix)")
if len(setup.token_id_prefix_name_space) != 0 and setup.language.upper() == "C":
error_msg("Token id prefix cannot contain a namespaces if '--language' is set to 'C'.")
# (*) Output programming language
setup.language = setup.language.upper()
verify_word_in_list(setup.language,
quex_core_engine_generator_languages_db.keys(),
"Programming language '%s' is not supported." % setup.language)
setup.language_db = quex_core_engine_generator_languages_db[setup.language]
setup.extension_db = global_extension_db[setup.language]
# Is the output file naming scheme provided by the extension database
# (Validation must happen immediately)
if setup.extension_db.has_key(setup.output_file_naming_scheme) == False:
error_msg("File extension scheme '%s' is not provided for language '%s'.\n" \
% (setup.output_file_naming_scheme, setup.language) + \
"Available schemes are: %s." % repr(setup.extension_db.keys())[1:-1])
# Before file names can be prepared, determine the output directory
# If 'source packaging' is enabled and no output directory is specified
# then take the directory of the source packaging.
if setup.source_package_directory != "" and setup.output_directory == "":
setup.output_directory = setup.source_package_directory
# (*) Output files
prepare_file_names(setup)
if setup.buffer_byte_order == "<system>":
setup.buffer_byte_order = sys.byteorder
setup.byte_order_is_that_of_current_system_f = True
else:
setup.byte_order_is_that_of_current_system_f = False
if setup.buffer_element_size == "wchar_t":
error_msg("Since Quex version 0.53.5, 'wchar_t' can no longer be specified\n"
"with option '--buffer-element-size' or '-bes'. Please, specify\n"
"'--buffer-element-type wchar_t' or '--bet'.")
if setup.buffer_element_type == "wchar_t":
setup.converter_ucs_coding_name = "WCHAR_T"
if setup.buffer_codec != "":
setup.buffer_element_size_irrelevant = True
make_numbers(setup)
# (*) Determine buffer element type and size (in bytes)
if setup.buffer_element_size == -1:
if global_character_type_db.has_key(setup.buffer_element_type):
setup.buffer_element_size = global_character_type_db[setup.buffer_element_type][3]
elif setup.buffer_element_type == "":
setup.buffer_element_size = 1
else:
# If the buffer element type is defined, then here we know that it is 'unknown'
# and Quex cannot know its size on its own.
setup.buffer_element_size = -1
if setup.buffer_element_type == "":
if setup.buffer_element_size in [1, 2, 4]:
setup.buffer_element_type = {
1: "uint8_t", 2: "uint16_t", 4: "uint32_t",
}[setup.buffer_element_size]
elif setup.buffer_element_size == -1:
pass
else:
error_msg("Buffer element type cannot be determined for size '%i' which\n" \
% setup.buffer_element_size +
"has been specified by '-b' or '--buffer-element-size'.")
if setup.buffer_codec in ["utf8", "utf16"]:
setup.buffer_codec_transformation_info = setup.buffer_codec + "-state-split"
elif setup.buffer_codec != "":
setup.buffer_codec_transformation_info = codec_db.get_codec_transformation_info(setup.buffer_codec)
setup.converter_f = False
if setup.converter_iconv_f or setup.converter_icu_f:
setup.converter_f = True
# The only case where no converter helper is required is where ASCII
# (Unicode restricted to [0, FF] is used.
setup.converter_helper_required_f = True
if setup.converter_f == False and setup.buffer_element_size == 1 and setup.buffer_codec == "":
setup.converter_helper_required_f = False
validation.do(setup, command_line, argv)
if setup.converter_ucs_coding_name == "":
if global_character_type_db.has_key(setup.buffer_element_type):
if setup.buffer_byte_order == "little": index = 1
else: index = 2
setup.converter_ucs_coding_name = global_character_type_db[setup.buffer_element_type][index]
if setup.token_id_foreign_definition_file != "":
CommentDelimiterList = [["//", "\n"], ["/*", "*/"]]
# Regular expression to find '#include <something>' and extract the 'something'
# in a 'group'. Note that '(' ')' cause the storage of parts of the match.
IncludeRE = "#[ \t]*include[ \t]*[\"<]([^\">]+)[\">]"
#
parse_token_id_file(setup.token_id_foreign_definition_file,
setup.token_id_prefix,
CommentDelimiterList, IncludeRE)
if setup.token_id_prefix_plain != setup.token_id_prefix:
# The 'plain' name space less token indices are also supported
parse_token_id_file(setup.token_id_foreign_definition_file,
setup.token_id_prefix_plain,
CommentDelimiterList, IncludeRE)
# (*) return setup ___________________________________________________________________
return True
def prepare(command_line, argv):
"""RETURN: True, if process needs to be started.
False, if job is done.
"""
global Setup
# (*) Classes and their namespace
__setup_analyzer_class(Setup)
__setup_token_class(Setup)
__setup_token_id_prefix(Setup)
__setup_lexeme_null(Setup) # Requires 'token_class_name_space'
# (*) Output programming language
Setup.language = Setup.language.upper()
error.verify_word_in_list(
Setup.language, output_language_db.keys(),
"Programming language '%s' is not supported." % Setup.language)
Setup.language_db = output_language_db[Setup.language]
Setup.extension_db = global_extension_db[Setup.language]
# Is the output file naming scheme provided by the extension database
# (Validation must happen immediately)
if Setup.extension_db.has_key(Setup.output_file_naming_scheme) == False:
error.log("File extension scheme '%s' is not provided for language '%s'.\n" \
% (Setup.output_file_naming_scheme, Setup.language) + \
"Available schemes are: %s." % repr(Setup.extension_db.keys())[1:-1])
if Setup.buffer_byte_order == "<system>":
Setup.buffer_byte_order = sys.byteorder
Setup.byte_order_is_that_of_current_system_f = True
else:
Setup.byte_order_is_that_of_current_system_f = False
Setup.buffer_element_specification_prepare()
if Setup.buffer_codec_name == "utf8": module = utf8_state_split
elif Setup.buffer_codec_name == "utf16": module = utf16_state_split
else: module = None
Setup.buffer_codec_prepare(Setup.buffer_codec_name,
Setup.buffer_codec_file, module)
# AFTER: Setup.buffer_codec_prepare() !!!
if Setup.language not in ["DOT"]:
prepare_file_names(Setup)
type_info = global_character_type_db.get(Setup.buffer_element_type)
if type_info is not None and len(type_info) >= 4 \
and type_info[3] != -1 and Setup.buffer_element_size != -1 \
and type_info[3] != Setup.buffer_element_size:
error.log("\nBuffer element type ('--bet' or '--buffer-element-type') was set to '%s'.\n" \
% Setup.buffer_element_type \
+ "It is well known to be of size %s[byte]. However, the buffer element size\n" \
% type_info[3] \
+ "('-b' or '--buffer-element-type') was specified as '%s'.\n\n" \
% Setup.buffer_element_size \
+ "Quex can continue, but the result is questionable.\n", \
DontExitF=True)
Setup.converter_f = False
if Setup.converter_iconv_f or Setup.converter_icu_f or len(
Setup.converter_user_new_func) != 0:
Setup.converter_f = True
# The only case where no converter helper is required is where ASCII
# (Unicode restricted to [0, FF] is used.
Setup.converter_helper_required_f = True
if Setup.converter_f == False and Setup.buffer_element_size == 1 and Setup.buffer_codec.name == "unicode":
Setup.converter_helper_required_f = False
validation.do(Setup, command_line, argv)
if Setup.converter_ucs_coding_name == "":
if global_character_type_db.has_key(Setup.buffer_element_type):
if Setup.buffer_byte_order == "little": index = 1
else: index = 2
Setup.converter_ucs_coding_name = global_character_type_db[
Setup.buffer_element_type][index]
if len(Setup.token_id_foreign_definition) != 0:
if len(Setup.token_id_foreign_definition) > 3:
error.log(
"Option '--foreign-token-id-file' received > 3 followers.\n"
"Found: %s" % str(Setup.token_id_foreign_definition)[1:-1])
if len(Setup.token_id_foreign_definition) > 1:
Setup.token_id_foreign_definition_file_region_begin_re = \
__compile_regular_expression(Setup.token_id_foreign_definition[1], "token id region begin")
if len(Setup.token_id_foreign_definition) > 2:
Setup.token_id_foreign_definition_file_region_end_re = \
__compile_regular_expression(Setup.token_id_foreign_definition[2], "token id region end")
Setup.token_id_foreign_definition_file = \
Setup.token_id_foreign_definition[0]
CommentDelimiterList = [["//", "\n"], ["/*", "*/"]]
token_id_file_parse(Setup.token_id_foreign_definition_file,
CommentDelimiterList)
# (*) Compression Types
compression_type_list = []
for name, ctype in [
("compression_template_f", E_Compression.TEMPLATE),
("compression_template_uniform_f", E_Compression.TEMPLATE_UNIFORM),
("compression_path_f", E_Compression.PATH),
("compression_path_uniform_f", E_Compression.PATH_UNIFORM)
]:
if command_line_args_defined(command_line, name):
compression_type_list.append(
(command_line_arg_position(name), ctype))
compression_type_list.sort(key=itemgetter(0))
Setup.compression_type_list = map(lambda x: x[1], compression_type_list)
# (*) return Setup ___________________________________________________________________
return True
def do(argv):
"""RETURN: True, if process needs to be started.
False, if job is done.
"""
global setup
# (*) Interpret Command Line (A) _____________________________________________________
command_line = GetPot(argv)
if command_line.search("--version", "-v"):
print "Quex - Fast Universal Lexical Analyzer Generator"
print "Version " + QUEX_VERSION
print "(C) 2006-2010 Frank-Rene Schaefer"
print "ABSOLUTELY NO WARRANTY"
return False
if command_line.search("--help", "-h"):
print "Quex - Fast Universal Lexical Analyzer Generator"
print "Please, consult the quex documentation for further help, or"
print "visit http://quex.org"
print "(C) 2006-2010 Frank-Rene Schaefer"
print "ABSOLUTELY NO WARRANTY"
return False
if command_line.search("--plot-format-list"):
print quex.output.graphviz.interface.get_supported_graphic_format_description(
)
return False
for variable_name, info in SETUP_INFO.items():
# Some parameters are not set on the command line. Their entry is not associated
# with a description list.
if type(info) != list: continue
if info[1] == FLAG:
setup.__dict__[variable_name] = command_line.search(info[0])
elif info[1] == NEGATED_FLAG:
setup.__dict__[variable_name] = not command_line.search(info[0])
elif info[1] == LIST:
if not command_line.search(info[0]):
setup.__dict__[variable_name] = []
else:
the_list = command_line.nominus_followers(info[0])
if the_list == []:
error_msg("Option %s\nnot followed by anything." %
repr(info[0])[1:-1])
if setup.__dict__.has_key(variable_name):
setup.__dict__[variable_name].extend(the_list)
else:
setup.__dict__[variable_name] = the_list
elif command_line.search(info[0]):
if not command_line.search(info[0]):
setup.__dict__[variable_name] = info[1]
else:
value = command_line.follow("--EMPTY--", info[0])
if value == "--EMPTY--":
error_msg("Option %s\nnot followed by anything." %
repr(info[0])[1:-1])
setup.__dict__[variable_name] = value
# (*) Classes and their namespace
setup.analyzer_class_name, \
setup.analyzer_name_space, \
setup.analyzer_name_safe = \
read_namespaced_name(setup.analyzer_class_name,
"analyzer engine (options -o, --engine, --analyzer-class)")
setup.analyzer_derived_class_name, \
setup.analyzer_derived_class_name_space, \
setup.analyzer_derived_class_name_safe = \
read_namespaced_name(setup.analyzer_derived_class_name,
"derived analyzer class (options --derived-class, --dc)",
AllowEmptyF=True)
if setup.analyzer_name_space == []:
setup.analyzer_name_space = ["quex"]
if setup.token_class_name == "":
setup.token_class_name = "%s::Token" % reduce(
lambda a, b: a + "::" + b, setup.analyzer_name_space)
# Token classes and derived classes have the freedom not to open a namespace,
# thus no check 'if namespace == []'.
setup.token_class_name, \
setup.token_class_name_space, \
setup.token_class_name_safe = \
read_namespaced_name(setup.token_class_name,
"token class (options --token-class, --tc)")
if setup.token_class_file != "":
lexer_mode.token_type_definition = \
ManualTokenClassSetup(setup.token_class_file,
setup.token_class_name,
setup.token_class_name_space,
setup.token_class_name_safe,
setup.token_id_type)
if setup.token_class_name_space == []:
setup.token_class_name_space = deepcopy(setup.analyzer_name_space)
setup.token_id_prefix_plain, \
setup.token_id_prefix_name_space, \
dummy = \
read_namespaced_name(setup.token_id_prefix,
"token prefix (options --token-prefix)")
if len(setup.token_id_prefix_name_space) != 0 and setup.language.upper(
) == "C":
error_msg(
"Token id prefix cannot contain a namespaces if '--language' is set to 'C'."
)
# (*) Output programming language
setup.language = setup.language.upper()
verify_word_in_list(
setup.language, quex_core_engine_generator_languages_db.keys(),
"Programming language '%s' is not supported." % setup.language)
setup.language_db = quex_core_engine_generator_languages_db[setup.language]
setup.extension_db = global_extension_db[setup.language]
# Is the output file naming scheme provided by the extension database
# (Validation must happen immediately)
if setup.extension_db.has_key(setup.output_file_naming_scheme) == False:
error_msg("File extension scheme '%s' is not provided for language '%s'.\n" \
% (setup.output_file_naming_scheme, setup.language) + \
"Available schemes are: %s." % repr(setup.extension_db.keys())[1:-1])
# Before file names can be prepared, determine the output directory
# If 'source packaging' is enabled and no output directory is specified
# then take the directory of the source packaging.
if setup.source_package_directory != "" and setup.output_directory == "":
setup.output_directory = setup.source_package_directory
# (*) Output files
prepare_file_names(setup)
if setup.buffer_byte_order == "<system>":
setup.buffer_byte_order = sys.byteorder
setup.byte_order_is_that_of_current_system_f = True
else:
setup.byte_order_is_that_of_current_system_f = False
if setup.buffer_element_size == "wchar_t":
error_msg(
"Since Quex version 0.53.5, 'wchar_t' can no longer be specified\n"
"with option '--buffer-element-size' or '-bes'. Please, specify\n"
"'--buffer-element-type wchar_t' or '--bet'.")
if setup.buffer_element_type == "wchar_t":
setup.converter_ucs_coding_name = "WCHAR_T"
if setup.buffer_codec != "":
setup.buffer_element_size_irrelevant = True
make_numbers(setup)
# (*) Determine buffer element type and size (in bytes)
if setup.buffer_element_size == -1:
if global_character_type_db.has_key(setup.buffer_element_type):
setup.buffer_element_size = global_character_type_db[
setup.buffer_element_type][3]
elif setup.buffer_element_type == "":
setup.buffer_element_size = 1
else:
# If the buffer element type is defined, then here we know that it is 'unknown'
# and Quex cannot know its size on its own.
setup.buffer_element_size = -1
if setup.buffer_element_type == "":
if setup.buffer_element_size in [1, 2, 4]:
setup.buffer_element_type = {
1: "uint8_t",
2: "uint16_t",
4: "uint32_t",
}[setup.buffer_element_size]
elif setup.buffer_element_size == -1:
pass
else:
error_msg("Buffer element type cannot be determined for size '%i' which\n" \
% setup.buffer_element_size +
"has been specified by '-b' or '--buffer-element-size'.")
if setup.buffer_codec in ["utf8", "utf16"]:
setup.buffer_codec_transformation_info = setup.buffer_codec + "-state-split"
elif setup.buffer_codec != "":
setup.buffer_codec_transformation_info = codec_db.get_codec_transformation_info(
setup.buffer_codec)
setup.converter_f = False
if setup.converter_iconv_f or setup.converter_icu_f:
setup.converter_f = True
# The only case where no converter helper is required is where ASCII
# (Unicode restricted to [0, FF] is used.
setup.converter_helper_required_f = True
if setup.converter_f == False and setup.buffer_element_size == 1 and setup.buffer_codec == "":
setup.converter_helper_required_f = False
validation.do(setup, command_line, argv)
if setup.converter_ucs_coding_name == "":
if global_character_type_db.has_key(setup.buffer_element_type):
if setup.buffer_byte_order == "little": index = 1
else: index = 2
setup.converter_ucs_coding_name = global_character_type_db[
setup.buffer_element_type][index]
if setup.token_id_foreign_definition_file != "":
CommentDelimiterList = [["//", "\n"], ["/*", "*/"]]
# Regular expression to find '#include <something>' and extract the 'something'
# in a 'group'. Note that '(' ')' cause the storage of parts of the match.
IncludeRE = "#[ \t]*include[ \t]*[\"<]([^\">]+)[\">]"
#
parse_token_id_file(setup.token_id_foreign_definition_file,
setup.token_id_prefix, CommentDelimiterList,
IncludeRE)
if setup.token_id_prefix_plain != setup.token_id_prefix:
# The 'plain' name space less token indices are also supported
parse_token_id_file(setup.token_id_foreign_definition_file,
setup.token_id_prefix_plain,
CommentDelimiterList, IncludeRE)
# (*) return setup ___________________________________________________________________
return True
def __perform_setup(command_line, argv):
"""RETURN: True, if process needs to be started.
False, if job is done.
"""
global setup
# (*) Classes and their namespace
__setup_analyzer_class(setup)
__setup_token_class(setup)
__setup_token_id_prefix(setup)
__setup_lexeme_null(setup) # Requires 'token_class_name_space'
# (*) Output programming language
setup.language = setup.language.upper()
verify_word_in_list(
setup.language, quex_core_engine_generator_languages_db.keys(),
"Programming language '%s' is not supported." % setup.language)
setup.language_db = quex_core_engine_generator_languages_db[setup.language]
setup.extension_db = global_extension_db[setup.language]
# Is the output file naming scheme provided by the extension database
# (Validation must happen immediately)
if setup.extension_db.has_key(setup.output_file_naming_scheme) == False:
error_msg("File extension scheme '%s' is not provided for language '%s'.\n" \
% (setup.output_file_naming_scheme, setup.language) + \
"Available schemes are: %s." % repr(setup.extension_db.keys())[1:-1])
# Before file names can be prepared, determine the output directory
# If 'source packaging' is enabled and no output directory is specified
# then take the directory of the source packaging.
if setup.source_package_directory != "" and setup.output_directory == "":
setup.output_directory = setup.source_package_directory
if setup.buffer_codec in ["utf8", "utf16"]:
setup.buffer_codec_transformation_info = setup.buffer_codec + "-state-split"
elif setup.buffer_codec_file != "":
try:
setup.buffer_codec = os.path.splitext(
os.path.basename(setup.buffer_codec_file))[0]
except:
error_msg("cannot interpret string following '--codec-file'")
setup.buffer_codec_transformation_info = codec_db.get_codec_transformation_info(
FileName=setup.buffer_codec_file)
elif setup.buffer_codec != "unicode":
setup.buffer_codec_transformation_info = codec_db.get_codec_transformation_info(
setup.buffer_codec)
if setup.buffer_codec != "unicode":
setup.buffer_element_size_irrelevant = True
# (*) Output files
if setup.language not in ["DOT"]:
prepare_file_names(setup)
if setup.buffer_byte_order == "<system>":
setup.buffer_byte_order = sys.byteorder
setup.byte_order_is_that_of_current_system_f = True
else:
setup.byte_order_is_that_of_current_system_f = False
if setup.buffer_element_size == "wchar_t":
error_msg(
"Since Quex version 0.53.5, 'wchar_t' can no longer be specified\n"
"with option '--buffer-element-size' or '-bes'. Please, specify\n"
"'--buffer-element-type wchar_t' or '--bet'.")
if setup.buffer_element_type == "wchar_t":
setup.converter_ucs_coding_name = "WCHAR_T"
make_numbers(setup)
# (*) Determine buffer element type and size (in bytes)
if setup.buffer_element_size == -1:
if global_character_type_db.has_key(setup.buffer_element_type):
setup.buffer_element_size = global_character_type_db[
setup.buffer_element_type][3]
elif setup.buffer_element_type == "":
setup.buffer_element_size = 1
else:
# If the buffer element type is defined, then here we know that it is 'unknown'
# and Quex cannot know its size on its own.
setup.buffer_element_size = -1
if setup.buffer_element_type == "":
if setup.buffer_element_size in [1, 2, 4]:
setup.buffer_element_type = {
1: "uint8_t",
2: "uint16_t",
4: "uint32_t",
}[setup.buffer_element_size]
elif setup.buffer_element_size == -1:
pass
else:
error_msg("Buffer element type cannot be determined for size '%i' which\n" \
% setup.buffer_element_size +
"has been specified by '-b' or '--buffer-element-size'.")
setup.converter_f = False
if setup.converter_iconv_f or setup.converter_icu_f:
setup.converter_f = True
# The only case where no converter helper is required is where ASCII
# (Unicode restricted to [0, FF] is used.
setup.converter_helper_required_f = True
if setup.converter_f == False and setup.buffer_element_size == 1 and setup.buffer_codec == "unicode":
setup.converter_helper_required_f = False
validation.do(setup, command_line, argv)
if setup.converter_ucs_coding_name == "":
if global_character_type_db.has_key(setup.buffer_element_type):
if setup.buffer_byte_order == "little": index = 1
else: index = 2
setup.converter_ucs_coding_name = global_character_type_db[
setup.buffer_element_type][index]
if setup.token_id_foreign_definition_file != "":
CommentDelimiterList = [["//", "\n"], ["/*", "*/"]]
# Regular expression to find '#include <something>' and extract the 'something'
# in a 'group'. Note that '(' ')' cause the storage of parts of the match.
IncludeRE = "#[ \t]*include[ \t]*[\"<]([^\">]+)[\">]"
#
parse_token_id_file(setup.token_id_foreign_definition_file,
setup.token_id_prefix, CommentDelimiterList,
IncludeRE)
if setup.token_id_prefix_plain != setup.token_id_prefix:
# The 'plain' name space less token indices are also supported
parse_token_id_file(setup.token_id_foreign_definition_file,
setup.token_id_prefix_plain,
CommentDelimiterList, IncludeRE)
# (*) Compression Types
compression_type_list = []
for name, ctype in [
("compression_template_f", E_Compression.TEMPLATE),
("compression_template_uniform_f", E_Compression.TEMPLATE_UNIFORM),
("compression_path_f", E_Compression.PATH),
("compression_path_uniform_f", E_Compression.PATH_UNIFORM)
]:
if command_line_args_defined(command_line, name):
compression_type_list.append(
(command_line_arg_position(name), ctype))
compression_type_list.sort(key=itemgetter(0))
setup.compression_type_list = map(lambda x: x[1], compression_type_list)
# (*) return setup ___________________________________________________________________
return True
def __perform_setup(command_line, argv):
"""RETURN: True, if process needs to be started.
False, if job is done.
"""
global setup
# (*) Classes and their namespace
__setup_analyzer_class(setup)
__setup_token_class(setup)
__setup_token_id_prefix(setup)
__setup_lexeme_null(setup) # Requires 'token_class_name_space'
# (*) Output programming language
setup.language = setup.language.upper()
verify_word_in_list(setup.language,
quex_core_engine_generator_languages_db.keys(),
"Programming language '%s' is not supported." % setup.language)
setup.language_db = quex_core_engine_generator_languages_db[setup.language]
setup.extension_db = global_extension_db[setup.language]
# Is the output file naming scheme provided by the extension database
# (Validation must happen immediately)
if setup.extension_db.has_key(setup.output_file_naming_scheme) == False:
error_msg("File extension scheme '%s' is not provided for language '%s'.\n" \
% (setup.output_file_naming_scheme, setup.language) + \
"Available schemes are: %s." % repr(setup.extension_db.keys())[1:-1])
# Before file names can be prepared, determine the output directory
# If 'source packaging' is enabled and no output directory is specified
# then take the directory of the source packaging.
if setup.source_package_directory != "" and setup.output_directory == "":
setup.output_directory = setup.source_package_directory
if setup.buffer_codec in ["utf8", "utf16"]:
setup.buffer_codec_transformation_info = setup.buffer_codec + "-state-split"
elif setup.buffer_codec_file != "":
try:
setup.buffer_codec = os.path.splitext(os.path.basename(setup.buffer_codec_file))[0]
except:
error_msg("cannot interpret string following '--codec-file'")
setup.buffer_codec_transformation_info = codec_db.get_codec_transformation_info(FileName=setup.buffer_codec_file)
elif setup.buffer_codec != "unicode":
setup.buffer_codec_transformation_info = codec_db.get_codec_transformation_info(setup.buffer_codec)
if setup.buffer_codec != "unicode":
setup.buffer_element_size_irrelevant = True
# (*) Output files
if setup.language not in ["DOT"]:
prepare_file_names(setup)
if setup.buffer_byte_order == "<system>":
setup.buffer_byte_order = sys.byteorder
setup.byte_order_is_that_of_current_system_f = True
else:
setup.byte_order_is_that_of_current_system_f = False
if setup.buffer_element_size == "wchar_t":
error_msg("Since Quex version 0.53.5, 'wchar_t' can no longer be specified\n"
"with option '--buffer-element-size' or '-bes'. Please, specify\n"
"'--buffer-element-type wchar_t' or '--bet'.")
if setup.buffer_element_type == "wchar_t":
setup.converter_ucs_coding_name = "WCHAR_T"
make_numbers(setup)
# (*) Determine buffer element type and size (in bytes)
if setup.buffer_element_size == -1:
if global_character_type_db.has_key(setup.buffer_element_type):
setup.buffer_element_size = global_character_type_db[setup.buffer_element_type][3]
elif setup.buffer_element_type == "":
setup.buffer_element_size = 1
else:
# If the buffer element type is defined, then here we know that it is 'unknown'
# and Quex cannot know its size on its own.
setup.buffer_element_size = -1
if setup.buffer_element_type == "":
if setup.buffer_element_size in [1, 2, 4]:
setup.buffer_element_type = {
1: "uint8_t", 2: "uint16_t", 4: "uint32_t",
}[setup.buffer_element_size]
elif setup.buffer_element_size == -1:
pass
else:
error_msg("Buffer element type cannot be determined for size '%i' which\n" \
% setup.buffer_element_size +
"has been specified by '-b' or '--buffer-element-size'.")
setup.converter_f = False
if setup.converter_iconv_f or setup.converter_icu_f:
setup.converter_f = True
# The only case where no converter helper is required is where ASCII
# (Unicode restricted to [0, FF] is used.
setup.converter_helper_required_f = True
if setup.converter_f == False and setup.buffer_element_size == 1 and setup.buffer_codec == "unicode":
setup.converter_helper_required_f = False
validation.do(setup, command_line, argv)
if setup.converter_ucs_coding_name == "":
if global_character_type_db.has_key(setup.buffer_element_type):
if setup.buffer_byte_order == "little": index = 1
else: index = 2
setup.converter_ucs_coding_name = global_character_type_db[setup.buffer_element_type][index]
if setup.token_id_foreign_definition_file != "":
CommentDelimiterList = [["//", "\n"], ["/*", "*/"]]
# Regular expression to find '#include <something>' and extract the 'something'
# in a 'group'. Note that '(' ')' cause the storage of parts of the match.
IncludeRE = "#[ \t]*include[ \t]*[\"<]([^\">]+)[\">]"
#
parse_token_id_file(setup.token_id_foreign_definition_file,
setup.token_id_prefix,
CommentDelimiterList, IncludeRE)
if setup.token_id_prefix_plain != setup.token_id_prefix:
# The 'plain' name space less token indices are also supported
parse_token_id_file(setup.token_id_foreign_definition_file,
setup.token_id_prefix_plain,
CommentDelimiterList, IncludeRE)
# (*) Compression Types
compression_type_list = []
for name, ctype in [("compression_template_f", E_Compression.TEMPLATE),
("compression_template_uniform_f", E_Compression.TEMPLATE_UNIFORM),
("compression_path_f", E_Compression.PATH),
("compression_path_uniform_f", E_Compression.PATH_UNIFORM)]:
if command_line_args_defined(command_line, name):
compression_type_list.append((command_line_arg_position(name), ctype))
compression_type_list.sort(key=itemgetter(0))
setup.compression_type_list = map(lambda x: x[1], compression_type_list)
# (*) return setup ___________________________________________________________________
return True
