file modules/BOSS/modules/utils.py
[No description available]
Namespaces
| Name |
|---|
| modules |
| modules::utils |
Source code
################################
# #
# Utility functions for BOSS #
# #
################################
from __future__ import print_function
import xml.etree.ElementTree as ET
from collections import OrderedDict
from operator import itemgetter
import os
import copy
import modules.active_cfg as active_cfg
exec("import configs." + active_cfg.module_name + " as cfg")
import modules.gb as gb
import subprocess
import modules.exceptions as exceptions
import modules.infomsg as infomsg
import shlex
import tempfile
# ====== isComplete ========
def isComplete(class_el):
is_complete = True
# Check that class is complete (not only forward declared):
if ('incomplete' in class_el.keys()) and (class_el.get('incomplete') == '1'):
is_complete = False
return is_complete
# ====== END: isComplete ========
# ====== isLoadable ========
def isLoadable(class_el, print_warning=False, check_pure_virtual_members=True):
import modules.classutils as classutils
is_loadable = True
# - Any loadable class should have a "name" XML entry
if not 'name' in class_el.keys():
is_loadable = False
return is_loadable
class_name = classutils.getClassNameDict(class_el)
# - Check if class should be ditched. If yes, return right away.
if class_name['long_templ'] in cfg.ditch:
is_loadable = False
return is_loadable
# - Check if class is a template class. BOSS cannot handle this yet.
if isTemplateClass(class_el):
is_loadable = False
if print_warning:
reason = "This is a template class. BOSS cannot yet handle this."
infomsg.ClassNotLoadable(class_name['long_templ'], reason).printMessage()
return is_loadable
# - Check that class is complete (not only forward declared).
if not isComplete(class_el):
is_loadable = False
if print_warning:
reason = "Class is incomplete, at least based on XML file %s" % (gb.xml_file_name)
infomsg.ClassNotLoadable(class_name['long_templ'], reason).printMessage()
return is_loadable
# - Check that class has at least one public constructor.
constructor_elements = classutils.getAcceptableConstructors(class_el, skip_copy_constructors=True)
if len(constructor_elements) == 0:
is_loadable = False
if print_warning:
reason = "No (acceptable) public constructors identified."
infomsg.ClassNotLoadable(class_name['long_templ'], reason).printMessage()
return is_loadable
# - Check for pure virtual members.
if check_pure_virtual_members:
pure_virtual_members = classutils.pureVirtualMembers(class_el)
if len(pure_virtual_members) > 0:
gb.contains_pure_virtual_members.append(class_name['long_templ'])
return is_loadable
# ====== END: isLoadable ========
# ====== isFundamental ========
def isFundamental(el):
is_fundamental = False
if el.tag == 'FundamentalType':
is_fundamental = True
return is_fundamental
# ====== END: isFundamental ========
# ====== isKnownClass ========
def isKnownClass(el, class_name=None):
import modules.classutils as classutils
is_known = False
type_dict = findType(el)
type_el = type_dict['el']
# - Any known class should have a "name" XML entry
if not 'name' in type_el.keys():
is_known = False
return is_known
# Get class_name dict if it is not passed in as an argument
if class_name is None:
class_name = classutils.getClassNameDict(type_el)
# Check if standard library class
if isStdType(el, class_name=class_name):
is_known = True
return is_known
# Check if listed among the user-specified known types
if isInList(class_name['long_templ'], cfg.known_classes.keys(), return_index=False, ignore_whitespace=True):
is_known = True
elif isInList(class_name['long'], cfg.known_classes.keys(), return_index=False, ignore_whitespace=True):
is_known = True
return is_known
# ====== END: isKnownClass ========
# ====== isTemplateClass ========
def isTemplateClass(class_el):
import modules.classutils as classutils
is_template = False
class_name = classutils.getClassNameDict(class_el)
if '<' in class_name['long_templ']:
is_template = True
return is_template
# ====== END: isTemplateClass ========
# ====== isTemplateFunction ========
def isTemplateFunction(func_el):
import modules.funcutils as funcutils
is_template = False
func_name = funcutils.getFunctionNameDict(func_el)
if '<' in func_name['long_templ']:
is_template = True
return is_template
# ====== END: isTemplateFunction ========
# ====== isEnumeration ========
def isEnumeration(el):
is_enumeration = False
if el.tag == 'Enumeration':
is_enumeration = True
return is_enumeration
# ====== END: isEnumeration ========
# ====== isNative ========
def isNative(el):
# Makes use of global variables: base_paths
is_native = False
can_check_tags = ['Class', 'Constructor', 'Converter', 'Destructor', 'Enumeration',
'Field', 'File', 'Function', 'Method', 'OperatorFunction',
'OperatorMethod', 'Struct', 'Typedef', 'Union', 'Variable']
cannot_check_tags = ['Unimplemented']
if el.tag == 'FundamentalType':
is_native = False
elif el.tag in can_check_tags:
if el.tag == 'File':
file_el = el
else:
file_el = gb.id_dict[el.get('file')]
check_path = file_el.get('name')
is_native = False
for accepted_path in cfg.base_paths:
if accepted_path in os.path.dirname(check_path):
is_native = True
break
elif el.tag in cannot_check_tags:
pass
else:
raise Exception('Cannot check whether XML element with id="%s" and tag "%s" is native.' % (el.get('id'), el.tag))
return is_native
# ====== END: isNative ========
# ====== isStdType ========
def isStdType(el, class_name=None):
# Makes use of global variables: base_paths
is_std = False
can_check_tags = ['Class', 'Struct', 'Union', 'Enumeration']
if el.tag in can_check_tags:
# Use the optional class_name dict?
if class_name is not None:
if len(class_name['long_templ']) >= 5:
if class_name['long_templ'][0:5] == 'std::':
is_std = True
elif 'name' in el.keys():
namespaces_list = getNamespaces(el, include_self=True)
if namespaces_list[0] == 'std':
is_std = True
else:
is_std = False
return is_std
# ====== END: isStdType ========
# ====== isConstFunction ========
def isConstFunction(func_el):
is_const_func = False
if ('const' in func_el.keys()) and (func_el.get('const')=='1'):
is_const_func = True
return is_const_func
# ====== END: isConstFunction ========
# ====== getTemplateBracket ========
def getTemplateBracket(el):
src_file_name = gb.id_dict[el.get('file')].get('name')
line_number = int(el.get('line'))
f = open(src_file_name, 'r')
file_content = f.read()
f.close()
file_content_nocomments = removeComments(file_content, insert_blanks=True)
# Find index of the \n in line number line_number
count = 0
prev_pos = 0
for index,char in enumerate(file_content_nocomments):
if char=='\n':
count += 1
if count == line_number:
break
if char=='\n': # STUPID HACK
prev_pos = index
newline_pos = index
# Find the template parameter bracket, e.g. <typename A, typename B>
search_content = file_content_nocomments[:newline_pos]
start_pos = 0
end_pos = search_content.rfind('>')
if end_pos != -1:
balance = -1
for i in range(end_pos-1, -1, -1):
char = search_content[i]
if char == '>':
balance -= 1
elif char == '<':
balance += 1
if (balance == 0):
start_pos = i
break
template_bracket = search_content[start_pos:end_pos+1]
else:
template_bracket = '<>'
# print('TEMPLATE BRACKET: ', template_bracket)
# Isolate only the template variable names (last word in each entry)
if template_bracket == '<>':
temp_var_list = []
else:
temp_var_list = template_bracket[1:-1].split(',')
temp_var_list = [ e.strip() for e in temp_var_list]
temp_var_list = [ e.split()[-1] for e in temp_var_list]
# Return result
return template_bracket, temp_var_list
# ====== END: getTemplateBracket ========
# ====== getSpecTemplateTypes ========
def getSpecTemplateTypes(input_type, byname=False):
# If input is a string
if byname:
input_name = input_type
# If input is an xml element
else:
el = input_type
# Classes and functions must be treated differently
if el.tag in ['Class', 'Struct']:
input_name = el.get('name')
elif el.tag in ['Function', 'Method', 'OperatorMethod', 'OperatorFunction']:
namespaces_list = getNamespaces(el, include_self=True)
input_name = '::'.join(namespaces_list)
else:
raise Exception("Don't know how to get template types from XML element with tag: %s" % el.tag)
# Standardize the spacing between template brackets to simplify the parsing
while "<<" in input_name:
input_name = input_name.replace("<<", "< <")
while ">>" in input_name:
input_name = input_name.replace(">>", "> >")
input_name_no_templ, templ_bracket = removeTemplateBracket(input_name, return_bracket=True)
spec_types = templ_bracket.strip().lstrip('<').rstrip('>').strip()
# Identify the correct commas
pos = []
balance = 0
for i,c in enumerate(spec_types):
if c == '<':
balance += 1
if c == '>':
balance -= 1
if (balance==0) and (c == ','):
pos.append(i)
# Construct list of arguments
spec_types_list = []
prev_p = 0
for p in pos:
spec_types_list.append(spec_types[prev_p:p].strip())
prev_p = p+1
spec_types_list.append(spec_types[prev_p:].strip())
# Return result
if spec_types_list == ['']:
return []
else:
return spec_types_list
# ====== END: getSpecTemplateTypes ========
# ====== unpackAllSpecTemplateTypes ========
def unpackAllSpecTemplateTypes(input_bracket, result_list):
# # Help subsequent parsing by standardizing the spacing between template brackets
# while "<<" in input_bracket:
# input_bracket = input_bracket.replace("<<", "< <")
# while ">>" in input_bracket:
# input_bracket = input_bracket.replace(">>", "> >")
spec_types = getSpecTemplateTypes(input_bracket, byname=True)
for type_name in spec_types:
if '<' in type_name:
result_list.append(type_name)
unpackAllSpecTemplateTypes(type_name, result_list=result_list)
else:
result_list.append(type_name)
# ====== END: unpackAllSpecTemplateTypes ========
# ====== getAllTemplateTypes ========
def getAllTemplateTypes(type_name):
type_name_parts = []
current_type_name = type_name
while True:
namespace, short_type_name = removeNamespace(current_type_name, return_namespace=True)
type_name_parts.append(short_type_name)
if namespace == "":
break
else:
current_type_name = namespace
all_template_types = []
for type_part in type_name_parts:
unpackAllSpecTemplateTypes(type_part, all_template_types)
return all_template_types
# ====== END: getAllTemplateTypes ========
# ====== getBasicTypeName ========
def getBasicTypeName(type_name):
# If type name contains a template brackets
if '<' in type_name:
type_name_notempl, templ_bracket = removeTemplateBracket(type_name, return_bracket=True)
before_bracket, after_bracket = type_name.rsplit(templ_bracket,1)
if (len(after_bracket) > 0) and (after_bracket[0] == ' '):
space_after_bracket = True
else:
space_after_bracket = False
# Remove asterix and/or ampersand
before_bracket = before_bracket.replace('*', '').replace('&', '')
after_bracket = after_bracket.replace('*', '').replace('&', '')
# Remove 'const' and 'volatile'
before_bracket_list = before_bracket.split()
before_bracket_list = [item for item in before_bracket_list if item != 'const']
before_bracket_list = [item for item in before_bracket_list if item != 'volatile']
before_bracket = ' '.join(before_bracket_list)
after_bracket_list = after_bracket.split()
after_bracket_list = [item for item in after_bracket_list if item != 'const']
after_bracket_list = [item for item in after_bracket_list if item != 'volatile']
after_bracket = ' '.join(after_bracket_list)
basic_type_name = before_bracket + templ_bracket + ' '*space_after_bracket + after_bracket
# If no template bracket
else:
basic_type_name = type_name
# Remove asterix and/or ampersand
basic_type_name = basic_type_name.replace('*', '').replace('&', '')
# Remove 'const' and 'volatile'
basic_type_name_list = basic_type_name.split()
basic_type_name_list = [item for item in basic_type_name_list if item != 'const']
basic_type_name_list = [item for item in basic_type_name_list if item != 'volatile']
basic_type_name = ' '.join(basic_type_name_list)
# Return result
return basic_type_name
# ====== END: getBasicTypeName ========
# ====== removeComments ========
def removeComments(content, insert_blanks=False):
# Prepare list for storing tuples of the form: (start_position, stop_position)
content_lenght = len(content)
comment_sections = []
#
# Locate comments:
#
# -- One-line comments
temp_startpos = 0
while True:
# Find start of comment
search_pos = content[temp_startpos:].find('//')
if search_pos == -1:
break
else:
comment_start = temp_startpos + search_pos
# Find end of comment
search_pos = content[comment_start:].find('\n')
if search_pos == -1:
comment_end = content_lenght - 1
else:
comment_end = comment_start + search_pos
# Store positions
comment_sections.append( (comment_start, comment_end) )
# Update loop variable
temp_startpos = comment_end
# -- Multi-line comments
temp_startpos = 0
while True:
# Find start of comment
search_pos = content[temp_startpos:].find('/*')
# Are we done?
if search_pos == -1:
break
# Check for the potentially confusing case of comments starting with "//*"
elif (search_pos > 0) and (content[search_pos-1] == '/'):
# This is really a single-line comment which has been dealt with above,
# so we don't add it to the list of comment positions
comment_start= temp_startpos + search_pos
search_pos = content[comment_start:].find('\n')
if search_pos == -1:
comment_end = content_lenght - 1
else:
comment_end = comment_start + search_pos
# Update loop variable
temp_startpos = comment_end
# Now for the proper multi-line comments
else:
comment_start = temp_startpos + search_pos
# Find end of comment
search_pos = content[comment_start:].find('*/')
if search_pos == -1:
comment_end = content_lenght - 1
else:
comment_end = comment_start + search_pos + 1
# Store positions
comment_sections.append( (comment_start, comment_end) )
# Update loop variable
temp_startpos = comment_end
# Sort comment_sections from last to first, depending on stop position
comment_sections = sorted(comment_sections, key=itemgetter(1), reverse=True)
# Remove comments
prev_start_pos = 0
prev_stop_pos = 0
for start_pos, stop_pos in comment_sections:
new_lenght = len(content)
# Skip if the current comment was contained within the previous removed comment
if (start_pos > prev_start_pos) and (stop_pos < prev_stop_pos):
continue
# If not, go on to remove comment
else:
# Insert whitespace?
if insert_blanks == False:
content = content.replace( content[start_pos:stop_pos+1], '')
else:
# Construct string of spaces and newlines to replace comments
insert_string = ''
for char in content[start_pos:stop_pos+1]:
insert_string += ' '*(char!='\n') + '\n'*(char=='\n')
# Perform replacement
content = content.replace( content[start_pos:stop_pos+1], insert_string )
# Update loop variables
prev_start_pos = start_pos
prev_stop_pos = stop_pos
return content
# ====== END: removeComments ========
# ====== findType ========
def findType(el_input):
# check_keywords = ['const']
# additional_keywords = []
cv_qualifiers = []
is_reference = False
pointerness = 0
found_function_pointer = False
is_array = False
array_limits = []
el = el_input
if el.tag in ['FundamentalType', 'Class', 'Struct', 'Enumeration']:
type_id = el.get('id')
elif el.tag in ['Constructor']:
type_id = el.get('context')
el = gb.id_dict[type_id]
else:
type_id = el.get('type')
prev_tag = ''
while ('type' in el.keys()) or ('returns' in el.keys()):
# Get xml id to move further through the xml file
if el.tag in ['FunctionType', 'Function', 'Method', 'OperatorMethod']:
if (el.tag == 'FunctionType') and (prev_tag == 'PointerType'):
found_function_pointer = True
type_id = el.get('returns')
else:
type_id = el.get('type')
# id='_0' refer to elements not resolved by CastXML. Should be safe to stop here.
if type_id == '_0':
break
# Check for reference or pointer type
if el.tag == 'ReferenceType':
is_reference = True
if el.tag == 'PointerType':
pointerness += 1
# Pick up any extra keywords (e.g. 'const') from non-function xml elements
if el.tag == 'CvQualifiedType':
if 'const' in el.keys():
if 'const' not in cv_qualifiers:
cv_qualifiers.append('const')
if 'volatile' in el.keys():
if 'volatile' not in cv_qualifiers:
cv_qualifiers.append('volatile')
# Pick up any array indices
if el.tag == 'ArrayType':
is_array = True
if el.get('max') != '':
max_index_str = el.get('max')
max_index = int(max_index_str.strip('u'))
array_limits.append(max_index + 1)
# Store tag (to identify function pointers)
prev_tag = el.tag
# change xlm element 'el'
el = gb.id_dict[type_id]
# When we exit the loop, 'el' is at the final element.
# Now get the full name, including any namespaces.
name_and_namespaces = getNamespaces(el, include_self=True)
typename = '::'.join(name_and_namespaces)
type_dict = OrderedDict([])
type_dict['name'] = typename
type_dict['cv_qualifiers'] = cv_qualifiers
type_dict['is_reference'] = is_reference
type_dict['pointerness'] = pointerness
type_dict['id'] = type_id
type_dict['el'] = el
type_dict['is_function_pointer'] = found_function_pointer
type_dict['is_array'] = is_array
type_dict['array_limits'] = tuple(array_limits)
return type_dict
# ====== END: findType ========
# ====== findNewLinePos ========
def findNewLinePos(content, line_number):
count = 0
for index,char in enumerate(content):
if char=='\n':
count += 1
if count == line_number:
break
newline_pos = index
return newline_pos
# ====== END: findNewLinePos ========
# ====== getBracketPositions ========
def getBracketPositions(content, delims=['{','}']):
# Input:
# - Content string
# - List of left and right delimiters
#
# Output:
# - List of bracket positions: [l_pos, r_pos]
#
l_delim, r_delim = delims
# Cannot handle identical left and right delimiters
if l_delim == r_delim:
raise Exception('Left and right delimiters cannot be identical.')
# Prepare search
bracket_count = 0
start_count = False
balance = False
l_pos = 0
r_pos = 0
# Search
for i,c in enumerate(content):
if c == l_delim:
bracket_count += 1
if not start_count:
l_pos = i
start_count = True
if start_count and c == r_delim:
bracket_count -= 1
r_pos = i
if start_count and bracket_count == 0:
balance = True
break
# If brackets did not balance, raise exception
if not balance:
raise ValueError("No matching right delimiter for the first left delimiter.")
# Else, return the found bracket positions
else:
return [l_pos, r_pos]
# ====== END: getBracketPositions ========
# ====== addIndentation ========
def addIndentation(content, indent):
if indent == 0:
new_content = content
else:
lines = content.split('\n')
new_content = '\n'.join( [' '*indent + line for line in lines] )
# If the last char in content was a newline,
# remove the indentation that was added after that newline
if lines[-1] == '':
new_content = new_content[:-indent]
return new_content
# ====== END: addIndentation ========
# ====== getNamespaces ========
def getNamespaces(xml_el, include_self=False, xml_file_name=''):
namespaces = []
if include_self:
if 'name' in xml_el.keys():
namespaces.append(xml_el.get('name'))
else:
namespaces.append('')
current_xml_el = xml_el
while 'context' in current_xml_el.keys():
context_id = current_xml_el.get('context')
if xml_file_name == '':
context_xml_el = gb.id_dict[context_id]
else:
context_xml_el = gb.all_id_dict[xml_file_name][context_id]
# if 'name' in current_xml_el.keys():
if 'name' in context_xml_el.keys():
context_name = context_xml_el.get('name')
namespaces.append(context_name)
else:
break
#continue
current_xml_el = context_xml_el
namespaces.reverse()
# If present, remove the first, default namespace in the XML file from the list (with name='::')
if (len(namespaces) > 0) and (namespaces[0] == '::'):
namespaces = namespaces[1:]
return namespaces
# ====== END: getNamespaces ========
# ====== removeTemplateBracket ========
def removeTemplateBracket(type_name, return_bracket=False):
if ('<' in type_name) and ('>' in type_name):
r_pos = type_name.rfind('>')
if r_pos <= 0:
raise Exception("Unbalanced template brackets in type name '%s'" % type_name)
pos = r_pos-1
count = 1
while pos > -1:
if type_name[pos] == '>':
count += 1
elif type_name[pos] == '<':
count -= 1
if count == 0:
break
pos -= 1
if count != 0:
raise Exception("Unbalanced template brackets in type name '%s'" % type_name)
l_pos = pos
type_name_notempl = type_name[:l_pos] + type_name[r_pos+1:]
template_bracket = type_name[l_pos:r_pos+1]
else:
type_name_notempl = type_name
template_bracket = ''
if return_bracket:
return type_name_notempl, template_bracket
else:
return type_name_notempl
# ====== END: removeTemplateBracket ========
# ====== removeNamespace ========
def removeNamespace(type_name, return_namespace=False):
type_name_notempl = removeTemplateBracket(type_name)
if '::' in type_name_notempl:
namespace = type_name_notempl.rsplit('::',1)[0]
new_type_name = type_name.replace(namespace+'::','',1)
else:
new_type_name = type_name
namespace = ''
if return_namespace:
return namespace, new_type_name
else:
return new_type_name
# ====== END: removeNamespace ========
# ====== removeArgumentBracket ========
def removeArgumentBracket(func_signature, return_args_bracket=False):
args_bracket_start = func_signature.rfind('(')
func_signature_noargs = func_signature[:args_bracket_start]
args_bracket = func_signature[args_bracket_start:]
if return_args_bracket:
return func_signature_noargs, args_bracket
else:
return func_signature_noargs
# ====== END: removeArgumentBracket ========
# ====== isAcceptedType ========
def isAcceptedType(input_el):
is_accepted_type = False
type_dict = findType(input_el)
type_el = type_dict['el']
type_name = type_dict['name']
pointerness = type_dict['pointerness']
is_ref = type_dict['is_reference']
is_array = type_dict['is_array']
array_limits = type_dict['array_limits']
# BOSS cannot yet handle delarations of arrays with unspecified length
if (is_array) and (len(array_limits) == 0):
reason = "BOSS cannot yet handle arrays declared with unspecified length."
infomsg.TypeNotAccepted(input_el.get('name'), reason).printMessage()
is_accepted_type = False
return is_accepted_type
# BOSS cannot yet handle arrays of loaded types
if isLoadedClass(type_el) and is_array:
reason = "BOSS cannot yet handle arrays of a loaded type."
infomsg.TypeNotAccepted(input_el.get('name'), reason).printMessage()
is_accepted_type = False
return is_accepted_type
# BOSS cannot yet handle function pointers
if type_dict['is_function_pointer']:
reason = "BOSS cannot yet handle function pointers."
infomsg.TypeNotAccepted(input_el.get('name'), reason).printMessage()
is_accepted_type = False
return is_accepted_type
if type_el.tag in ['Class', 'Struct']:
namespaces_list = getNamespaces(type_el, include_self=True)
full_name = '::'.join(namespaces_list)
if (full_name in gb.accepted_types) or (full_name.replace(' ','') in gb.accepted_types):
is_accepted_type = True
elif type_el.tag in ['FundamentalType', 'Enumeration']:
type_name = type_el.get('name')
if type_name in gb.accepted_types:
is_accepted_type = True
else:
reason = "Cannot determine if XML element with id='%s' and tag '%s' corresponds to an accepted type. Assuming it does not." % (input_el.get('id'), input_el.tag)
infomsg.TypeNotAccepted(input_el.get('id'), reason).printMessage()
is_accepted_type = False
return is_accepted_type
# ====== END: isAcceptedType ========
# ====== isLoadedClass ========
def isLoadedClass(input_type, byname=False, class_name=None):
is_loaded_class = False
# If the class_name dict is passed as an argument, use it.
if class_name is not None:
if class_name['long_templ'] in cfg.load_classes:
is_loaded_class = True
else:
if byname:
type_name = input_type
# Remove '*' and '&'
type_name = type_name.replace('*','').replace('&','')
# Remove template bracket
type_name = type_name.split('<')[0]
# Check against cfg.load_classes
if type_name in cfg.load_classes:
is_loaded_class = True
else:
type_dict = findType(input_type)
type_el = type_dict['el']
if type_el.tag in ['Class', 'Struct']:
if type_dict['name'] in cfg.load_classes:
is_loaded_class = True
return is_loaded_class
# ====== END: isLoadedClass ========
# ====== isParentOfLoadedType ========
def isParentOfLoadedType(input_el):
is_parent = False
return is_parent
# ====== END: isParentOfLoadedType ========
# ====== constrAbsForwardDeclHeader ========
def constrAbsForwardDeclHeader(file_output_path):
import modules.classutils as classutils
# If this is the first time this function is executed, read initial code from header_templates/ folder
if file_output_path not in gb.new_code.keys():
f = open(gb.boss_dir+'/header_templates/standard_header_template.hpp')
initial_code = f.read()
f.close()
initial_code_tuple = (0, initial_code)
gb.new_code[file_output_path] = {'code_tuples':[initial_code_tuple], 'add_include_guard':True}
current_code = gb.new_code[file_output_path]['code_tuples'][0][1]
insert_code = ''
tag_pos = current_code.find('__INSERT_CODE_HERE__')
current_namespaces = []
for class_name_long, class_el in gb.loaded_classes_in_xml.items():
# print([class_name_full], [class_name_full.split('<',1)[0]], [class_name_full.split('<',1)[0].rsplit('::',1)[-1]])
namespaces = getNamespaces(class_el)
has_namespace = bool(len(namespaces))
namespace_str = '::'.join(namespaces) + '::'*has_namespace
# class_name = classutils.getClassNameDict(class_el)
abstr_class_name = classutils.getClassNameDict(class_el, abstract=True)
if namespaces != current_namespaces:
# close current namespace
insert_code += constrNamespace(current_namespaces, 'close', indent=cfg.indent)
# open new namespace
insert_code += constrNamespace(namespaces, 'open', indent=cfg.indent)
# update current namespace
current_namespaces = namespaces
# Add forward decls
n_indents = len(namespaces)
full_indent = ' '*n_indents*cfg.indent
if '<' in abstr_class_name['long_templ']:
is_template = True
else:
is_template = False
if is_template:
template_bracket = getTemplateBracket(class_el)[0]
if is_template:
insert_code += full_indent + 'template ' + template_bracket + '\n'
insert_code += full_indent + 'class ' + abstr_class_name['short'] + ';\n'
insert_code += full_indent + 'class ' + abstr_class_name['short_templ'] + ';\n'
else:
insert_code += full_indent + 'class ' + abstr_class_name['short_templ'] + ';\n'
# Close current namespace
insert_code += constrNamespace(current_namespaces, 'close', indent=cfg.indent)
insert_code += '\n'
new_code = current_code[:tag_pos] + insert_code + current_code[tag_pos:]
# Replace other code tags
new_code = replaceCodeTags(new_code)
new_code_tuple = (0,new_code)
# Overwrite existing code tuple
gb.new_code[file_output_path]['code_tuples'] = [(new_code_tuple)]
# ====== END: constrAbsForwardDeclHeader ========
# ====== constrWrpForwardDeclHeader ========
def constrWrpForwardDeclHeader(file_output_path):
import modules.classutils as classutils
# If this is the first time this function is executed, read initial code from header_templates/ folder
if file_output_path not in gb.new_code.keys():
f = open(gb.boss_dir+'/header_templates/standard_header_template.hpp')
initial_code = f.read()
f.close()
initial_code_tuple = (0, initial_code)
gb.new_code[file_output_path] = {'code_tuples':[initial_code_tuple], 'add_include_guard':True}
current_code = gb.new_code[file_output_path]['code_tuples'][0][1]
insert_code = ''
tag_pos = current_code.find('__INSERT_CODE_HERE__')
for class_name in gb.loaded_classes_in_xml.keys():
namespace, class_name_short = removeNamespace(class_name, return_namespace=True)
if namespace == '':
namespace_list = []
else:
namespace_list = namespace.split('::')
n_indents = len(namespace_list)
# - Open namespace
insert_code += constrNamespace(namespace_list, 'open')
# - Forward declaration
insert_code += ' '*n_indents*cfg.indent + 'class ' + class_name_short + ';\n'
# - Close namespace
insert_code += constrNamespace(namespace_list, 'close')
new_code = current_code[:tag_pos] + insert_code + current_code[tag_pos:]
# Replace other code tags
new_code = replaceCodeTags(new_code)
new_code_tuple = (0,new_code)
# Overwrite existing code tuple
gb.new_code[file_output_path]['code_tuples'] = [(new_code_tuple)]
# ====== END: constrWrpForwardDeclHeader ========
# ====== getParentClasses ========
def getParentClasses(class_el, only_native_classes=False, only_loaded_classes=False):
import modules.classutils as classutils
parent_classes = []
sub_el_list = class_el.findall('Base')
for sub_el in sub_el_list:
base_id = sub_el.get('type')
base_el = gb.id_dict[base_id]
if (only_loaded_classes) and (not isLoadedClass(base_el)):
continue
elif (only_native_classes) and (not isNative(base_el)):
continue
else:
base_access = sub_el.get('access')
base_virtual = bool( int( sub_el.get('virtual') ) )
base_name_dict = classutils.getClassNameDict(base_el)
abstr_base_name_dict = classutils.getClassNameDict(base_el, abstract=True)
is_accepted_type = isAcceptedType(base_el)
is_native = isNative(base_el)
is_fundamental = isFundamental(base_el)
is_std = isStdType(base_el)
is_loaded_class = isLoadedClass(base_el)
temp_dict = OrderedDict([])
temp_dict['class_name'] = base_name_dict
temp_dict['abstr_class_name'] = abstr_base_name_dict
temp_dict['wrapper_name'] = classutils.toWrapperType(base_name_dict['long'])
temp_dict['access'] = base_access
temp_dict['virtual'] = base_virtual
temp_dict['id'] = base_id
temp_dict['accepted'] = is_accepted_type
temp_dict['native'] = is_native
temp_dict['fundamental'] = is_fundamental
temp_dict['std'] = is_std
temp_dict['loaded'] = is_loaded_class
parent_classes.append(temp_dict)
return parent_classes
# ====== END: getParentClasses ========
# ====== getAllParentClasses ========
def getAllParentClasses(class_el, only_native_classes=True, only_loaded_classes=False, return_dicts=False, reverse_order=False):
import modules.classutils as classutils
parent_classes = []
done_parent_classes = []
temp_class_list = [class_el]
while len(temp_class_list) > 0:
current_class = temp_class_list.pop()
if 'bases' in current_class.keys():
for parent_class_id in current_class.get('bases').split():
# Remove accessor info from id, e.g. "private:_123" --> "_123"
parent_class_id = parent_class_id.split(':')[-1]
parent_class_el = gb.id_dict[parent_class_id]
if only_loaded_classes and not isLoadedClass(parent_class_el):
continue
elif only_native_classes and not isNative(parent_class_el):
continue
else:
if parent_class_el not in done_parent_classes:
temp_class_list.append(parent_class_el)
if return_dicts:
base_name_dict = classutils.getClassNameDict(parent_class_el)
abstr_base_name_dict = classutils.getClassNameDict(parent_class_el, abstract=True)
is_accepted_type = isAcceptedType(parent_class_el)
is_native = isNative(parent_class_el)
is_fundamental = isFundamental(parent_class_el)
is_std = isStdType(parent_class_el)
is_loaded_class = isLoadedClass(parent_class_el)
temp_dict = OrderedDict([])
temp_dict['class_name'] = base_name_dict
temp_dict['abstr_class_name'] = abstr_base_name_dict
temp_dict['wrapper_name'] = classutils.toWrapperType(base_name_dict['long'])
temp_dict['id'] = parent_class_id
temp_dict['accepted'] = is_accepted_type
temp_dict['native'] = is_native
temp_dict['fundamental'] = is_fundamental
temp_dict['std'] = is_std
temp_dict['loaded'] = is_loaded_class
parent_classes.append(temp_dict)
done_parent_classes.append(parent_class_el)
else:
parent_classes.append(parent_class_el)
done_parent_classes.append(parent_class_el)
if reverse_order:
return parent_classes[::-1]
else:
return parent_classes
# ====== END: getAllParentClasses ========
# ====== getAllTypesInClass ========
def getAllTypesInClass(class_el, include_parents=False):
import modules.classutils as classutils
import modules.funcutils as funcutils
all_types = []
check_member_elements = getMemberElements(class_el)
class_id = class_el.get('id')
for mem_el in check_member_elements:
if mem_el.tag in ['Constructor', 'Destructor', 'Method', 'OperatorMethod']:
args_list = funcutils.getArgs(mem_el)
for arg_dict in args_list:
arg_type_el = gb.id_dict[arg_dict['id']]
arg_type_name = classutils.getClassNameDict(arg_type_el)
arg_type_dict = OrderedDict([])
arg_type_dict['class_name'] = arg_type_name
arg_type_dict['el'] = arg_type_el
all_types.append(arg_type_dict)
if ('type' in mem_el.keys()) or ('returns' in mem_el.keys()):
mem_type_dict = findType(mem_el)
type_el = mem_type_dict['el']
type_name = classutils.getClassNameDict(type_el)
type_dict = OrderedDict([])
type_dict['class_name'] = type_name
type_dict['el'] = type_el
all_types.append(type_dict)
if include_parents:
parent_classes = getParentClasses(class_el, only_native_classes=False, only_loaded_classes=False)
for parent_dict in parent_classes:
small_parent_dict = OrderedDict([])
small_parent_dict['class_name'] = parent_dict['class_name']
small_parent_dict['el'] = gb.id_dict[parent_dict['id']]
all_types.append(small_parent_dict)
return all_types
# ====== END: getAllTypesInClass ========
# ====== getMemberElements ========
def getMemberElements(el, include_artificial=False):
member_elements = []
if 'members' in el.keys():
for mem_id in el.get('members').split():
mem_el = gb.id_dict[mem_id]
# Check if this member element should be ditched
if mem_el.tag in ['Method', 'OperatorMethod', 'Field', 'Variable']:
if 'name' in mem_el.keys():
namespaces_list = getNamespaces(mem_el, include_self=True)
full_name = '::'.join(namespaces_list)
if full_name in cfg.ditch:
continue
if include_artificial:
member_elements.append(mem_el)
else:
if not 'artificial' in mem_el.keys():
member_elements.append(mem_el)
return member_elements
# ====== END: getMemberElements ========
# ====== getMemberFunctions ========
def getMemberFunctions(class_el, include_artificial=False, include_inherited=False, only_accepted=True, limit_pointerness=True, include_operators=False):
import modules.funcutils as funcutils
all_classes = [class_el]
all_members = []
all_functions = []
# If include_inherited=True, append all (native) parent classes
# the list 'all_classes'
if include_inherited:
parent_classes = getAllParentClasses(class_el, only_loaded_classes=True)
all_classes = all_classes + parent_classes
# Get all member elements
for el in all_classes:
class_members = getMemberElements(el, include_artificial=include_artificial)
all_members = all_members + class_members
# Extract only regular member functions (no variables, constructors, destructors, ...)
for mem_el in all_members:
if (mem_el.tag == 'Method' or (include_operators==True and mem_el.tag=='OperatorMethod')) and (mem_el.get('access') == 'public'):
if only_accepted and funcutils.ignoreFunction(mem_el, limit_pointerness=limit_pointerness):
method_name = mem_el.get('name')
if mem_el.tag=='OperatorMethod':
method_name = 'operator' + method_name
reason = "Makes use of a non-accepted type."
infomsg.IgnoredFunction(method_name, reason).printMessage()
continue
else:
all_functions.append(mem_el)
return all_functions
# ====== END: getMemberFunctions ========
# ====== getAllTypesInFunction ========
def getAllTypesInFunction(func_el):
import modules.classutils as classutils
import modules.funcutils as funcutils
all_types = []
func_id = func_el.get('id')
if func_el.tag in ['Function', 'Constructor']:
args_list = funcutils.getArgs(func_el)
for arg_dict in args_list:
arg_type_el = gb.id_dict[arg_dict['id']]
arg_type_name = classutils.getClassNameDict(arg_type_el)
arg_type_dict = OrderedDict([])
arg_type_dict['class_name'] = arg_type_name
arg_type_dict['el'] = arg_type_el
all_types.append(arg_type_dict)
if ('type' in func_el.keys()) or ('returns' in func_el.keys()) or (func_el.tag=='Constructor' and 'context' in func_el.keys()):
mem_type_dict = findType(func_el)
type_el = mem_type_dict['el']
type_name = classutils.getClassNameDict(type_el)
type_dict = OrderedDict([])
type_dict['class_name'] = type_name
type_dict['el'] = type_el
all_types.append(type_dict)
return all_types
# ====== END: getAllTypesInFunction ========
# ====== constrNamespace ========
def constrNamespace(namespaces, open_or_close, indent=cfg.indent):
if len(namespaces) > 0:
if namespaces[0] == '::':
namespaces = namespaces[1:]
code = ''
if open_or_close == 'open':
n_indents = 0
for ns in namespaces:
code += ' '*n_indents*indent + 'namespace ' + ns + '\n'
code += ' '*n_indents*indent + '{' + '\n'
n_indents += 1
elif open_or_close == 'close':
n_indents = len(namespaces)
for ns in namespaces:
n_indents -= 1
code += ' '*n_indents*indent + '}' + '\n'
else:
raise ValueError("Second argument must be either 'open' or 'close'.")
return code
# ====== END: constrNamespace ========
# ====== pointerAndRefCheck ========
def pointerAndRefCheck(input_type, byname=False):
#
# Input type should either be an XML element (byname=False)
# or a string (byname=True)
#
if byname:
type_name = input_type
# Remove template bracket
if '<' in type_name:
type_name = type_name.split('<',1)[0] + type_name.rsplit('>',1)[-1]
# Check pointerness
pointerness = type_name.count('*')
# Check reference
is_reference = bool('&' in type_name)
else:
type_dict = findType(input_type)
pointerness = type_dict['pointerness']
is_reference = type_dict['is_reference']
return pointerness, is_reference
# ====== END: pointerAndRefCheck ========
# ====== addIncludeGuard ========
def addIncludeGuard(code, file_name, prefix='', suffix='', uppercase=False):
file_name = file_name.rstrip('.FOR_GAMBIT')
if suffix == '':
guard_var = '__' + (prefix + '__')*bool(len(prefix)) + file_name.replace('.','_') + '__'
else:
file_name_no_ext, file_ext = os.path.splitext(file_name)
guard_var = '__' + (prefix + '__')*bool(len(prefix)) + file_name_no_ext.replace('.','_') + '_' + suffix + file_ext.replace('.','_') + '__'
# file_name_no_ext, file_ext = os.path.splitext(os.path.basename(file_name))
# guard_var = '__' + (prefix + '__')*bool(len(prefix)) + file_name_no_ext.replace('.','_') + ('__' + suffix)*bool(len(suffix)) + '__'
if uppercase:
guard_var = guard_var.upper()
guard_code_top = '#ifndef ' + guard_var + '\n' + '#define ' + guard_var + '\n'
guard_code_bottom = '#endif /* ' + guard_var + ' */\n'
new_code = guard_code_top + '\n' + code + '\n' + guard_code_bottom
return new_code
# ====== END: addIncludeGuard ========
# ====== identifyIncludedHeaders ========
def identifyIncludedHeaders(content, only_native=True):
return_dict = OrderedDict()
# Remove comments
content = removeComments(content, insert_blanks=True)
content_list = content.split('\n')
# Parse content and identify header file names
headers_in_file = []
for line in content_list:
line = line.strip()
if line[0:8] == '#include':
# Make sure there's a whitespace after '#include'
if line[8] != ' ':
line = line[0:8] + ' ' + line[8:]
header_file_name = line.split()[1]
# Skip standard headers (of the form: #include <FILENAME>)
if header_file_name[0] == '<':
continue
else:
header_file_name = header_file_name.strip('"')
headers_in_file.append(os.path.basename(header_file_name))
else:
continue
# Connect with XML elements
for check_file_path, file_el in gb.file_dict.items():
# - If only_native=True, check for match with cfg.base_paths
if only_native:
is_native_file = False
for accepted_path in cfg.base_paths:
if accepted_path in os.path.dirname(check_file_path):
is_native_file = True
break
if not is_native_file:
continue
# - Cut down to file name only
check_file_name = os.path.basename(check_file_path)
# - Keep XML id if the corresponding file name mathces with an identified header
if check_file_name in headers_in_file:
return_dict[check_file_name] = file_el.get('id')
return return_dict
# ====== END: identifyIncludedHeaders ========
# ====== isHeader ========
def isHeader(file_el):
is_header = False
file_name = file_el.get('name')
extension = os.path.splitext(file_name)[1]
if extension != '':
if extension.lower() in ['.hpp', '.h', '.hh', '.hxx', cfg.header_extension.lower()]:
is_header = True
return is_header
# ====== END: isHeader ========
# ====== getIncludeStatements ========
def getIncludeStatements(input_el, convert_loaded_to='none', exclude_types=[],
input_element='class', forward_declared='exclude',
use_full_path=False, include_parents=False):
include_statements = []
# Check string arguments
convert_loaded_to = convert_loaded_to.lower()
input_element = input_element.lower()
forward_declared = forward_declared.lower()
if convert_loaded_to not in ['none', 'abstract', 'wrapper', 'wrapper_decl', 'wrapper_def']:
raise Exception("getIncludeStatements: Keyword argument 'convert_loaded_to=' must be either 'none', 'abstract', 'wrapper', 'wrapper_decl' or 'wrapper_def'.")
if input_element not in ['class', 'function']:
raise Exception("getIncludeStatements: Keyword argument 'input_element=' must be either 'class' or 'function'.")
if forward_declared not in ['include', 'exclude', 'only']:
raise Exception("getIncludeStatements: Keyword argument 'forward_declared=' must be either 'include', 'exclude' or 'only'.")
# Get list of all types used in this class/function (each entry is a dict)
if input_element == 'class':
all_types = getAllTypesInClass(input_el, include_parents=include_parents)
elif input_element == 'function':
all_types = getAllTypesInFunction(input_el)
# Get file name and line number of the current class/function
start_line_number = int( input_el.get('line') )
start_file_el = gb.id_dict[ input_el.get('file') ]
start_file_path = start_file_el.get('name')
# Read file from beginning to position of class/function definition
start_file = open(start_file_path, 'r')
start_file_content = start_file.readlines()[0:start_line_number]
start_file_content = ''.join(start_file_content)
start_file.close()
# Identify included header files from this file (utils.identifyIncludedHeaders returns a dict of the form {header_file_name: xml_id})
included_headers_dict = identifyIncludedHeaders(start_file_content, only_native=True)
# Move up the header tree and identify all the relevant (native) included headers
header_paths = [ gb.id_dict[file_id].get('name') for file_id in included_headers_dict.values() ]
header_paths_done = []
while len(header_paths) > 0:
header_path = header_paths.pop()
# Read header
header = open(header_path, 'r')
header_content = header.read()
header.close()
# Identify new headers
new_included_headers = identifyIncludedHeaders(header_content, only_native=True)
# Add any new headers to included_headers_dict
for file_name, file_id in new_included_headers.items():
if file_name not in included_headers_dict.keys():
included_headers_dict[file_name] = file_id
# Add any new headers to the list of header files to check
new_header_paths = [ gb.id_dict[header_id].get('name') for header_id in new_included_headers.values() ]
for new_path in new_header_paths:
if (new_path not in header_paths) and (new_path not in header_paths_done):
header_paths.append(new_path)
# Keep track of headers we've done
header_paths_done.append(header_path)
# Determine what include statements to generate:
for type_dict in all_types:
type_el = type_dict['el']
type_name = type_dict['class_name']
if type_name in exclude_types:
continue
if isAcceptedType(type_el):
if isFundamental(type_el):
pass
elif isEnumeration(type_el):
pass
elif isLoadedClass(type_el):
# For each loaded class used in this class/function, check whether the corresponding class definition can be
# found in the current file (above current class/function) or among the included headers. If no such class
# definition is found, it must be a case of simply using forward declaration.
# TODO: Why isn't it enough just to check for the 'incomplete' key in the type_el?
# Need to check this again...
type_file_id = type_el.get('file')
type_line_number = int(type_el.get('line'))
if ('incomplete' in type_el.keys() and type_el.get('incomplete')=='1'):
is_incomplete = True
else:
is_incomplete = False
if (type_file_id in included_headers_dict.values()) :
type_definition_found = True
elif (type_file_id == input_el.get('file')) and (type_line_number < start_line_number):
type_definition_found = True
else:
type_definition_found = False
if (not type_definition_found) and (forward_declared=='exclude'):
# This must be a case of a type that is only forward declared. Don't include any header (as this will typically lead to a 'header loop').
continue
elif (type_definition_found) and (forward_declared=='only'):
# This must be a case of a type that *is* fully declared, so we ignore it if forward_declared=='only'.
continue
else:
if convert_loaded_to == 'none':
type_file_el = gb.id_dict[type_file_id]
type_file_full_path = type_file_el.get('name')
# If the xml element we have for the type is only for a forward declaration
# we must search all other xml files for the complete type declaration.
if is_incomplete:
for xml_file_name in gb.all_name_dict.keys():
try:
new_type_el = gb.all_name_dict[xml_file_name][type_name['long_templ']]
except KeyError:
new_type_el = None
if new_type_el is not None:
if 'incomplete' not in new_type_el.keys():
new_type_file_id = new_type_el.get('file')
new_type_file_el = gb.all_id_dict[xml_file_name][new_type_file_id]
# Set new header path and break the loop
type_file_full_path = new_type_file_el.get('name')
break
if isHeader(type_file_el):
use_path = shortenHeaderPath(type_file_full_path)
include_statements.append( '#include "' + use_path + '"')
else:
reason = "Found declaration of loaded type '%s' in file '%s', but this file is not recognized as a header file." % (type_name['long_templ'], type_file_full_path)
infomsg.NoIncludeStatementGenerated(type_name['long_templ'], reason).printMessage()
else:
if use_full_path:
header_key = convert_loaded_to + '_fullpath'
else:
header_key = convert_loaded_to
include_statements.append('#include "' + gb.new_header_files[type_name['long']][header_key] + '"')
elif isStdType(type_el):
if type_name['long'] in gb.std_headers:
header_name = gb.std_headers[type_name['long']].strip()
if (header_name[0] == '<') and (header_name[-1] == '>'):
include_statements.append('#include ' + gb.std_headers[type_name['long']])
else:
include_statements.append('#include "' + gb.std_headers[type_name['long']] + '"')
else:
reason = "The standard type '%s' has no specified header file. Please update modules/gb.py." % type_name['long_templ']
infomsg.NoIncludeStatementGenerated(type_name['long_templ'], reason).printMessage()
else:
is_known, index = isInList(type_name['long_templ'], cfg.known_classes.keys(), return_index=True, ignore_whitespace=True)
if not is_known:
is_known, index = isInList(type_name['long'], cfg.known_classes.keys(), return_index=True, ignore_whitespace=True)
if is_known:
header_name = list(cfg.known_classes.values())[index]
if (header_name[0] == '<') and (header_name[-1] == '>'):
include_statements.append('#include ' + header_name)
else:
include_statements.append('#include "' + header_name + '"')
else:
reason = "The type '%s' has no specified header file. Please update the 'known_classes' dictionary in the config file." % type_name['long_templ']
infomsg.NoIncludeStatementGenerated(type_name['long_templ'], reason).printMessage()
else:
infomsg.NoIncludeStatementGenerated( type_name['long_templ'] ).printMessage()
# Remove duplicates and return list (ordered)
include_statements = list( OrderedDict.fromkeys(include_statements) )
include_statements = orderIncludeStatements(include_statements)
return include_statements
# ====== END: getIncludeStatements ========
# ====== getOriginalHeaderPath ========
def getOriginalHeaderPath(el, full_path=False):
file_id = el.get('file')
file_el = gb.id_dict[file_id]
file_full_path = file_el.get('name')
if isHeader(file_el):
if full_path:
header_path = file_full_path
else:
header_path = os.path.basename(file_full_path)
else:
raise exceptions.ReturnError("The file %s is not recognized as a header file." % file_full_path)
return header_path
# ====== END: getOriginalHeaderPath ========
# ====== shortenHeaderPath ========
def shortenHeaderPath(full_path):
shorter_path = full_path
# Split path into individual parts
path_parts = pathSplitAll(full_path)
for incl_path in cfg.include_paths:
incl_path_parts = pathSplitAll(incl_path)
n = len(incl_path_parts)
if len(path_parts) >= n:
if path_parts[0:n] == incl_path_parts:
shorter_path = os.path.join(*path_parts[n:])
break
return shorter_path
# ====== END: shortenHeaderPath ========
# ====== constrNamespaceFromTags ========
def constrNamespaceFromTags(content, new_namespace, open_tag, close_tag):
new_namespace_list = new_namespace.split('::')
new_content = content
while True:
prev_content = new_content
# Find tag positions
open_pos = prev_content.find(open_tag)
close_pos = prev_content.find(close_tag)
if (open_pos == -1) and (close_pos == -1):
# No tags found. Do nothing more.
# return content
break
elif (open_pos == -1) or (close_pos == -1):
raise Exception('Matching pair of namespace tags %s and %s not found in given content.' % (open_tag, close_tag))
else:
pass
# Split content into three parts
content_before = prev_content[:open_pos]
content_within = prev_content[open_pos:close_pos]
content_after = prev_content[close_pos:]
# Remove the namespace tags
content_within = content_within.replace(open_tag, '')
content_after = content_after.replace(close_tag, '', 1)
# Add indentation to middle part
content_within = addIndentation(content_within, cfg.indent*len(new_namespace_list))
# Contruct new namespace and combine code
open_new_namespace_code = constrNamespace(new_namespace_list, 'open', indent=cfg.indent)
close_new_namespace_code = constrNamespace(new_namespace_list, 'close', indent=cfg.indent)
new_content = content_before + open_new_namespace_code + content_within + close_new_namespace_code + content_after
return new_content
# ====== END: constrNamespaceFromTags ========
# ====== replaceCodeTags ========
def replaceCodeTags(input, file_input=False, write_file=False):
# Input is either a file name or a string with content
if file_input:
f = open(input, 'r')
new_content = f.read()
f.close()
else:
new_content = input
# Replace various tags in template code with code specific for the current backend
new_content = new_content.replace('__BACKEND_NAME__' , cfg.gambit_backend_name)
new_content = new_content.replace('__BACKEND_VERSION__' , cfg.gambit_backend_version)
new_content = new_content.replace('__BACKEND_SAFE_VERSION__' , gb.gambit_backend_safeversion)
new_content = new_content.replace('__BACKEND_REFERENCE__' , cfg.gambit_backend_reference)
new_content = new_content.replace('__CODE_SUFFIX__' , gb.code_suffix)
new_content = new_content.replace('__PATH_TO_FRWD_DECLS_ABS_CLASSES_HEADER__', os.path.join(gb.backend_types_basedir, gb.gambit_backend_name_full, gb.frwd_decls_abs_fname + cfg.header_extension))
new_content = new_content.replace('__PATH_TO_IDENTIFICATION_HEADER__' , os.path.join(gb.backend_types_basedir, gb.gambit_backend_name_full, 'identification.hpp'))
new_content = new_content.replace('__PATH_TO_BACKEND_UNDEFS_HEADER__' , os.path.join(gb.gambit_backend_incl_dir, "backend_undefs.hpp"))
# Should a new file be written?
if file_input and write_file:
f.open(input,'w')
f.write(new_content)
f.close()
# Return code
return new_content
# ====== END: replaceCodeTags ========
# ====== removeCodeTags ========
def removeCodeTags(content, remove_tags_list):
new_content = content
# Remove tags
for tag in remove_tags_list:
new_content = new_content.replace(tag, '')
# Return code
return new_content
# ====== END: removeCodeTags ========
# ====== constrLoadedTypesHeaderContent ======
def constrLoadedTypesHeaderContent():
#
# Construct the code lines for the loaded classes, containg all the factory symbols and argument brackets for that class
#
class_lines = []
# Loop over all classes
for class_name in gb.classes_done:
if not class_name['long'] in gb.factory_info.keys():
reason = "Probably there are no public and accepted constructors."
infomsg.NoLoadedTypesEntry(class_name['long'], reason).printMessage()
else:
class_line = ' (( /*class*/'
namespace, class_name_short = removeNamespace(class_name['long'], return_namespace=True)
if namespace == '':
namespace_list = []
else:
namespace_list = namespace.split('::')
for ns_part in namespace_list:
class_line += '(' + ns_part + ')'
class_line += '(' + class_name['short'] + '),'
class_line += ' /*constructors*/'
for info_dict in gb.factory_info[ class_name['long'] ]:
class_line += '(("' + info_dict['name'] + '",' + info_dict['args_bracket'].replace(' ::', ' ').replace('(::', '(') + ')) '
class_line += ')) \\'
class_lines.append(class_line)
class_lines_code = ''
class_lines_code += '#define ' + gb.gambit_backend_name_full + '_all_data \\\n'
class_lines_code += '\n'.join(class_lines) + '\n'
#
# Construct include guards with additional ' 1' appended to the line starting with #define
#
incl_guard = addIncludeGuard('', 'loaded_types.hpp', prefix='', suffix=gb.gambit_backend_name_full)
incl_guard_lines = incl_guard.split('\n')
incl_guard_start = '\n'.join(incl_guard_lines[:2]) + ' 1\n'
incl_guard_end = incl_guard_lines[-2] + '\n'
#
# Construct include statements, surrounded by pragma directives
#
incl_statements_code = ''
for pragma_directive in cfg.pragmas_begin:
incl_statements_code += pragma_directive.strip() + '\n'
for class_name in gb.classes_done:
if class_name['long'] in gb.factory_info.keys():
namespace, class_name_short = removeNamespace(class_name['long'], return_namespace=True)
incl_statements_code += '#include "' + gb.wrapper_header_prefix + class_name['short'] + cfg.header_extension + '"\n'
incl_statements_code += '#include "identification.hpp"\n'
for pragma_directive in cfg.pragmas_end:
incl_statements_code += pragma_directive.strip() + '\n'
#
# Combine everything to construct header code
#
code = ''
code += incl_guard_start
code += '\n'
code += incl_statements_code
code += '\n'
code += '// Indicate which types are provided by this backend, and what the symbols of their factories are.\n'
code += class_lines_code
code += '\n'
code += '// If the default version has been loaded, set it as default.\n'
code += '#if ALREADY_LOADED(CAT_3(BACKENDNAME,_,CAT(Default_,BACKENDNAME)))\n'
code += ' SET_DEFAULT_VERSION_FOR_LOADING_TYPES(BACKENDNAME,SAFE_VERSION,CAT(Default_,BACKENDNAME))\n'
code += '#endif\n'
code += '\n'
code += '// Undefine macros to avoid conflict with other backends.\n'
code += '#include "' + os.path.join(gb.gambit_backend_incl_dir, "backend_undefs.hpp") + '"\n'
code += '\n'
code += incl_guard_end
return code
# ====== END: constrLoadedTypesHeaderContent ======
# ====== constrEnumDeclHeader ========
def constrEnumDeclHeader(enum_el_list, file_output_path):
import modules.classutils as classutils
# If this is the first time this function is executed, read initial code from header_templates/ folder
if file_output_path not in gb.new_code.keys():
f = open(gb.boss_dir+'/header_templates/standard_header_template.hpp')
initial_code = f.read()
f.close()
initial_code_tuple = (0, initial_code)
gb.new_code[file_output_path] = {'code_tuples':[initial_code_tuple], 'add_include_guard':True}
current_code = gb.new_code[file_output_path]['code_tuples'][0][1]
insert_code = ''
tag_pos = current_code.find('__INSERT_CODE_HERE__')
for enum_el in enum_el_list:
# Skip any enumerated type that is not native to the external code
if not isNative(enum_el):
continue
# Get enum names and values
enum_members_list = []
for sub_el in list(enum_el):
if sub_el.tag == 'EnumValue':
member_string = sub_el.get('name') + '=' + sub_el.get('init')
enum_members_list.append(member_string)
all_members_string = ', '.join(enum_members_list)
# Get namespace list
namespace_list = getNamespaces(enum_el)
n_indents = len(namespace_list)
# - Open namespace
insert_code += constrNamespace(namespace_list, 'open')
insert_code += ' '*n_indents*cfg.indent + 'typedef enum {' + all_members_string + '} ' + enum_el.get('name') + ';\n'
# - Close namespace
insert_code += constrNamespace(namespace_list, 'close')
new_code = current_code[:tag_pos] + insert_code + current_code[tag_pos:]
# Replace other code tags
new_code = replaceCodeTags(new_code)
new_code_tuple = (0,new_code)
# Overwrite existing code tuple
gb.new_code[file_output_path]['code_tuples'] = [(new_code_tuple)]
# ====== END: constrEnumDeclHeader ========
# ====== castxmlRunner ========
# Calls castxml from the shell (via modules.shelltimeout).
def castxmlRunner(input_file_path, include_paths_list, xml_output_path, use_castxml_path=None):
# Avoid including intel headers when in "gnu mode" by
# temporarily unsetting some environment variables
if 'gnu' in cfg.castxml_cc_id:
temp_env_vars = {}
for var_name in ['CPATH', 'C_INCLUDE_PATH', 'CPLUS_INCLUDE_PATH']:
try:
if 'intel' in os.environ[var_name].lower():
temp_env_vars[var_name] = str(os.environ[var_name])
os.environ[var_name] = ''
except KeyError:
pass
# Set the use_castxml_path if it is not already set
castxml_system_path = 'castxml'
castxml_local_path = os.path.join(gb.boss_dir,"castxml/bin/castxml")
if use_castxml_path is None:
if gb.has_castxml_system:
use_castxml_path = castxml_system_path
elif gb.has_castxml_local:
use_castxml_path = castxml_local_path
else:
raise Exception('No castxml binary found.')
# Construct castxml command to run
castxml_cmd = use_castxml_path + ' --castxml-gccxml -x c++'
# Add castxml settings from cfg file
castxml_cmd += ' --castxml-cc-' + cfg.castxml_cc_id + ' "(" ' + cfg.castxml_cc
if cfg.castxml_cc_opt != '':
castxml_cmd += ' ' + cfg.castxml_cc_opt
castxml_cmd += ' ")" '
# - Add include paths
for incl_path in include_paths_list:
castxml_cmd += ' -I' + incl_path
# - Add standard include paths
for std_incl_path in gb.std_include_paths:
castxml_cmd += ' -I' + std_incl_path
# - Add the input file path (full path)
castxml_cmd += ' ' + input_file_path
# - Add castxml option that specifies the xml output file: input_file_short_name.xml
castxml_cmd += ' -o ' + xml_output_path
# Run castxml
print(' Running command: ' + castxml_cmd)
did_fail = False
error_message = ''
output_tmpfile = tempfile.TemporaryFile()
try:
p = subprocess.Popen(shlex.split(castxml_cmd), stdout=output_tmpfile, stderr=output_tmpfile)
p.wait()
except subprocess.CalledProcessError as e:
did_fail = True
error_message = e.message
# Reset environment variables
if 'gnu' in cfg.castxml_cc_id:
for var_name, value in temp_env_vars.items():
os.environ[var_name] = value
# Get output from tempfile
output_tmpfile.seek(0)
output = output_tmpfile.read()
output_tmpfile.close()
# Any error that did not result in a CalledProcessError?
if (p.returncode != 0):
did_fail = True
if did_fail:
print(' ' + modifyText('ERROR: CastXML failed.','red'))
# Get error message
print()
print(modifyText('START CASTXML OUTPUT','underline'))
print()
print(output)
print(modifyText('END CASTXML OUTPUT','underline'))
print()
if error_message != '':
print("CalledProcessError.message:", error_message)
print()
# If it fails with the syste-wide castxml binary, try again with the local one
if (did_fail and use_castxml_path==castxml_system_path and gb.has_castxml_local):
print(' ' + modifyText('Will retry with castxml binary in ' + castxml_local_path,'yellow') )
did_fail = False
use_castxml_path = castxml_local_path
castxmlRunner(input_file_path, include_paths_list, xml_output_path, use_castxml_path=use_castxml_path)
# If it fails with icpc, try again with g++.
if ("gnu" in cfg.castxml_cc_id) and ("icpc" in cfg.castxml_cc) and did_fail:
print()
print(' ' + modifyText('Will retry with --castxml-cc=g++','yellow'))
print()
cfg.castxml_cc = 'g++'
castxmlRunner(input_file_path, include_paths_list, xml_output_path, use_castxml_path=use_castxml_path)
# Print error report
elif did_fail:
raise Exception('castxml failed')
else:
print(' ' + modifyText('Command finished successfully.','green'))
print()
# ====== END: castxmlRunner ========
# ====== pathSplitAll ========
def pathSplitAll(path):
all_parts = []
current_path = path
while True:
parts = os.path.split(current_path)
if parts[0] == current_path: # Stopping criterion for absolute paths
all_parts.insert(0, parts[0])
break
elif parts[1] == current_path: # Stopping criterion for relative paths
all_parts.insert(0, parts[1])
break
else:
current_path = parts[0]
all_parts.insert(0, parts[1])
return all_parts
# ====== pathSplitAll ========
# ====== fillAcceptedTypesList ========
def fillAcceptedTypesList():
import modules.classutils as classutils
# Sets to store type names
fundamental_types = set()
std_types = set()
known_classes = set()
# enumeration_types = set()
loaded_classes = set()
# Keep track of how many types have been checked
type_counter = 0
print()
#
# Collect names of all fundamental, std, enumeration, known and loaded types that are acceptable
#
for xml_file in gb.all_id_dict.keys():
# Reset some variables for each new xml file
new_fundamental_types = []
new_std_types = []
new_known_classes = []
# new_enumeration_types = []
new_loaded_classes = []
initGlobalXMLdicts(xml_file)
#
# Loop over all named elements in the xml file
#
for full_name, el in gb.name_dict.items():
# Only consider types
if el.tag not in ['Class', 'Struct', 'FundamentalType', 'Enumeration']:
continue
type_counter += 1
if type_counter%500 == 0:
print(' - %i types classified...' % (type_counter))
# To save a bit of time, construct class name dict once and pass to remaining checks
class_name = classutils.getClassNameDict(el)
# Skip problematic types
if isProblematicType(el):
continue
#
# Known class?
#
is_known_class = isKnownClass(el, class_name=class_name)
if is_known_class:
new_known_classes.append(full_name)
# Skip incomplete types
if ('incomplete' in el.keys()) and (el.get('incomplete') == '1'):
continue
#
# Fundamental type?
#
is_fundamental = isFundamental(el)
if is_fundamental:
new_fundamental_types.append(full_name)
#
# Std type?
#
is_std_type = isStdType(el, class_name=class_name)
if is_std_type:
new_std_types.append(full_name)
# #
# # Enumeration type?
# #
# is_enumeration = isEnumeration(el)
# if is_enumeration:
# new_enumeration_types.append( '::'.join( getNamespaces(el, include_self=True) ) )
#
# Loaded type?
#
is_loaded_class = isLoadedClass(el, byname=False, class_name=class_name)
if is_loaded_class:
new_loaded_classes.append(full_name)
#
# Update sets of types
#
fundamental_types = fundamental_types.union(set(new_fundamental_types))
std_types = std_types.union(set(new_std_types))
known_classes = known_classes.union(set(new_known_classes))
# enumeration_types = enumeration_types.union(set(new_enumeration_types))
loaded_classes = loaded_classes.union(set(new_loaded_classes))
# Print final number of types classified
print(' - %i types classified.' % (type_counter))
# Fill global list
gb.accepted_types = list(loaded_classes) + list(known_classes) + list(fundamental_types) + list(std_types)
# gb.accepted_types = list(loaded_classes) + list(fundamental_types) + list(std_types) + list(enumeration_types)
# ====== END: fillAcceptedTypesList ========
# ====== isProblematicType ========
# Use this to identify types that BOSS currently have problems with,
# but that (hopefully) should be fixed in the future.
def isProblematicType(el):
is_problematic = False
#
# Check: types that use native types as template arguments
#
if el.tag in ['Class', 'Struct', 'FundamentalType']:
# Get list of all template arguments (unpack any nested template arguments)
unpacked_template_args = getAllTemplateTypes(el.get('name'))
# If no template arguments, continue
if unpacked_template_args == []:
pass
else:
for templ_arg in unpacked_template_args:
# Remove asterix and/or ampersand
base_templ_arg = getBasicTypeName(templ_arg)
# Check that this type name is listed in gb.name_dict
try:
type_el = gb.name_dict[base_templ_arg]
except KeyError:
type_el = None
if type_el is not None:
# If this is a native type, the input type is problematic for BOSS
if isNative(type_el):
is_problematic = True
return is_problematic
return is_problematic
# ====== END: isProblematicType ========
# ====== addParentClasses ========
# Adds parent classes to cfg.load_classes.
def addParentClasses():
import modules.classutils as classutils
for xml_file in gb.all_id_dict.keys():
# If new xml file, initialise global dicts
if xml_file != gb.xml_file_name:
gb.xml_file_name = xml_file
initGlobalXMLdicts(xml_file, id_and_name_only=True)
# Loop over all named elements in the xml file
for full_name, el in gb.name_dict.items():
if el.tag in ['Class', 'Struct']:
if isLoadedClass(el):
parents_el_list = getAllParentClasses(el, only_native_classes=True)
for parent_el in parents_el_list:
# Skip classes that are not loadable (incomplete, abstract, ...)
if not isLoadable(el, print_warning=True):
continue
class_name = classutils.getClassNameDict(parent_el)
# - Update cfg.load_classes
if class_name['long_templ'] not in cfg.load_classes:
cfg.load_classes.append(class_name['long_templ'])
# ====== END: addParentClasses ========
# ====== fillParentsOfLoadedClassesList ========
# Adds parent classes to cfg.load_classes.
def fillParentsOfLoadedClassesList():
import modules.classutils as classutils
messages = []
for xml_file in gb.all_id_dict.keys():
# If new xml file, initialise global dicts
if xml_file != gb.xml_file_name:
gb.xml_file_name = xml_file
initGlobalXMLdicts(xml_file, id_and_name_only=True)
# Loop over all named elements in the xml file
for full_name, el in gb.name_dict.items():
if el.tag in ['Class', 'Struct']:
if isLoadedClass(el):
parents_el_list = getAllParentClasses(el, only_native_classes=True)
for parent_el in parents_el_list:
# Skip classes that are not loadable (incomplete, abstract, ...)
if not isLoadable(parent_el, print_warning=True, check_pure_virtual_members=False):
continue
class_name = classutils.getClassNameDict(parent_el)
# Append to gb.parents_of_loaded_classes
if class_name['long_templ'] not in gb.parents_of_loaded_classes:
gb.parents_of_loaded_classes.append(class_name['long_templ'])
# Print info
msg = ' - %s is parent of %s.' % (class_name['long_templ'], full_name)
if msg not in messages:
print(msg)
messages.append(msg)
# ====== END: fillParentsOfLoadedClassesList ========
# ====== xmlFilesToDicts ========
#
# Read all xml elements of all files and store in two dict of dicts:
#
# 1. all_id_dict: file name --> xml id --> xml element
# 2. all_name_dict: file name --> name --> xml element
#
def xmlFilesToDicts(xml_files):
for xml_file in xml_files:
gb.all_id_dict[xml_file] = OrderedDict()
gb.all_name_dict[xml_file] = OrderedDict()
tree = ET.parse(xml_file)
root = tree.getroot()
for el in list(root):
# Fill id-based dict
gb.all_id_dict[xml_file][el.get('id')] = el
for el in list(root):
# Determine name
if 'name' in el.keys():
namespaces_list = getNamespaces(el, include_self=True, xml_file_name=xml_file)
full_name = '::'.join(namespaces_list)
else:
# Skip elements that don't have a name
continue
# Fill name-based dict
gb.all_name_dict[xml_file][full_name] = el
# ====== END: xmlFilesToDicts ========
# ====== clearGlobalXMLdicts ========
def clearGlobalXMLdicts():
# Clear a bunch of global dicts
gb.id_dict.clear()
gb.name_dict.clear()
gb.file_dict.clear()
gb.std_types_dict.clear()
gb.typedef_dict.clear()
gb.loaded_classes_in_xml.clear()
gb.func_dict.clear()
# ====== END: clearGlobalXMLdicts ========
# ====== initGlobalXMLdicts ========
def initGlobalXMLdicts(xml_path, id_and_name_only=False):
import modules.classutils as classutils
import modules.funcutils as funcutils
# Clear dicts
clearGlobalXMLdicts()
# Set some global dicts directly
gb.id_dict = copy.deepcopy( gb.all_id_dict[xml_path] )
gb.name_dict = copy.deepcopy( gb.all_name_dict[xml_path] )
# Stop here?
if id_and_name_only:
return
#
# Loop over all elements in this xml file
# to fill the remaining dicts. (The order is important!)
#
for xml_id, el in gb.id_dict.items():
# Update global dict: file name --> file xml element
if el.tag == 'File':
gb.file_dict[el.get('name')] = el
# Update global dict: std type --> type xml element
if isStdType(el):
class_name = classutils.getClassNameDict(el)
gb.std_types_dict[class_name['long_templ']] = el
# Update global dict of loaded classes in this xml file: class name --> class xml element
if el.tag in ['Class', 'Struct']:
try:
class_name = classutils.getClassNameDict(el)
except KeyError:
continue
# Check if we have done this class already
if class_name in gb.classes_done:
infomsg.ClassAlreadyDone( class_name['long_templ'] ).printMessage()
continue
# Check that class is requested
if (class_name['long_templ'] in cfg.load_classes):
# Check that class is complete
if isComplete(el):
# Store class xml element
gb.loaded_classes_in_xml[class_name['long_templ']] = el
# Update global dict: typedef name --> typedef xml element
if el.tag == 'Typedef':
# Only accept native typedefs:
if isNative(el):
typedef_name = el.get('name')
type_dict = findType(el)
type_el = type_dict['el']
# If underlying type is a fundamental or standard type, accept it right away
if isFundamental(type_el) or isStdType(type_el):
gb.typedef_dict[typedef_name] = el
# If underlying type is a class/struct, check if it's acceptable
elif type_el.tag in ['Class', 'Struct']:
type_name = classutils.getClassNameDict(type_el)
if type_name['long_templ'] in cfg.load_classes:
gb.typedef_dict[typedef_name] = el
# If neither fundamental or class/struct, ignore it.
else:
pass
# Update global dict: function name --> function xml element
if el.tag == 'Function':
try:
func_name = funcutils.getFunctionNameDict(el)
except KeyError:
continue
# Check if we have done this function already
if func_name in gb.functions_done:
infomsg.FunctionAlreadyDone( func_name['long_templ_args'] ).printMessage()
continue
if func_name['long_templ_args'] in cfg.load_functions:
gb.func_dict[func_name['long_templ_args']] = el
# Add entries to global dict: new header files
if el in gb.loaded_classes_in_xml.values():
class_name = classutils.getClassNameDict(el)
class_name_short = class_name['short']
class_name_long = class_name['long']
if class_name_long not in gb.new_header_files.keys():
abstract_header_name = gb.abstr_header_prefix + class_name_short + cfg.header_extension
wrapper_header_name = gb.wrapper_header_prefix + class_name_short + cfg.header_extension
wrapper_decl_header_name = gb.wrapper_header_prefix + class_name_short + '_decl' + cfg.header_extension
wrapper_def_header_name = gb.wrapper_header_prefix + class_name_short + '_def' + cfg.header_extension
abstract_header_fullpath = os.path.join(gb.backend_types_basedir, gb.gambit_backend_name_full, gb.abstr_header_prefix + class_name_short + cfg.header_extension )
wrapper_header_fullpath = os.path.join(gb.backend_types_basedir, gb.gambit_backend_name_full, gb.wrapper_header_prefix + class_name_short + cfg.header_extension )
wrapper_decl_header_fullpath = os.path.join(gb.backend_types_basedir, gb.gambit_backend_name_full, gb.wrapper_header_prefix + class_name_short + '_decl' + cfg.header_extension )
wrapper_def_header_fullpath = os.path.join(gb.backend_types_basedir, gb.gambit_backend_name_full, gb.wrapper_header_prefix + class_name_short + '_def' + cfg.header_extension )
gb.new_header_files[class_name_long] = { 'abstract': abstract_header_name,
'wrapper': wrapper_header_name,
'wrapper_decl': wrapper_decl_header_name,
'wrapper_def': wrapper_def_header_name,
'abstract_fullpath': abstract_header_fullpath,
'wrapper_fullpath': wrapper_header_fullpath,
'wrapper_decl_fullpath': wrapper_decl_header_fullpath,
'wrapper_def_fullpath': wrapper_def_header_fullpath }
#
# END: Loop over all elements in this xml file
#
# ====== END: initGlobalXMLdicts ========
# ====== identifyStdIncludePaths ========
def identifyStdIncludePaths():
# Shell command: Pipe an include statement to the compiler and use
# verbose mode to print the header search paths.
command = 'echo "#include <iostream>" | ' + cfg.castxml_cc + ' -v -x c++ -c -'
# Run command
print(' Running command: ' + command)
did_fail = False
error_message = ''
output_tmpfile = tempfile.TemporaryFile()
try:
p = subprocess.Popen(shlex.split(command), stdout=output_tmpfile, stderr=output_tmpfile)
p.wait()
except subprocess.CalledProcessError as e:
did_fail = True
error_message = e.message
# Reset environment variables
if 'gnu' in cfg.castxml_cc_id:
for var_name, value in temp_env_vars.items():
os.environ[var_name] = value
# Get output from tempfile
output_tmpfile.seek(0)
output = output_tmpfile.read()
output_tmpfile.close()
# Any error that did not result in a CalledProcessError?
if p.returncode != 0:
did_fail = True
if did_fail:
print(' ' + modifyText('ERROR: Shell command failed.','red'))
# Get error message
print()
print(modifyText('START SHELL COMMAND OUTPUT','underline'))
print()
print(output)
print(modifyText('END SHELL COMMAND OUTPUT','red'))
print()
if error_message != '':
print("CalledProcessError.message:", error_message)
print()
raise Exception('Shell command failed')
else:
print(' ' + modifyText('Command finished successfully.','green'))
print()
std_include_paths = []
output_lines = output.split('\n')
try:
start_i = output_lines.index("#include <...> search starts here:")
end_i = output_lines.index("End of search list.")
except ValueError:
print(' ' + modifyText('WARNING: Could not identify standard include paths.\n Add them manually in the config file if necessary.','yellow'))
print()
else:
for line in output_lines[start_i+1:end_i]:
std_include_paths.append( line.strip().split()[0] )
# Filter out Intel-specific paths to avoid conflict with gnu headers
if (cfg.castxml_cc_id == 'gnu') or (cfg.castxml_cc_id == 'gnu-c'):
len_before_filter = len(std_include_paths)
std_include_paths = [path for path in std_include_paths if 'intel' not in path]
len_after_filter = len(std_include_paths)
if len_after_filter < len_before_filter:
print(' (Filtered out Intel paths to avoid conflicts with gcc headers.)')
print()
print(' Identified %i standard include paths:' % len(std_include_paths))
for path in std_include_paths:
print(' - ' + path)
print()
# Set global list
gb.std_include_paths = std_include_paths
# ====== END: identifyStdIncludePaths ========
# ====== isInList ========
def isInList(search_entry, search_list, return_index=True, ignore_whitespace=True):
# In case search_list is passed in as a dict_keys object (Python3)
# instead of as a regular list, convert it to a list
search_list = list(search_list)
# Search for entry
try:
i = search_list.index(search_entry)
if return_index:
return True, i
else:
return True
except ValueError:
pass
# Search for entry after removing all whitespace
if ignore_whitespace:
search_entry_no_ws = "".join(search_entry.split())
search_list_no_ws = ["".join(e.split()) for e in search_list]
try:
i = search_list_no_ws.index(search_entry_no_ws)
if return_index:
return True, i
else:
return True
except ValueError:
pass
# Entry not found
if return_index:
return False, -1
else:
return False
# ====== END: isInList ========
# ====== modifyText ========
def modifyText(msg, mod):
if mod not in gb.textmods.keys():
return msg
msg = gb.textmods[mod] + msg + gb.textmods['end']
return msg
# ====== END: modifyText ========
# ====== orderIncludeStatements ========
def orderIncludeStatements(include_statements):
ordered_include_statements = []
# This is not the fastest solution, but an easy way to
# to keep the existing order within each group of headers
# Add standard headers (not Boost headers)
for s in include_statements:
if "<" in s:
if "<boost/" not in s:
ordered_include_statements.append(s)
# Add BOSS-generated and/or backend-specific headers
for s in include_statements:
if "<" not in s:
ordered_include_statements.append(s)
# Add Boost headers
for s in include_statements:
if "<boost/" in s:
ordered_include_statements.append(s)
# Check that we haven't missed any include statements
assert len(ordered_include_statements) == len(include_statements)
# Return ordered list of include statements
return ordered_include_statements
# ====== END: orderIncludeStatements ========
Updated on 2022-08-03 at 12:58:09 +0000