add TW's vhdl parser (not coupled yet)

src/vhdl_parser.py

+#!/usr/bin/python
+
+def remove_gaps(buf, delims, gap_chars, lower_strings=False):
+	da={}
+	for d in delims:
+		da[d]=False
+	prev_is_gap = False
+	buf2=""
+	lines=[]
+	for c in buf:
+		for d in delims:
+			if(c==d):
+				da[d]=not da[d]
+		
+		within_string = any(da.values()) and not (c in delims)
+			
+		if(not within_string):
+			if(c in gap_chars):
+				if(not prev_is_gap):
+					prev_is_gap = True
+					buf2+=" "
+			else:
+				prev_is_gap = False
+				buf2+=c
+				if(c==";" or c=="\n"):
+					lines.append(buf2);
+					buf2=""
+		else:
+			buf2+=c;
+			prev_is_gap = False
+
+	return lines
+
+from dep_solver import DepParser, DepRelation, DepFile
+		
+class VHDLParser(DepParser):
+	def parse(self, f_deps, filename):
+		f = open(filename,"r")
+		buf = ""
+		# stage 1: strip comments
+		for l in f:
+			ci = l.find("--")
+			if(ci==0):
+				continue
+				
+			while(ci>0): 
+				nquotes = l[:ci].count('"') # ignore comments in strings
+				if(nquotes % 2 == 0):
+					l=l[:ci-1];
+					break
+				ci= l.find("--", ci+1)
+			buf+=l
+			
+		# stage 2: remove spaces, crs, lfs, strip strings (we don't need them)
+		buf2=""
+		string_literal = char_literal = False
+		prev_is_gap = False
+		gap_chars = " \r\n\t"
+		lines=[]
+		
+		for c in buf:
+			if(c == '"' and not char_literal):
+				string_literal = not string_literal
+			if(c == "'" and not string_literal):
+				char_literal = not char_literal
+		
+			within_string = (string_literal or char_literal) and (c != '"') and (c != "'")
+			
+			if(not within_string):
+				if(c in gap_chars):
+					if(not prev_is_gap):
+						prev_is_gap = True
+						buf2+=" "
+				else:
+					prev_is_gap = False
+					buf2+=c.lower()
+					if(c==";"):
+						lines.append(buf2);
+						buf2=""
+			else:
+				prev_is_gap = False
+
+		import re
+
+
+		patterns = {
+			"use" : "^ *use *(\w+) *\. *(\w+) *. *\w+ *;",
+			"entity" : "^ *entity +(\w+) +is +(port|generic)",
+			"package" : "^ *package +(\w+) +is",
+			"arch_begin" : "^ *architecture +(\w+) +of +(\w+) +is +",
+			"arch_end" : "^ *end +(\w+) +;",
+			"instance" : "^ *(\w+) *\: *(\w+) *(port|generic) *map"
+		}
+		
+		compiled_patterns = map(lambda p: (p,re.compile(patterns[p])), patterns)
+
+		within_architecture = False
+			
+		for l in lines:
+			matches = filter(lambda(k,v): v!=None, map(lambda (k,v): (k,re.match(v,l)), compiled_patterns))
+			if(not len(matches)):
+				continue
+					
+			what,g=matches[0]
+			
+			if(what == "use"):
+				f_deps.add_relation(DepRelation(g.group(1)+"."+g.group(2), DepRelation.USE, DepRelation.PACKAGE))
+			if(what == "package"):
+				f_deps.add_relation(DepRelation(g.group(1), DepRelation.PROVIDE, DepRelation.PACKAGE))
+			elif(what == "entity"):
+				f_deps.add_relation(DepRelation(g.group(1), DepRelation.PROVIDE, DepRelation.ENTITY))
+			elif(what == "package"):
+				f_deps.add_relation(DepRelation(g.group(1), DepRelation.PROVIDE, DepRelation.PACKAGE))
+			elif(what == "arch_begin"):
+				arch_name = g.group(1)
+				within_architecture = True
+			elif(what == "arch_end" and within_architecture and g.group(1) == arch_name):
+				within_architecture = False
+			elif(what == "instance" and within_architecture):
+				f_deps.add_relation(DepRelation(g.group(1), DepRelation.USE, DepRelation.ENTITY))
+
+