Sync the fixes for bug #211949 from trunk.

svn path=/main/branches/2.1.2/; revision=9465

1 files changed, 78 insertions, 17 deletions

diff --git a/bin/filter-bash-environment.py b/bin/filter-bash-environment.py
index 691d406da..d3434ff0e 100755
--- a/bin/filter-bash-environment.py
+++ b/bin/filter-bash-environment.py
@@ -5,22 +5,62 @@
 
 import os, re, sys
 
-egrep_compat_map = {
-	"[:alnum:]" : r'\w',
-	"[:digit:]" : r'\d',
-	"[:space:]" : r'\s',
-}
-
 here_doc_re = re.compile(r'.*\s<<[-]?(\w+)$')
+func_start_re = re.compile(r'^[-\w]+\s*\(\)\s*$')
+func_end_re = re.compile(r'^\}$')
+
+var_assign_re = re.compile(r'(^|^declare\s+-\S+\s+|^declare\s+|^export\s+)([^=\s]+)=("|\')?.*$')
+close_quote_re = re.compile(r'(\\"|"|\')\s*$')
+readonly_re = re.compile(r'^declare\s+-(\S*)r(\S*)\s+')
 
-def compile_egrep_pattern(s):
-	for k, v in egrep_compat_map.iteritems():
-		s = s.replace(k, v)
-	return re.compile(s)
+def have_end_quote(quote, line):
+	"""
+	Check if the line has an end quote (useful for handling multi-line
+	quotes). This handles escaped double quotes that may occur at the
+	end of a line. The posix spec does not allow escaping of single
+	quotes inside of single quotes, so that case is not handled.
+	"""
+	close_quote_match = close_quote_re.search(line)
+	return close_quote_match is not None and \
+		close_quote_match.group(1) == quote
 
 def filter_bash_environment(pattern, file_in, file_out):
 	here_doc_delim = None
+	in_func = None
+	multi_line_quote = None
+	multi_line_quote_filter = None
 	for line in file_in:
+		if multi_line_quote is not None:
+			if not multi_line_quote_filter:
+				file_out.write(line)
+			if have_end_quote(multi_line_quote, line):
+				multi_line_quote = None
+				multi_line_quote_filter = None
+			continue
+		if here_doc_delim is None and in_func is None:
+			var_assign_match = var_assign_re.match(line)
+			if var_assign_match is not None:
+				quote = var_assign_match.group(3)
+				filter_this = pattern.match(var_assign_match.group(2)) \
+					is not None
+				if quote is not None and not have_end_quote(quote, line):
+					multi_line_quote = quote
+					multi_line_quote_filter = filter_this
+				if not filter_this:
+					readonly_match = readonly_re.match(line)
+					if readonly_match is not None:
+						declare_opts = ""
+						for i in (1, 2):
+							group = readonly_match.group(i)
+							if group is not None:
+								declare_opts += group
+						if declare_opts:
+							line = "declare -%s %s" % \
+								(declare_opts, line[readonly_match.end():])
+						else:
+							line = "declare " + line[readonly_match.end():]
+					file_out.write(line)
+				continue
 		if here_doc_delim is not None:
 			if here_doc_delim.match(line):
 				here_doc_delim = None
@@ -31,15 +71,29 @@ def filter_bash_environment(pattern, file_in, file_out):
 			here_doc_delim = re.compile("^%s$" % here_doc.group(1))
 			file_out.write(line)
 			continue
-		if pattern.match(line) is None:
+		# Note: here-documents are handled before functions since otherwise
+		# it would be possible for the content of a here-document to be
+		# mistaken as the end of a function.
+		if in_func:
+			if func_end_re.match(line) is not None:
+				in_func = None
+			file_out.write(line)
+			continue
+		in_func = func_start_re.match(line)
+		if in_func is not None:
 			file_out.write(line)
+			continue
+		# This line is not recognized as part of a variable assignment,
+		# function definition, or here document, so just allow it to
+		# pass through.
+		file_out.write(line)
 
 if __name__ == "__main__":
-	description = "Filter out any lines that match a given PATTERN " + \
-		"while leaving bash here-documents intact. The PATTERN should " + \
-		"use python regular expression syntax but [:space:] and " + \
-		"[:alnum:] character classes will be automatically translated " + \
-		"for compatibility with egrep syntax."
+	description = "Filter out variable assignments for varable " + \
+		"names matching a given PATTERN " + \
+		"while leaving bash function definitions and here-documents " + \
+		"intact. The PATTERN is a space separated list of variable names" + \
+		" and it supports python regular expression syntax."
 	usage = "usage: %s PATTERN" % os.path.basename(sys.argv[0])
 	from optparse import OptionParser
 	parser = OptionParser(description=description, usage=usage)
@@ -48,6 +102,13 @@ if __name__ == "__main__":
 		parser.error("Missing required PATTERN argument.")
 	file_in = sys.stdin
 	file_out = sys.stdout
+	var_pattern = args[0].split()
+
+	# Filter invalid variable names that are not supported by bash.
+	var_pattern.append(r'\d.*')
+	var_pattern.append(r'.*\W.*')
+
+	var_pattern = "^(%s)$" % "|".join(var_pattern)
 	filter_bash_environment(
-		compile_egrep_pattern(args[0]), file_in, file_out)
+		re.compile(var_pattern), file_in, file_out)
 	file_out.flush()