Some notes/comment upgrades on split_args.

lib/ansible/module_utils/basic.py

@@ -170,39 +170,65 @@ def split_args(args):
     '''
     Splits args on whitespace, but intelligently reassembles
     those that may have been split over a jinja2 block or quotes.
-    When used in a module, we won't ever have to be concerned about
+
+    When used in a remote module, we won't ever have to be concerned about
     jinja2 blocks, however this function is/will be used in the
     core portions as well before the args are templated.
+
+    example input: a=b c=d
+    example output: dict(a='b', c='d')
+
+    Basically this is a variation shlex that has some more intelligence for
+    how Ansible needs to use it.
     '''
 
+    # FIXME: refactoring into smaller functions
+
     # the list of params parsed out of the arg string
+    # this is going to be the result value when we are donei
     params = []
+
     # here we encode the args, so we have a uniform charset to
     # work with, and split on white space
     args = args.encode('utf-8')
     items = args.split()
 
     # iterate over the items, and reassemble any that may have been
-    # split on a space inside a jinja2 block. These variables are used
+    # split on a space inside a jinja2 block. 
+    # ex if tokens are "{{", "foo", "}}" these go together
+
+    # These variables are used
     # to keep track of the state of the parsing, since blocks and quotes
     # may be nested within each other.
+
     inside_quotes = False
     quote_char = None
     split_print_depth = 0
     split_block_depth = 0
     split_comment_depth = 0
+
     # now we loop over each split item, coalescing items if the white space
     # split occurred within quotes or a jinja2 block of some kind
+
     for item in items:
+
         item = item.strip()
+
         # store the previous quoting state for checking later
         was_inside_quotes = inside_quotes
+
         # determine the current quoting state
-        for i in range(0, len(item)):
+        # the goal of this block is to determine if the quoted string
-            c = item[i]
+        # is unterminated in which case it needs to be put back together
-            bc = None
+
+        bc = None # before_char
+        for i in range(0, len(item)):  # use enumerate
+
+            c = item[i]  # current_char
+
             if i > 0:
                 bc = item[i-1]
+
             if c in ('"', "'"):
                 if inside_quotes:
                     if c == quote_char and bc != '\\':
@@ -211,29 +237,42 @@ def split_args(args):
                 else:
                     inside_quotes = True
                     quote_char = c
+
         # multiple conditions may append a token to the list of params,
         # so we keep track with this flag to make sure it only happens once
+        # append means add to the end of the list, don't append means concatenate
+        # it to the end of the last token
         appended = False
+
         # if we're inside quotes now, but weren't before, append the item
         # to the end of the list, since we'll tack on more to it later
+
         if inside_quotes and not was_inside_quotes:
             params.append(item)
             appended = True
+
         # otherwise, if we're inside any jinja2 block, inside quotes, or we were
         # inside quotes (but aren't now) concat this item to the last param
-        elif ((split_print_depth + split_block_depth + split_comment_depth) > 0 or inside_quotes or was_inside_quotes):
+        # FIXME: just or these all together
+        elif (split_print_depth or split_block_depth or split_comment_depth or inside_quotes or was_inside_quotes):
             params[-1] = "%s %s" % (params[-1], item)
             appended = True
+
         # these variables are used to determine the current depth of each jinja2
         # block type, by counting the number of openings and closing tags
+        # FIXME: assumes Jinja2 seperators aren't changeable (also true elsewhere in ansible ATM)
+
         num_print_open    = item.count('{{')
         num_print_close   = item.count('}}')
         num_block_open    = item.count('{%')
         num_block_close   = item.count('%}')
         num_comment_open  = item.count('{#')
         num_comment_close = item.count('#}')
-        # if the number is not the same, the depth has changed, so we calculate that here
+
+        # if the number of paired block tags is not the same, the depth has changed, so we calculate that here
         # and may append the current item to the params (if we haven't previously done so)
+
+        # FIXME: DRY a bit
         if num_print_open != num_print_close:
             split_print_depth += (num_print_open - num_print_close)
             if not appended:
@@ -241,6 +280,7 @@ def split_args(args):
                 appended = True
             if split_print_depth < 0:
                 split_print_depth = 0
+
         if num_block_open != num_block_close:
             split_block_depth += (num_block_open - num_block_close)
             if not appended:
@@ -248,6 +288,7 @@ def split_args(args):
                 appended = True
             if split_block_depth < 0:
                 split_block_depth = 0
+
         if num_comment_open != num_comment_close:
             split_comment_depth += (num_comment_open - num_comment_close)
             if not appended:
@@ -255,14 +296,19 @@ def split_args(args):
                 appended = True
             if split_comment_depth < 0:
                 split_comment_depth = 0
+
         # finally, if we're at zero depth for all blocks and not inside quotes, and have not
         # yet appended anything to the list of params, we do so now
-        if (split_print_depth + split_block_depth + split_comment_depth) == 0 and not inside_quotes and not appended:
+
+        if not (split_print_depth or split_block_depth or split_comment_depth) and not inside_quotes and not appended:
             params.append(item)
+
     # If we're done and things are not at zero depth or we're still inside quotes,
     # raise an error to indicate that the args were unbalanced
-    if (split_print_depth + split_block_depth + split_comment_depth) != 0 or inside_quotes:
+
+    if (split_print_depth or split_block_depth or split_comment_depth) or inside_quotes:
         raise Exception("error while splitting arguments, either an unbalanced jinja2 block or quotes")
+
     # finally, we decode each param back to the unicode it was in the arg string
     params = [x.decode('utf-8') for x in params]
     return params