mirror of
https://github.com/ansible-collections/community.general.git
synced 2024-09-14 20:13:21 +02:00
Some notes/comment upgrades on split_args.
This commit is contained in:
parent
43154e5101
commit
b5d64fdb36
1 changed files with 55 additions and 9 deletions
|
@ -222,39 +222,65 @@ def split_args(args):
|
||||||
'''
|
'''
|
||||||
Splits args on whitespace, but intelligently reassembles
|
Splits args on whitespace, but intelligently reassembles
|
||||||
those that may have been split over a jinja2 block or quotes.
|
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
|
jinja2 blocks, however this function is/will be used in the
|
||||||
core portions as well before the args are templated.
|
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
|
# the list of params parsed out of the arg string
|
||||||
|
# this is going to be the result value when we are donei
|
||||||
params = []
|
params = []
|
||||||
|
|
||||||
# here we encode the args, so we have a uniform charset to
|
# here we encode the args, so we have a uniform charset to
|
||||||
# work with, and split on white space
|
# work with, and split on white space
|
||||||
args = args.encode('utf-8')
|
args = args.encode('utf-8')
|
||||||
items = args.split()
|
items = args.split()
|
||||||
|
|
||||||
# iterate over the items, and reassemble any that may have been
|
# 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
|
# to keep track of the state of the parsing, since blocks and quotes
|
||||||
# may be nested within each other.
|
# may be nested within each other.
|
||||||
|
|
||||||
inside_quotes = False
|
inside_quotes = False
|
||||||
quote_char = None
|
quote_char = None
|
||||||
split_print_depth = 0
|
split_print_depth = 0
|
||||||
split_block_depth = 0
|
split_block_depth = 0
|
||||||
split_comment_depth = 0
|
split_comment_depth = 0
|
||||||
|
|
||||||
# now we loop over each split item, coalescing items if the white space
|
# now we loop over each split item, coalescing items if the white space
|
||||||
# split occurred within quotes or a jinja2 block of some kind
|
# split occurred within quotes or a jinja2 block of some kind
|
||||||
|
|
||||||
for item in items:
|
for item in items:
|
||||||
|
|
||||||
item = item.strip()
|
item = item.strip()
|
||||||
|
|
||||||
# store the previous quoting state for checking later
|
# store the previous quoting state for checking later
|
||||||
was_inside_quotes = inside_quotes
|
was_inside_quotes = inside_quotes
|
||||||
|
|
||||||
# determine the current quoting state
|
# 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:
|
if i > 0:
|
||||||
bc = item[i-1]
|
bc = item[i-1]
|
||||||
|
|
||||||
if c in ('"', "'"):
|
if c in ('"', "'"):
|
||||||
if inside_quotes:
|
if inside_quotes:
|
||||||
if c == quote_char and bc != '\\':
|
if c == quote_char and bc != '\\':
|
||||||
|
@ -263,29 +289,42 @@ def split_args(args):
|
||||||
else:
|
else:
|
||||||
inside_quotes = True
|
inside_quotes = True
|
||||||
quote_char = c
|
quote_char = c
|
||||||
|
|
||||||
# multiple conditions may append a token to the list of params,
|
# 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
|
# 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
|
appended = False
|
||||||
|
|
||||||
# if we're inside quotes now, but weren't before, append the item
|
# 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
|
# to the end of the list, since we'll tack on more to it later
|
||||||
|
|
||||||
if inside_quotes and not was_inside_quotes:
|
if inside_quotes and not was_inside_quotes:
|
||||||
params.append(item)
|
params.append(item)
|
||||||
appended = True
|
appended = True
|
||||||
|
|
||||||
# otherwise, if we're inside any jinja2 block, inside quotes, or we were
|
# 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
|
# 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)
|
params[-1] = "%s %s" % (params[-1], item)
|
||||||
appended = True
|
appended = True
|
||||||
|
|
||||||
# these variables are used to determine the current depth of each jinja2
|
# these variables are used to determine the current depth of each jinja2
|
||||||
# block type, by counting the number of openings and closing tags
|
# 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_open = item.count('{{')
|
||||||
num_print_close = item.count('}}')
|
num_print_close = item.count('}}')
|
||||||
num_block_open = item.count('{%')
|
num_block_open = item.count('{%')
|
||||||
num_block_close = item.count('%}')
|
num_block_close = item.count('%}')
|
||||||
num_comment_open = item.count('{#')
|
num_comment_open = item.count('{#')
|
||||||
num_comment_close = 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)
|
# 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:
|
if num_print_open != num_print_close:
|
||||||
split_print_depth += (num_print_open - num_print_close)
|
split_print_depth += (num_print_open - num_print_close)
|
||||||
if not appended:
|
if not appended:
|
||||||
|
@ -293,6 +332,7 @@ def split_args(args):
|
||||||
appended = True
|
appended = True
|
||||||
if split_print_depth < 0:
|
if split_print_depth < 0:
|
||||||
split_print_depth = 0
|
split_print_depth = 0
|
||||||
|
|
||||||
if num_block_open != num_block_close:
|
if num_block_open != num_block_close:
|
||||||
split_block_depth += (num_block_open - num_block_close)
|
split_block_depth += (num_block_open - num_block_close)
|
||||||
if not appended:
|
if not appended:
|
||||||
|
@ -300,6 +340,7 @@ def split_args(args):
|
||||||
appended = True
|
appended = True
|
||||||
if split_block_depth < 0:
|
if split_block_depth < 0:
|
||||||
split_block_depth = 0
|
split_block_depth = 0
|
||||||
|
|
||||||
if num_comment_open != num_comment_close:
|
if num_comment_open != num_comment_close:
|
||||||
split_comment_depth += (num_comment_open - num_comment_close)
|
split_comment_depth += (num_comment_open - num_comment_close)
|
||||||
if not appended:
|
if not appended:
|
||||||
|
@ -307,14 +348,19 @@ def split_args(args):
|
||||||
appended = True
|
appended = True
|
||||||
if split_comment_depth < 0:
|
if split_comment_depth < 0:
|
||||||
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
|
# 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
|
# 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)
|
params.append(item)
|
||||||
|
|
||||||
# If we're done and things are not at zero depth or we're still inside quotes,
|
# 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
|
# 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")
|
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
|
# finally, we decode each param back to the unicode it was in the arg string
|
||||||
params = [x.decode('utf-8') for x in params]
|
params = [x.decode('utf-8') for x in params]
|
||||||
return params
|
return params
|
||||||
|
|
Loading…
Reference in a new issue