def test_api(self):
raw_text = 'abc- "d" ""ef"" g'
stream = StringStream(raw_text)
assert stream.index == 0
assert stream.len == len(raw_text)
buf = stream.read(1)
assert_match(buf, "a")
assert stream.index == 1
stream.seek(-1)
assert stream.index == 0
buf = stream.advance_past_chars(['"'])
buf = stream.advance_past_string_with_gdb_escapes()
assert_match(buf, "d")
buf = stream.advance_past_chars(['"'])
buf = stream.advance_past_chars(['"'])
buf = stream.advance_past_string_with_gdb_escapes()
assert_match(buf, "ef")
# read way past end to test it gracefully returns the
# remainder of the string without failing
buf = stream.read(50)
assert_match(buf, '" g')
def _parse_dict(stream: StringStream):
"""Parse dictionary, with optional starting character '{'
return (tuple):
Number of characters parsed from to_parse
Parsed dictionary
"""
obj: Dict[str, Union[str, list]] = {}
logger.debug("%s", fmt_green("parsing dict"))
while True:
c = stream.read(1)
if c in _WHITESPACE:
pass
elif c in ["{", ","]:
pass
elif c in ["}", ""]:
# end of object, exit loop
break
else:
stream.seek(-1)
key, val = _parse_key_val(stream)
if key in obj:
# This is a gdb bug. We should never get repeated keys in a dict!
# See https://sourceware.org/bugzilla/show_bug.cgi?id=22217
# and https://github.com/cs01/pygdbmi/issues/19
# Example:
# thread-ids={thread-id="1",thread-id="2"}
# Results in:
# thread-ids: {{'thread-id': ['1', '2']}}
# Rather than the lossy
# thread-ids: {'thread-id': 2} # '1' got overwritten!
if isinstance(obj[key], list):
obj[key].append(val) # type: ignore
else:
obj[key] = [obj[key], val]
else:
obj[key] = val
look_ahead_for_garbage = True
c = stream.read(1)
while look_ahead_for_garbage:
if c in ["}", ",", ""]:
look_ahead_for_garbage = False
else:
# got some garbage text, skip it. for example:
# name="gdb"gargage # skip over 'garbage'
# name="gdb"\n # skip over '\n'
logger.debug("skipping unexpected charcter: " + c)
c = stream.read(1)
stream.seek(-1)
logger.debug("parsed dict")
logger.debug("%s", fmt_green(obj))
return obj
def _parse_array(stream: StringStream):
"""Parse an array, stream should be passed the initial [
returns:
Parsed array
"""
logger.debug("parsing array")
arr = []
while True:
c = stream.read(1)
if c in _GDB_MI_VALUE_START_CHARS:
stream.seek(-1)
val = _parse_val(stream)
arr.append(val)
elif c in _WHITESPACE:
pass
elif c == ",":
pass
elif c == "]":
# Stop when this array has finished. Note
# that elements of this array can be also be arrays.
break
logger.debug("parsed array:")
logger.debug("%s", fmt_green(arr))
return arr
def _parse_val(stream: StringStream):
"""Parse value from string
returns:
Parsed value (either a string, array, or dict)
"""
logger.debug("parsing value")
while True:
c = stream.read(1)
if c == "{":
# Start object
val = _parse_dict(stream)
break
elif c == "[":
# Start of an array
val = _parse_array(stream)
break
elif c == '"':
# Start of a string
val = stream.advance_past_string_with_gdb_escapes()
break
elif _DEBUG:
raise ValueError("unexpected character: %s" % c)
else:
logger.warn(f'unexpected character: "{c}" ({ord(c)}). Continuing.')
val = "" # this will be overwritten if there are more characters to be read
logger.debug("parsed value:")
logger.debug("%s", fmt_green(val))
return val