On Fri, 23 Aug 2013 17:59:21 +0900 (JST)
yamam...@valinux.co.jp (YAMAMOTO Takashi) wrote:
>> On Fri, 23 Aug 2013 13:42:54 +0900 (JST)
>> yamam...@valinux.co.jp (YAMAMOTO Takashi) wrote:
>>
>>>> On Fri, 23 Aug 2013 12:18:31 +0900 (JST)
>>>> yamam...@valinux.co.jp (YAMAMOTO Takashi) wrote:
>>>>
>>>>>> On Fri, 16 Aug 2013 15:29:22 +0900
>>>>>> YAMAMOTO Takashi <yamam...@valinux.co.jp> wrote:
>>>>>>
>>>>>>> Signed-off-by: YAMAMOTO Takashi <yamam...@valinux.co.jp>
>>>>>>> ---
>>>>>>> ryu/ofproto/ofproto_v1_0_parser.py | 6 +++++-
>>>>>>> 1 file changed, 5 insertions(+), 1 deletion(-)
>>>>>>>
>>>>>>> diff --git a/ryu/ofproto/ofproto_v1_0_parser.py
>>>>>>> b/ryu/ofproto/ofproto_v1_0_parser.py
>>>>>>> index 3fae86d..e1951c8 100644
>>>>>>> --- a/ryu/ofproto/ofproto_v1_0_parser.py
>>>>>>> +++ b/ryu/ofproto/ofproto_v1_0_parser.py
>>>>>>> @@ -18,6 +18,7 @@ import struct
>>>>>>> import binascii
>>>>>>>
>>>>>>> from ofproto_parser import StringifyMixin, MsgBase, msg_pack_into,
>>>>>>> msg_str_attr
>>>>>>> +from ryu.lib import addrconv
>>>>>>> from ryu.lib import mac
>>>>>>> from . import ofproto_parser
>>>>>>> from . import ofproto_v1_0
>>>>>>> @@ -86,6 +87,9 @@ class
>>>>>>> OFPPhyPort(ofproto_parser.namedtuple('OFPPhyPort', (
>>>>>>> def parser(cls, buf, offset):
>>>>>>> port = struct.unpack_from(ofproto_v1_0.OFP_PHY_PORT_PACK_STR,
>>>>>>> buf, offset)
>>>>>>> + i = cls._fields.index('hw_addr')
>>>>>>> + port = list(port)
>>>>>>> + port[i] = addrconv.mac.bin_to_text(port[i])
>>>>>>> return cls(*port)
>>>>>>
>>>>>> Can we use unicode type here? Then _encode_value() in stringfy.py just
>>>>>
>>>>> are you talking about "name", not "hw_addr"?
>>>>
>>>> hw_addr.
>>>
>>> then i'm not sure if i understand what you mean.
>>> do you mean to use different python types (str vs unicode) to
>>> tell the encoder do base64 or not? it can work for encoding
>>> but not for decoding.
>>
>> _JSON_FORMATER thing in this patchset describes the type of members in
>> a class. Then you can know how to decode.
>>
>> I thought that we could use unicode for 'string' type however looks
>> like it needs more changes than I want. Scratch that.
>
> IMO, if how to decode is specified _JSON_FORMATTER or _STRING,
> it's better to use the same mechanism to specify how to encode.
>
>>
>> I like to simply use 'type' attribute rather than allowing own
>> encoder/decoder. We use JSON so it's just about string or binary. I
>> like a simpler like the following. I don't like to use class attribute
>> though. I prefer to use the attribute of each member.
>
> do you like something like the following?
>
> python:
>
> class Binary(str):
> pass
>
> class Hoge(StringifyMixin):
> def __init__(self):
> self.field1 = Binary('foo')
> self.field2 = 'bar'
>
> json:
>
> { "Hoge": {
> "field1": { "Binary": "xxx base64 encoded 'foo' xxx" },
> "field2": "bar"
> }}
Not sure. For me, the above JSON looks unfamiliar. If we do the above,
I think that we should do all the field in the same way for
consistency (that is, adding 'u32', 'u64', etc description). And
surely I don't like such.
I'm not sure we would add other types except for 'string' but I'm
inclined to something like the following.
=
>From b5ddaabfbb4ad5500d70fdd350f616ae357058f4 Mon Sep 17 00:00:00 2001
From: FUJITA Tomonori <fujita.tomon...@lab.ntt.co.jp>
Date: Sat, 24 Aug 2013 08:54:16 +0900
Subject: [PATCH 3/6] stop be64 encode/decode for strings in JSON
representation
Our JSON handles binary and string in the same way (use be64
always). Had better to handle string in human readable way (string is
supposed to be utf-8 encoded/decoded by default in JSON).
Signed-off-by: FUJITA Tomonori <fujita.tomon...@lab.ntt.co.jp>
---
ryu/lib/stringify.py | 15 +++++++++++++--
1 file changed, 13 insertions(+), 2 deletions(-)
diff --git a/ryu/lib/stringify.py b/ryu/lib/stringify.py
index 8b755c5..050aa0a 100644
--- a/ryu/lib/stringify.py
+++ b/ryu/lib/stringify.py
@@ -79,9 +79,18 @@ class StringifyMixin(object):
return False
@classmethod
+ def _is_string_type(cls, k):
+ if hasattr(cls, '_TYPE'):
+ if 'string' in cls._TYPE and k in cls._TYPE['string']:
+ return True
+ return False
+
+ @classmethod
def _encode_value(cls, k, v, encode_string=base64.b64encode):
encode = lambda x: cls._encode_value(k, x, encode_string)
- if isinstance(v, (bytes, unicode)):
+ if cls._is_string_type(k):
+ json_value = v.encode('utf')
+ elif isinstance(v, (bytes, unicode)):
json_value = encode_string(v)
elif isinstance(v, list):
json_value = map(encode, v)
@@ -124,7 +133,9 @@ class StringifyMixin(object):
@classmethod
def _decode_value(cls, k, json_value, decode_string=base64.b64decode):
decode = lambda x: cls._decode_value(k, x, decode_string)
- if isinstance(json_value, (bytes, unicode)):
+ if cls._is_string_type(k):
+ v = json_value
+ elif isinstance(json_value, (bytes, unicode)):
v = decode_string(json_value)
elif isinstance(json_value, list):
v = map(decode, json_value)
--
1.7.12.4 (Apple Git-37)
------------------------------------------------------------------------------
Introducing Performance Central, a new site from SourceForge and
AppDynamics. Performance Central is your source for news, insights,
analysis and resources for efficient Application Performance Management.
Visit us today!
http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk
_______________________________________________
Ryu-devel mailing list
Ryu-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ryu-devel
