Re: [Ryu-devel] [PATCH] ofproto: change the way __str__ is implemented for of-wire messages

2013-05-09 Thread Isaku Yamahata
On Fri, May 10, 2013 at 03:01:13PM +0900, YAMAMOTO Takashi wrote: > instead of explicit listing of of-wire attributes, use a heuristics > to exclude internal attributes. (eg. buf, datapath, etc) > > this commit changes __str__ outputs. update a test case accordingly. > > Signed-off-by: YAMAMOTO

[Ryu-devel] [PATCH] ofproto: change the way __str__ is implemented for of-wire messages

2013-05-09 Thread YAMAMOTO Takashi
instead of explicit listing of of-wire attributes, use a heuristics to exclude internal attributes. (eg. buf, datapath, etc) this commit changes __str__ outputs. update a test case accordingly. Signed-off-by: YAMAMOTO Takashi --- ryu/ofproto/ofproto_parser.py | 27 +

Re: [Ryu-devel] [PATCH 3/3] msgpack-rpc: unit tests

2013-05-09 Thread YAMAMOTO Takashi
> On Fri, May 10, 2013 at 10:17:28AM +0900, YAMAMOTO Takashi wrote: >> > Hi. Basically tests looks good. >> > nose.tools.assert_xxx should be used instead of assert. >> >> why? > > For consistency. It seems other tests use nose.tools.*. then you prefer eq_ than assert_xxx? > Since this test con

Re: [Ryu-devel] [PATCH 3/3] msgpack-rpc: unit tests

2013-05-09 Thread Isaku Yamahata
On Fri, May 10, 2013 at 10:17:28AM +0900, YAMAMOTO Takashi wrote: > > Hi. Basically tests looks good. > > nose.tools.assert_xxx should be used instead of assert. > > why? For consistency. It seems other tests use nose.tools.*. Since this test consistently uses assert, I don't claim so strongly.

Re: [Ryu-devel] [PATCH] Support bind IPv6 addresses in StreamServer

2013-05-09 Thread Can Zhang
That's apparently OK. 2013/5/10 FUJITA Tomonori > On Fri, 3 May 2013 17:39:02 +0800 > Can Zhang wrote: > > > Use a very loose condition to determine if the address is an IPv6 > address or not. > > > > Signed-off-by: Can Zhang > > --- > > ryu/lib/hub.py | 8 +++- > > 1 file changed, 7 in

Re: [Ryu-devel] [PATCH 3/3] msgpack-rpc: unit tests

2013-05-09 Thread YAMAMOTO Takashi
> Hi. Basically tests looks good. > nose.tools.assert_xxx should be used instead of assert. why? YAMAMOTO Takashi > > thanks, -- Learn Graph Databases - Download FREE O'Reilly Book "Graph Databases" is the definitive n

Re: [Ryu-devel] [PATCH 2/3] msgpack-rpc fixes and improvements

2013-05-09 Thread YAMAMOTO Takashi
> On Fri, 10 May 2013 09:31:55 +0900 (JST) > yamam...@valinux.co.jp (YAMAMOTO Takashi) wrote: > +class Client(object): +"""a convenient class for a pure rpc client +*sock* is a socket-like. it should be blocking. +""" >>> >>> I'm not sure about having pure 'Client

Re: [Ryu-devel] [PATCH 2/3] msgpack-rpc fixes and improvements

2013-05-09 Thread FUJITA Tomonori
On Fri, 10 May 2013 09:31:55 +0900 (JST) yamam...@valinux.co.jp (YAMAMOTO Takashi) wrote: >>> +class Client(object): >>> +"""a convenient class for a pure rpc client >>> +*sock* is a socket-like. it should be blocking. >>> +""" >> >> I'm not sure about having pure 'Client' class. Bot

Re: [Ryu-devel] [PATCH 2/3] msgpack-rpc fixes and improvements

2013-05-09 Thread YAMAMOTO Takashi
>> +class Client(object): >> +"""a convenient class for a pure rpc client >> +*sock* is a socket-like. it should be blocking. >> +""" > > I'm not sure about having pure 'Client' class. Both end can be client > and server once a connection is established. i think a pure synchronous cl

Re: [Ryu-devel] Event Timing

2013-05-09 Thread FUJITA Tomonori
On Fri, 3 May 2013 10:35:12 -0400 Kevin Han wrote: > Are there any events in ryu which can be timed to run at certain intervals. > For example, would I be able to set up an event to time out cache entries > in my tables every second? I have a plan to add such service which sends notificaiton via

Re: [Ryu-devel] [PATCH 2/3] msgpack-rpc fixes and improvements

2013-05-09 Thread FUJITA Tomonori
On Wed, 8 May 2013 15:32:36 +0900 YAMAMOTO Takashi wrote: > - make create_request returns msgid > - fix msgid wrap around > - rename classes > - convenient transport classes for socket-like > - update requirements > - copyright notice, comments, and assertions > > Signed-off-by: YAMAMOTO Takash

Re: [Ryu-devel] using StreamServer defined in hub.py to listen ipv6 protocol

2013-05-09 Thread FUJITA Tomonori
On Fri, 3 May 2013 22:26:26 +0800 jian lee wrote: > Hi,I used the StreamServer you defined in hub.py to listen ipv6 > protocol,codes is bellow: > > from ryu.lib.hub import StreamServer > > listen_info = (('::', 179), 10, 50) > server = StreamServer(listen_inf

Re: [Ryu-devel] [PATCH] Support bind IPv6 addresses in StreamServer

2013-05-09 Thread FUJITA Tomonori
On Fri, 3 May 2013 17:39:02 +0800 Can Zhang wrote: > Use a very loose condition to determine if the address is an IPv6 address or > not. > > Signed-off-by: Can Zhang > --- > ryu/lib/hub.py | 8 +++- > 1 file changed, 7 insertions(+), 1 deletion(-) Thanks, due to other patches, I updated

Re: [Ryu-devel] [PATCH] add eventlet to test requirement

2013-05-09 Thread FUJITA Tomonori
On Wed, 8 May 2013 06:56:36 +0900 (JST) yamam...@valinux.co.jp (YAMAMOTO Takashi) wrote: > hi, > > > Signed-off-by: FUJITA Tomonori > > --- > > tools/test-requires | 1 + > > 1 file changed, 1 insertion(+) > > > > diff --git a/tools/test-requires b/tools/test-requires > > index 6f21bff..ebbb1

Re: [Ryu-devel] [PATCH 0/4] hub related patches

2013-05-09 Thread FUJITA Tomonori
On Wed, 8 May 2013 15:24:41 +0900 YAMAMOTO Takashi wrote: > YAMAMOTO Takashi (4): > hub: comment > hub: handle GreenletExit internally > hub: add some more tests > controller: stop catching GreenletExit > > ryu/controller/controller.py | 6 -- > ryu/lib/hub.py |

Re: [Ryu-devel] [PATCH 0/3] NXAction clean ups

2013-05-09 Thread FUJITA Tomonori
On Thu, 9 May 2013 16:20:30 +0900 Isaku Yamahata wrote: > It is reported that ofctl_v1_0.actions_to_str depends on that action has > cls_action_type attribute. > This isn't the case for NXActions. > Make it correct and followed by some realted clean up. > > Isaku Yamahata (3): > ofproto_v1_0_

Re: [Ryu-devel] [PATCH 3/3] msgpack-rpc: unit tests

2013-05-09 Thread Isaku Yamahata
Hi. Basically tests looks good. nose.tools.assert_xxx should be used instead of assert. thanks, On Wed, May 08, 2013 at 03:32:37PM +0900, YAMAMOTO Takashi wrote: > > Signed-off-by: YAMAMOTO Takashi > --- > ryu/tests/unit/lib/test_rpc.py | 226 > + > 1 f

[Ryu-devel] [PATCH 0/3] NXAction clean ups

2013-05-09 Thread Isaku Yamahata
It is reported that ofctl_v1_0.actions_to_str depends on that action has cls_action_type attribute. This isn't the case for NXActions. Make it correct and followed by some realted clean up. Isaku Yamahata (3): ofproto_v1_0_parser: NXActionHeader should be subclass of OFPActionVendor ofprot

[Ryu-devel] [PATCH 3/3] ofproto v1.0: simplify NXAction initialization

2013-05-09 Thread Isaku Yamahata
Signed-off-by: Isaku Yamahata --- ryu/ofproto/ofproto_v1_0_parser.py | 144 1 file changed, 64 insertions(+), 80 deletions(-) diff --git a/ryu/ofproto/ofproto_v1_0_parser.py b/ryu/ofproto/ofproto_v1_0_parser.py index 95c60b7..7081c31 100644 --- a/ryu/ofprot

[Ryu-devel] [PATCH 2/3] ofproto_v1_0_parser: simplify OFPActionVendor

2013-05-09 Thread Isaku Yamahata
vendor value is available as cls_vendor. Signed-off-by: Isaku Yamahata --- ryu/ofproto/ofproto_v1_0_parser.py |6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/ryu/ofproto/ofproto_v1_0_parser.py b/ryu/ofproto/ofproto_v1_0_parser.py index 96781ff..95c60b7 100644 --- a/

[Ryu-devel] [PATCH 1/3] ofproto_v1_0_parser: NXActionHeader should be subclass of OFPActionVendor

2013-05-09 Thread Isaku Yamahata
Because NXActionHeader is extention to OFPActionVendor. Signed-off-by: Isaku Yamahata --- ryu/ofproto/ofproto_v1_0_parser.py |5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/ryu/ofproto/ofproto_v1_0_parser.py b/ryu/ofproto/ofproto_v1_0_parser.py index aa0eaea..96781ff