[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAD4GDZyx9eVcGorgOgdSinM_pGRNhqqLCNX-SAvdZAXqTCy+KQ@mail.gmail.com>
Date: Wed, 21 Feb 2024 11:40:52 +0000
From: Donald Hunter <donald.hunter@...il.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: Jiri Pirko <jiri@...nulli.us>, netdev@...r.kernel.org, pabeni@...hat.com,
davem@...emloft.net, edumazet@...gle.com, jacob.e.keller@...el.com,
swarupkotikalapudi@...il.com, sdf@...gle.com, lorenzo@...nel.org,
alessandromarcolini99@...il.com
Subject: Re: [patch net-next 03/13] tools: ynl: allow user to pass enum string
instead of scalar value
On Wed, 21 Feb 2024 at 01:59, Jakub Kicinski <kuba@...nel.org> wrote:
>
> Some of the sharp edges in Python YNL are because I very much
> anticipated the pyroute2 maintainer to do a proper implementation,
> and this tool was just a very crude PoC :D
Hah yeah, I looked at pyroute2 a while back and thought there was a
bit of an impedance mismatch between the dynamic schema driven
approach of ynl and the declarative / procedural code in iproute2. I
think a code generator would be the way to target iproute2.
https://github.com/svinota/pyroute2/blob/34d0768f89fd232126c49e2f7c94e6da6582795b/pyroute2/netlink/rtnl/rtmsg.py#L102-L139
I find ynl to be a very useful tool when writing and testing spec
files and have been happy to contribute to it for that purpose. I
think we have started to remove some of the sharp edges, but there is
more to do :-)
Powered by blists - more mailing lists