[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 1 Feb 2018 20:08:37 -0700
From: David Ahern <dsahern@...il.com>
To: Stephen Hemminger <stephen@...workplumber.org>
Cc: netdev@...r.kernel.org, Stephen Hemminger <sthemmin@...rosoft.com>
Subject: Re: [RFC iproute2-next 00/16] iproute printing enhancements
On 2/1/18 6:19 PM, Stephen Hemminger wrote:
> The first group of patches refactor the printout to breakup
> excessively long print_route function and the last few implement
> JSON and color output format for routes.
>
> The one thing that maybe controversial is changing the default
> json output format to be compact. The value of JSON is for programattic
> output therefore pretty-printing should be an option.
>
> One change I didn't make (but considered) is not calling fflush
> after every single entry. When dumping millions of routes, letting
> stdio buffer output would be a performance win.
>
> Stephen Hemminger (16):
> iproute: refactor printing flags
> iproute: make printing icmpv6 a function
> iproute: make printing IPv4 cache flags a function
> iproute: refactor cacheinfo printing
> iproute: refactor metrics print
> iproute: refactor printing flow info
> iproute2: refactor newdst, gateway and via printing
> iproute: refactor multipath print
> iproute: refactor printing of interface
> iproute: whitespace fixes
> iproute: don't do assignment in condition
> iproute: make flush a separate function
> iproute: implement JSON and color output
> json: make pretty printing optional
> man: add documentation for json and pretty flags
> json: fix newline at end of array
>
The first 12 look fine to me; familiar territory. Why not send those
outside of the json patches which require a closer look?
Powered by blists - more mailing lists