[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <c15b9c71-1c37-a82d-a167-b56eace6a3b0@nvidia.com>
Date: Thu, 3 Jun 2021 19:14:34 +0300
From: Roi Dayan <roid@...dia.com>
To: Stephen Hemminger <stephen@...workplumber.org>
CC: <netdev@...r.kernel.org>, Paul Blakey <paulb@...dia.com>,
David Ahern <dsahern@...il.com>
Subject: Re: [PATCH iproute2-next v2 1/1] police: Add support for json output
On 2021-06-03 6:49 PM, Stephen Hemminger wrote:
> On Thu, 3 Jun 2021 10:33:45 +0300
> Roi Dayan <roid@...dia.com> wrote:
>
>> if (tb[TCA_POLICE_TBF] == NULL) {
>> - fprintf(f, "[NULL police tbf]");
>> + print_string(PRINT_FP, NULL, "%s", "[NULL police tbf]");
>> return 0;
>> }
>> #ifndef STOOPID_8BYTE
>> if (RTA_PAYLOAD(tb[TCA_POLICE_TBF]) < sizeof(*p)) {
>> - fprintf(f, "[truncated police tbf]");
>> + print_string(PRINT_FP, NULL, "%s", "[truncated police tbf]");
>
> These are errors, and you should just print them to stderr.
> That way if program is using JSON they can see the output on stdout;
> and look for non-structured errors on stderr.
>
right. thanks. i'll fix it.
I think I looked in m_ct.c example for the print and there it's using
print_string(). so i'll do a commit also to fix it there.
other actions use fprintf() but they print to stdout and not stderr.
i guess those should be fixed as well.
any reason also why the original return is 0? i think it should be -1
also in that case.
Powered by blists - more mailing lists