Re: [dpdk-dev] [PATCH] log:Change magic number on RTE_LOG_LEVEL to a define

["Wiles, Keith" <keith.wiles@intel.com>]

On 8/2/15, 3:44 PM, "Thomas Monjalon" <thomas.monjalon@6wind.com> wrote:

>2015-08-02 19:10, Wiles, Keith:
>> On 8/2/15, 12:15 PM, "Thomas Monjalon" <thomas.monjalon@6wind.com>
>>wrote:
>> >2015-06-06 19:04, Keith Wiles:
>> >> +# Log level use: RTE_LOG_XXX
>> >> +#   XXX = NOOP, EMERG, ALERT, CRIT, ERR, WARNING, NOTICE, INFO or
>>DEBUG
>> >> +#   Look in rte_log.h for others if any.
>> >> +#
>> >
>> >I think this comment is useless.
>> 
>> I do not think the comment is useless as some may not understand what
>> values the Log level can be set too in the future. Not commenting the
>> change would be a problem IMO. This is also why the line was moved.
>
>It is already documented in the API doc.
>I agree having some comments in the config files would be convenient but:
>	- this one is 3 lines long
>	- currently comments are only used to separate sections
>Maybe you can do a oneline:
>	# Minimum log level compiled: DEBUG, INFO, NOTICE, WARNING or ERR
>I think it is important to tell it is a minimum log level, i.e. compiled
>logs.
>And probably it is not needed to suggest a minimum level higher than ERR.

I will reduce the comment to one line, should I move the LOG_HISTORY down
to under LOG_LEVEL?
>
>> >> +enum {
>> >> +	RTE_LOG_NOOP = 0,   /**< Noop not used (zero entry)        */
>> >
>> >NOOP is useless: EMERG may be = 1
>> 
>> Does it really matter if I used RTE_LOG_NOOP, just to make sure someone
>> did not try to use zero here. Instead of setting the RTE_LOG_EMERG=1, I
>> can change it to be the way you suggest, but I think it does not hurt
>> anything does it?
>
>We avoid adding code without real use which could cause confusion.

I will remove NOOP and set EMERG=1.
>
>> >> +	RTE_LOG_EMERG,      /**< System is unusable.               */
>> >> +	RTE_LOG_ALERT,      /**< Action must be taken immediately. */
>> >> +	RTE_LOG_CRIT,       /**< Critical conditions.              */
>> >> +	RTE_LOG_ERR,        /**< Error conditions.                 */
>> >> +	RTE_LOG_WARNING,    /**< Warning conditions.               */
>> >> +	RTE_LOG_NOTICE,     /**< Normal but significant condition. */
>> >> +	RTE_LOG_INFO,       /**< Informational.                    */
>> >> +	RTE_LOG_DEBUG       /**< Debug-level messages.             */
>> >> +};
>> >
>> >What is the benefit of this change?
>> 
>> The change is to use a enum in place of using magic numbers, plus you
>>get
>> the benefit of seeing the enum name in the debugger instead of a number.
>> It makes the code more readable IMHO.
>
>OK so a comment in the commit message could give the debugger
>justification.

OK will add the debugger comment to the commit log.
>
>


‹ 
Regards,
++Keith
Intel Corporation