Re: [Xen-devel] [PATCH] perfc: Print a system time in a convenient format

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 11.09.18 at 11:24, <george.dunlap@xxxxxxxxxx> wrote:
> On 09/11/2018 09:50 AM, Andrii Anisov wrote:
>> Hello Jan,
>> 
>> 
>> On 11.09.18 11:27, Jan Beulich wrote:
>>> NAK, for two reasons: I'm not of the opinion that reading a 15 or more
>>> digit decimal number without any separators is any easier than the
>>> current format.
>> It's quite subjective. IMHO timestamps measured in ns easier to
>> understand in decimals rather than in separated 32-bit hex-es. No matter
>> how many digital number they have.
>> Even post processing of perfc output is easier in case of decimal
>> timestamps. You should not parse hexes and odd separators to calculate
>> the time elapsed between two samples.
> 
> I agree with this.  Printing in hexadecimal is useful if you need to see
> the bitwise structure of the content; that's of zero use when it comes
> to timestamps.  On the contrary, you almost always want to be able to do
> a subtraction in your head and get at least an order-of-magnitude
> difference between the times, something which is far too difficult in hex.

I disagree - it's a matter of who looks at those numbers. Personally
I'm quite fine doing order-of-magnitude hex math.

> The fact that we even have PRI_stime is evidence that the developer
> community as a whole agrees.
> 
>>>   And then if you/we were to change this, then please
>>> uniformly everywhere - there's even a second instance right in this
>>> same file you change, not to speak of similar ones elsewhere
> 
> Presumably the fact that we have PRI_stime defined means that *most*
> places already use this format, and the places with `:` in the middle
> are outliers.  It's perfectly acceptable to change them one by one;
> changing them all should only be a suggestion.

Well, what should I say - I disagree. I would agree if it would be a
whole lot of work, but there's not that many key handlers to begin
with, and far not all of them print a time stamp in their headlines.

Jan



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel