Re: [PATCH] Limit log field width

From: Alex Rousskov <rousskov_at_measurement-factory.com>
Date: Wed, 05 Oct 2011 11:13:02 -0600

On 10/05/2011 10:35 AM, Tsantilas Christos wrote:
> On 09/06/2011 09:45 PM, Alex Rousskov wrote:
>> On 04/16/2011 11:28 PM, Amos Jeffries wrote:
>>> Clearing up my queue of waiting patches this old one pops up as approved
>>> but not merged.
>>>
>>> Alex, it is waiting on your commit and resolution of a TODO (separate
>>> patch was fine). If it is still going to happen you my want to do it
>>> soon before any more polishing changes alter the logging area.
>>
>> Done as trunk r11709. A separate polishing patch to remove no longer
>> needed type castings will follow.
>
> And this is the policing patch, which
> - converts type of the Token::[width|precision] members
> from "unsigned int" to "int"
> - renames the Token::[width|precision] members to
> Token::[widthMin/widthMax]
> - removes unneeded typecastings

Looks good to me.

> case LFT_TIME_SUBSECOND:
> divisor = 1000;
>
> - if (precision) {
> + if (widthMax > 0) {
> int i;
> divisor = 1000000;
>
> - for (i = precision; i > 1; i--)
> + for (i = widthMax; i > 1; i--)
> divisor /= 10;
>
> if (!divisor)
> divisor = 0;
> }
> break;

> ts Seconds since epoch
> tu subsecond time (milliseconds)

I do not think there is a problem with the above code, but it looks like
we could merge %ts.%tu into something like %.3t. That would be a minor
usability improvement and out of scope for this patch, of course.

Thank you,

Alex.
Received on Wed Oct 05 2011 - 17:13:14 MDT

This archive was generated by hypermail 2.2.0 : Thu Oct 06 2011 - 12:00:04 MDT