Re: GUC time unit spelling a bit inconsistent

Heikki Linnakangas <heikki@enterprisedb.com>

From: Heikki Linnakangas <heikki@enterprisedb.com>
To: Dave Page <dpage@postgresql.org>
Cc: Michael Paesold <mpaesold@gmx.at>, Bruce Momjian <bruce@momjian.us>, Simon Riggs <simon@2ndquadrant.com>, Peter Eisentraut <peter_e@gmx.net>, Tom Lane <tgl@sss.pgh.pa.us>, pgsql-hackers@postgresql.org
Date: 2007-06-21T09:23:08Z
Lists: pgsql-hackers
Dave Page wrote:
> Michael Paesold wrote:
>> It's not about a certain standard. There are so many different ways in 
>> the world to write time units, so in a certain context a standard is 
>> really useful to constrain the format/syntax, but...
>>
>> This all was about usability of a configuration file, wasn't it? Now, 
>> Peter, you improved that very much with this change. But do you at the 
>> same time want to cripple the usefulness again by insisting on a 
>> certain _syntax_, while the _semantics_ are completely clear to 
>> (guessing) 99% of the people who will changes these settings?
> 
> FWIW, I agree entirely.

My 2c on this:

The way I was taught in school is that "min" is for minute and "mon" is 
for month. Specifically, not "m".

I just had to download ISO 8601 and it seems irrelevant here. It talks 
about using certain characters *in place* of digits, like "hh:mm", and 
it talks about time periods, but that syntax is completely different, 
like P1H2M, meaning 5 hour and 2 minutes.

A HINT listing the valid units is a reasonable compromise.

-- 
   Heikki Linnakangas
   EnterpriseDB   http://www.enterprisedb.com