Re: [PATCH] Cleanup of GUC units code

Simon Riggs <simon(at)2ndQuadrant(dot)com> writes:

> Peter's objection is reasonable, as far as most people have replied.
> Marko's proposal is also reasonable to most people, since they do not
> wish fat fingers to cause any amount of downtime. ISTM that if you've
> done this, you appreciate the feature, if not it seems less important.

My objection isn't down-time at all, it's the insultingly user-hostile
attitude. I normally am setting work_mem by hand in a psql session and *every*
time I do it I swear at Postgres for being annoyingly pedantic here.

I'm all for using the correct acronyms in all messages and documentation. What
I find annoying is the:

postgres=# set work_mem = '1g';
ERROR: invalid value for parameter "work_mem": "1g"
HINT: It's perfectly clear what you want but I'm going to refuse to do
it until you type it exactly as I say: "GB"

> * Marko should change patch to put WARNINGs in place so people know they
> got it wrong

That's only slightly less insulting than an error.

> * we make sure the case is always shown correctly in all other aspects
> of Postgres server and docs (no relaxation at all there)

I believe we're already fairly stringent about this as we should be.

> * in the longer term, we look for the solution to be a config checker

I don't think a config checker directly addresses the same problem. I never
set work_mem in a config and it still annoys the hell out of me.

--
Gregory Stark
EnterpriseDB http://www.enterprisedb.com
Ask me about EnterpriseDB's On-Demand Production Tuning