Re: alpha5 - Mailing list pgsql-hackers

From Andrew Dunstan
Subject Re: alpha5
Date
Msg-id 4D909BC2.1070005@dunslane.net
Whole thread Raw
In response to Re: alpha5  (Robert Haas <robertmhaas@gmail.com>)
List pgsql-hackers

On 03/28/2011 09:25 AM, Robert Haas wrote:
> On Sat, Mar 26, 2011 at 9:41 PM, Robert Haas<robertmhaas@gmail.com>  wrote:
>> Per previous discussion, I'm going to wrap alpha5 Monday morning
>> Eastern time, barring objections.
> It seems that the 'make distcheck' build is broken.  Apparently we
> don't have any automated testing of this?  Anyone know what to fix
> here?
>
>

There is a bunch of things we don't test in a way that shows up 
obviously, if at all. I'm thinking of making a buildfarm enhancement 
that will allow performance of optional steps with some extra timing and 
branch params. Obvious candidates include building the docs, looking for 
typedefs, and benchmark runs as per Greg's post yesterday. The extra 
config would look something like:
    optional_steps => {         find_typedefs => { branches => ['HEAD'], min_hours_since => 23 },         benchmark =>
{dow=> [0,2], min_hour => 3, max_hour => 6 },         build_docs => {min_hours_since => 72},    },
 

Adding a docs build to the buildfarm would at least show up errors like 
the one you found pretty quickly. It's going to have to be optional, 
though, as by no means does everyone have the required toolset installed.

cheers

andrew


pgsql-hackers by date:

Previous
From: "Kevin Grittner"
Date:
Subject: Re: Set hint bits upon eviction from BufMgr
Next
From: Gurjeet Singh
Date:
Subject: Re: Triggers on system catalog