Re: Change default of jit to off - Mailing list pgsql-hackers

From Tomas Vondra
Subject Re: Change default of jit to off
Date
Msg-id ebfe51f8-ac31-4714-97ab-2dcd70bffb16@vondra.me
Whole thread Raw
In response to Re: Change default of jit to off  (Tomas Vondra <tomas@vondra.me>)
List pgsql-hackers
On 4/2/26 21:46, Tomas Vondra wrote:
> 
> 
> On 4/2/26 15:10, Andrew Dunstan wrote:
>>
>> On 2026-04-02 Th 8:50 AM, Thomas Munro wrote:
>>> On Fri, Apr 3, 2026 at 12:52 AM Tomas Vondra <tomas@vondra.me> wrote:
>>>> Pushed, after rewording the commit message a little bit, and removing
>>>> the sentence in the docs saying
>>>>
>>>>      (before PostgreSQL 19 the default was on)
>>> How can we keep it enabled in the build farm?
>>>
>>>
>>
>> No easy way to do it globally - I guess we could default it to on in the
>> next client release. Individual owners can enable it in their buildfarm
>> configs, in the extra_config section.
>>
> 
> That's a good idea. Maybe we should ping the buildfarm owners (those
> that are already running --with-llvm) to add jit=on.
> 
> I'll do that on my two clang animals dikkop/jay, and maybe on some of
> the gcc ones too. But I'll wait a bit, I don't want to upset them right
> before feature freeze.
> 

FWIW I've now done this on my 5 animals - dikkop, turaco, widowbird, jay
and hippopotamus. None of them was even building with LLVM before. On
the last two it was a bit of a struggle, due to having old gcc and LLVM.

regards

-- 
Tomas Vondra




pgsql-hackers by date:

Previous
From: Andreas Karlsson
Date:
Subject: Re: Improve pgindent's formatting named fields in struct literals and varidic functions
Next
From: David Rowley
Date:
Subject: Re: Use stack allocated StringInfoDatas, where possible (round 2)