On 2025-02-14 Fr 11:54 AM, Andres Freund wrote:
> Hi,
>
> On 2025-02-14 08:14:45 -0500, Andrew Dunstan wrote:
>> It's been discussed before, but I'd really really like to get rid of
>> BackgroundPsql. It's ugly, non-intuitive and fragile.
> I agree, unfortunately we're stuck with this until we have a better
> alternative in tree :(
>
>
>> Last year I did some work on this. I was going to present it at Athens but
>> illness prevented me, and then other life events managed to get in my way.
>> But the basic work is around. See
<https://github.com/adunstan/test-pq/commit/98518e4929e80fb96f210bbc5aab9fdcea058512>
>> This introduces a libpq session object (PostgreSQL::Test::Session) which can
>> be backed either by FFI or a small XS wrapper - the commit has recipes for
>> both. Missing is a meson.build file for the XS wrapper. There are
>> significant performance gains to be had too (poll_query_until is much nicer,
>> for example, as are most uses of safe_psql). If there is interest I will
>> bring the work up to date, and maybe we can introduce this early in the v19
>> cycle. It would significantly reduce our dependency on IPC::Run, especially
>> the pump() stuff.
> I definitely am interested.
Here we are almost exactly a year later. I returned to this work quite
recently, and reached a milestone, namely the removal of all calls to
BackgroundPsql, with all the TAP tests passing. The XS module still has
some problems, and I think I'm inclined not to pursue it, and just rely
on the FFI mapping.
The current state of this is attached. People who are interested can hit
me up for info in Vancouver if not before. I'll work on turning this
into a set of commitable patches.
cheers
andrew
--
Andrew Dunstan
EDB: https://www.enterprisedb.com
