Conversation

Notices

1. Oneesan succubus (lain@pleroma.soykaf.com@pleroma.soykaf.com)'s status on Tuesday, 25-Feb-2020 19:09:18 JST Oneesan succubus

   • Oneesan succubus
   Update to my recent article about performance issues with prepared queries in postgresql:
   
   It came to my attention that you can give the prepare: :unnamed option to your Ecto database configuration to achieve a similar effect on Postgres versions before 12. This works universally, but it only gives you about 80% of the performance of the force_custom_plan option in my testing, presumably because we lose the performance boost that skipping the parsing step for prepared queries gives us.
   • Oneesan succubus (lain@pleroma.soykaf.com@pleroma.soykaf.com)'s status on Tuesday, 25-Feb-2020 19:16:05 JST Oneesan succubus

     • Haelwenn /ɛlwən/ ?
     @lanodan working on the MR right now :)
   • Oneesan succubus (lain@pleroma.soykaf.com@pleroma.soykaf.com)'s status on Tuesday, 25-Feb-2020 21:10:11 JST Oneesan succubus

     • CrunkLord420
     @crunklord420 idk, did you report it?
   • Oneesan succubus (lain@pleroma.soykaf.com@pleroma.soykaf.com)'s status on Tuesday, 25-Feb-2020 21:11:14 JST Oneesan succubus

     • CrunkLord420
     @crunklord420 then i have no meaningful way of answering your question
   • Oneesan succubus (lain@pleroma.soykaf.com@pleroma.soykaf.com)'s status on Tuesday, 25-Feb-2020 21:18:12 JST Oneesan succubus

     • CrunkLord420
     @crunklord420 i mean, even just reporting your issues can make sense. sometimes people chime in with 'oh yeah i had the same, here's a fix i shared with nobody'.