Re: BackgroundPsql swallowing errors on windows
От | Andres Freund |
---|---|
Тема | Re: BackgroundPsql swallowing errors on windows |
Дата | |
Msg-id | 2mxs2zbon5orhp4mfu3yli6adndrhnf72otbvdgffimzo3ue2c@ucr2zcdqcvkx обсуждение исходный текст |
Ответ на | Re: BackgroundPsql swallowing errors on windows (Tom Lane <tgl@sss.pgh.pa.us>) |
Список | pgsql-hackers |
Hi, On 2025-02-14 12:14:55 -0500, Tom Lane wrote: > Andres Freund <andres@anarazel.de> writes: > > On 2025-02-14 08:14:45 -0500, Andrew Dunstan wrote: > >> ... 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. > > +1. Would there be a chance of removing our use of IPC::Run entirely? > There'd be a lot to like about that. Unfortunately I doubt it'd get us that close to that. We do have several tests that intentionally involve psql, e.g. 010_tab_completion.pl, 001_password.pl. Presumably those would continue to use something like BackgroundPsql.pm, even if we had a sane way to have longrunning connections in tap tests. There also are a bunch of things we use IPC::Run to run synchronously, we probably could replace most of those without too much pain. The hardest probably would be timeout handling. Greetings, Andres Freund
В списке pgsql-hackers по дате отправления: