On Sun, Nov 1, 2026 at 9:56 AM shihao zhong <[email protected]> wrote: > +1 to commit this patch first.
I've pushed the patch. Thanks! > It does not cover one more case. If the recovery target is not at a > checkpoint redo, a timeline 1 summary can go past the switch point, > and a truncation after that point still drops blocks. This is the > case from the "Assertion failure in WAL summarizer" thread [1], and > Manu's 0002 there fixes it on top of this patch. > > [1] https://postgr.es/m/[email protected] Yes, I'm planning to work on that. Regards, -- Fujii Masao
