Hi! On Tue, Nov 28, 2023 at 2:06 PM Aleksander Alekseev <aleksan...@timescale.com> wrote: > > >> > I think what's done in patch 0001 is just an extension of existing > >> > logic and moving it into separate function. > >> > >> That's right. I'm arguing that now is a good time to clean it up. > >> > >> I won't insist if Alexander prefers to commit it as it is, though. But > >> let's at least explain how this works in the comment, instead of the XXX. > > > > I agree with you that would be good to add a comment instead of XXX and > > commit. > > +1 > > One could argue that getting rid of short filenames entirely in the > long term (i.e. always long_segment_names == true) could be a better > strategy. Maybe it's not but I believe this should be discussed > separately from the patchset under question.
Heikki, thank you for catching this. This mess with file names formats already lasts quite long. I don't think we should hurry unifying this as long as we're anyway going to change that in near future. Please, find the revised patchset with relevant comment. ------ Regards, Alexander Korotkov
0003-Make-use-FullTransactionId-in-2PC-filenames-v62.patch
Description: Binary data
0002-Use-larger-segment-file-names-for-pg_notify-v62.patch
Description: Binary data
0004-Add-SLRU-tests-for-64-bit-page-case-v62.patch
Description: Binary data
0001-Index-SLRUs-by-64-bit-integers-rather-than-by-32-v62.patch
Description: Binary data
