Hi Pavel, > I think the direction is good and delivering 64-bit GUCs is very much worth > committing. > The patch itself looks good, but we could need to add locks against > concurrently modifying 64-bit values, which could be non-atomic on older > architectures.
Thanks for the feedback. > I think 12345678912345 is good enough. Underscore dividers make reading > little bit easier but look weird overall. I can't remember other places where > we output long numbers with dividers. We already support this in SQL: psql (18devel) =# SELECT 123_456; ?column? ---------- 123456 -- Best regards, Aleksander Alekseev
