On Wed, 2023-11-29 at 20:31 +0700, John Naylor wrote: > v5-0001 puts fash-hash as-is into a new header, named in a way to > convey in-memory use e.g. hash tables. > > v5-0002 does the minimal to allow dynash to use this for string_hash, > inlined but still calling strlen. > > v5-0003 shows one way to do a incremental interface. It might be okay > for simplehash with fixed length keys, but seems awkward for strings. > > v5-0004 shows a bytewise incremental interface, with implementations > for dynahash (getting rid of strlen) and guc hash.
I'm trying to follow the distinctions you're making between dynahash and simplehash -- are you saying it's easier to do incremental hashing with dynahash, and if so, why? If I understood what Andres was saying, the exposed hash state would be useful for writing a hash function like guc_name_hash(). But whether we use simplehash or dynahash is a separate question, right? Also, while the |= 0x20 is a nice trick for lowercasing, did we decide that it's better than my approach in patch 0004 here: https://www.postgresql.org/message-id/27a7a289d5b8f42e1b1e79b1bcaeef3a40583bd2.ca...@j-davis.com which optimizes exact hits (most GUC names are already folded) before trying case folding? Regards, Jeff Davis
