Dimitri Fontaine <dimitri@2ndQuadrant.fr> writes:
> All done in the attached. As usual, you get an incremental patch and a
> complete one. It's also published on github for easy browsing:
>
> https://github.com/dimitri/postgres/compare/33f285e67f...ba42279924
Same as usual, even if that's a very little adjustment this time:
https://github.com/dimitri/postgres/commit/5ec1ddb41489a89eea09e350f6ee39e726f9fb03
>> any rate having NameData involved seems like it's probably not too
>> good an idea; if for some reason we need to keep that hash, use a
>> NUL-terminated string and initialize the hash table with string_hash
>> instead of tag_hash. That'll be simpler and also allows the length of
>> the buffer to vary independently of NAMEDATALEN, which can only be to
>> the good.
Fixed now, using a char[] as with PortalHashTable for example.
Regards,
--
Dimitri Fontaine
http://2ndQuadrant.fr PostgreSQL : Expertise, Formation et Support