Re: testing 3.4.90 with PostgreSQL grammar

[Akim Demaille]

Hi John,

> Le 31 oct. 2019 à 13:11, John Naylor <address@hidden> a écrit :
> 
> Hi,
> 
> I built Postgres trunk with the recent Bison beta, and everything
> works fine, at least as far as our regression tests are concerned.

Thanks a lot, that's good to know.

> I also wanted to see if api.token.raw offered any speedup at all.
> After ripping out all the single-character literal tokens from the
> Postgres grammars, I got it to work and I can see that YYTRANSLATE is
> now a no-op. Using a micro-benchmark that isolates raw parsing (Flex +
> Bison), the api.token.raw option increases performance by about
> 1.7-1.9%.

Yeah, it's probably not very big, indeed.  But it's somewhat a useless
cost.  Of course it means more definitions in the parser, and more rules
in the scanner, so there's a tradeoff here.

Thanks for reporting your mileage!