Hi John! > Le 9 avr. 2020 à 08:22, John Naylor <[email protected]> a écrit : > > On Mon, Feb 24, 2020 at 2:11 AM Akim Demaille <[email protected]> wrote: >> Please, if you have some good benchmark somewhere, try the changes I >> suggested above, and report your mileage. It would be great if we could >> speed up the parsers by improving this. > > Hi, > > I benchmarked the Postgres parser with the latest Bison tip with and > without the nofinal patch, and the performance seems identical.
Thanks for reporting this, your input is valuable! Just to make sure: when you tried nofinal, you did pass -Dnofinal to bison, right? The patch installs the feature, but does not enable it by default. Alternatively, you can uncomment this line: #b4_percent_define_default([[nofinal]], [[true]]) in yacc.c. Cheers!
