Re: pure parsers and reentrant scanners

On 17.01.25 16:40, Peter Eisentraut wrote:
> Here are some more patches.  This should cover the last sub-topic of
> this topic: not passing the yyparse() result via global variables.  This
> uses techniques that are already in use in some parsers in the tree, for
> example cube and jsonpath.  The syncrep parser was a bit trickier,
> because there we need to pass the syncrep_parse_error variable all the
> way down to the scanner (not just the parser), but overall it's all
> still pretty compact and standard.

This has been committed.

I think this concludes this topic.

For those who haven't seen it, I wrote a blog post about this:
https://peter.eisentraut.org/blog/2025/01/21/implementing-thread-safe-scanners-and-parsers-in-postgresql