On Sat, 23 May 2026, Ludovic Courtès <[email protected]> wrote: > Olivier Dion <[email protected]> skribis: >> Also, I think that psyntax should maybe do it's own recursive traversal >> of the form for hashing. That would avoid the issue that Tomas >> mentionned with duplicated top-level identifiers. > > Recursive traversal could quickly become terrible though, making macro > expansion even more expensive than it already is.
It would be proportional to the expanded form. I don't think that this would be a problem in general, since I think the hashing is only done on top-level identifiers that are part of the matching syntax. Only way to know is to measure it I guess. But perhaps there are better way of generating top-level identifiers in a reproducible manner. > There’s also the caveat that due to this kind of use cases, ‘hash’ & > co. cannot be changed in a stable series because that’d indirectly be an > ABI break. Yes since this would change symbol naming. That would definitely not be part of the stable series. Thanks, Olivier -- Olivier Dion
