On Mon, Oct 14, 2024 at 09:21:50PM +0100, Gavin Smith wrote: > I may as well post what I've been able to come up with. The new > code is about 150 lines of C, in the file convert/hashmap.c (patch > below). Probably not perfect but hopefully simple enough to be > maintainable. > > If this approach looks ok I can try to tidy it up and commit it. > We could also experiment with different numbers of hash bins or the > memory allocator. I'll read a bit more about hash tables as I don't > know a lot about this subject.
This looks good to me, it is quite simple. Maybe the number of buckets could be set based on the number of sections + nodes + index entries? -- Pat
