3 ms·
I wonder if the file scanning is hitting a variant of the N+1 problem, and if there's even a way to avoid that using the file system APIs. There's an old paper
by cratermoon 4y ago
I wonder if the file scanning is hitting a variant of the N+1 problem, and if there's even a way to avoid that using the file system APIs. There's an old paper titled "New Need Not Be Slow"[1] that goes into the right and wrong ways to scan a filesystem for lots of files.
1 ftp://ftp.cs.toronto.edu/doc/programming/cnews.{ps,pdf}
- nicklaf 4y agoMinor correction: should be ftp://ftp.cs.toronto.edu/doc/programming/c-news.ps