Leo Antunes
Leo Antunes
@hjacobs just to complement: a similar race condition might also happen between the ingress and its DNS entry, if it's also being managed by k8s (via something like cloud provider...
getting `read tcp 127.0.0.1:33766->127.0.0.1:9125: i/o timeout` in the 1.19 tests. Maybe a flaky setup? :thinking:
Can we please keep this issue open? I understand not having the bandwidth to implement this, but it's still a valuable addition. Case in point: - [prometheus uses httptrace](https://github.com/prometheus/client_golang/blob/main/prometheus/promhttp/instrument_client.go#L163) to...
Is there maybe still a mistake in the benchmark that might be skewing the results?
@nigeltao > If you're talking about http.ServeContent, though, it may perform better to `os.Open` an `os.File` (because of the sendfile syscall) instead. Absolutely, that's exactly what we're currently doing (plus...
> if you're ultimately concerned about http.ServeContent performance, that calls io.CopyN, not io.Copy, and io.CopyN doesn't honor io.WriterTo oh boy, you're right, of course :disappointed: > In any case, I...
> if you're ultimately concerned about http.ServeContent performance, that calls io.CopyN, not io.Copy, and io.CopyN doesn't honor io.WriterTo oh, wait, not so sure: [`io.CopyN` uses `io.Copy`+`io.LimitReader`](https://cs.opensource.google/go/go/+/refs/tags/go1.19.2:src/io/io.go;l=362;drc=939f9fd64a740ef770a0ca3588f8c193dceebe83). So AFAICT this change...
> The WriteTo method cannot take the n argument passed to io.CopyN. :man_facepalming: yes, of course, sorry for the noise. (feel free to mark the last couple of msgs as...
For the record: my naive attempt at solving this involved using seteuid()/setegid() and euidaccess() to check for file access. That at least doesn't reinvent the wheel, but it still doesn't...
@pmontrasio the problem with this approach is that it creates (at least) an information leak. Unpriviledged users could create scripts that could create watches and monitor what's happening in folders...