Rob Figueiredo

The 40+ seconds happens on both Darwin & Linux, laptop and workstation, so I don't think it's a quirk of my setup. FWIW, the Go std library checks quickly for...

That's too bad. In theory, it should be sufficient to look only at the signatures of functions that are directly called (the first degree dependencies). That should be much faster...

FYI, I am implementing the [TechEmpower Benchmarks](techempower.com/benchmarks/) with a couple different Go ORMs and submitting the fastest one as the Revel entry. I would be happy to benchmark beedb if...

Yeah, I found that by googling and tried it, but didn't notice a difference (unless I did it wrong). But also, I don't think that's a reasonable requirement for fsnotify...

Even if you don't track deletes, won't it still cause the watcher to remove its watch? Also, that throttling routine looks like an explanation for the behavior of getting an...

I see.. in that case, I still have no idea :)

I definitely agree with #56 that a recursive watch is super common and should be supported. To your second point, is there a problem that it requires a framework, or...

(It certainly seems like being able to specify platform-specific flags to a watch call would be useful and not inappropriate. )

Related - https://github.com/bazelbuild/bazel-gazelle/issues/786 What was the reasoning not to delegate to path.Match if a pattern didn't include "**"?

Yes, I believe it was called 2 - 5M times. There is a list of patterns that is compared to every file in a repository. It could probably be more...