kbara

That looks like a pretty good example of some of the things I miss most working with Snabb. It'd be great to move more in that direction, especially as core...

Horrifyingly enough, for full compatibility we need to support an arbitrary number of '-'s. % tcpdump -d "len < 1 ---1" (000) ld #pktlen (001) jge #0x0 jt 2 jf...

It gets sillier: % tcpdump -d "len < 1 - - -- 1" (000) ld #pktlen (001) jge #0x2 jt 2 jf 3 Arbitrary spaces seem to be allowed.

Yep, you're entirely right. The relevant bits of libpcap source define gen_neg and UMINUS, and it's a unary operator.

Any preferred format/location for the documentation? @andywingo

Here's an initial draft. It uses the convert_ssa/convert_anf APIs with the extra optimize flag, although those are still in a PR. --- ``` The pflua API Pflua compiles pflang (libpcap...

GC plays a role in this, especially for the smaller benchmarks. Wingolog-2 master, with normal GC:  Wingolog-2 master, without GC (forcing a garbage collection, then stopping garbage collection...

We can mark it as low-priority and work on other things, but I think those variance pictures are from after fixing the unsunk CNEWI issue. Here's wingolog-2 from today, showing...

This bug is specific to the non-optimized bpf-lua pipeline; with --force-opt, the problem does not show up.

This is marked as high-priority, but hasn't been closed or merged despite being open for months. Is the underlying issue fixed, or does this patch need to be updated and...