Rip out nodex
Because we no longer use a generic cnode implementation, we can just wrap the cnode response in a {:myhtml_response, resp}
tuple and pattern match on it, so we don't need to start a separate process to receive a message without collisions.
Benchmarks (2019-11-6--17-48-15-utc
is master
):
##### With input document-large.html #####
Name ips average deviation median 99th %
Decoding 5.96 167.80 ms ±8.45% 162.95 ms 215.39 ms
Decoding (2019-11-6--17-48-15-utc) 4.66 214.46 ms ±4.91% 213.82 ms 245.64 ms
Comparison:
Decoding 5.96
Decoding (2019-11-6--17-48-15-utc) 4.66 - 1.28x slower +46.67 ms
##### With input document-medium.html #####
Name ips average deviation median 99th %
Decoding 343.03 2.92 ms ±26.71% 2.74 ms 6.04 ms
Decoding (2019-11-6--17-48-15-utc) 282.66 3.54 ms ±18.98% 3.52 ms 5.60 ms
Comparison:
Decoding 343.03
Decoding (2019-11-6--17-48-15-utc) 282.66 - 1.21x slower +0.62 ms
##### With input document-small.html #####
Name ips average deviation median 99th %
Decoding 865.17 1.16 ms ±49.53% 1.04 ms 4.25 ms
Decoding (2019-11-6--17-48-15-utc) 850.08 1.18 ms ±28.47% 1.10 ms 2.02 ms
Comparison:
Decoding 865.17
Decoding (2019-11-6--17-48-15-utc) 850.08 - 1.02x slower +0.0205 ms
##### With input fragment-large.html #####
Name ips average deviation median 99th %
Decoding 578.09 1.73 ms ±30.80% 1.62 ms 4.49 ms
Decoding (2019-11-6--17-48-15-utc) 533.21 1.88 ms ±29.82% 1.84 ms 3.70 ms
Comparison:
Decoding 578.09
Decoding (2019-11-6--17-48-15-utc) 533.21 - 1.08x slower +0.146 ms
##### With input fragment-small.html #####
Name ips average deviation median 99th %
Decoding (2019-11-6--17-48-15-utc) 2.10 K 475.45 μs ±62.68% 451.38 μs 1010.31 μs
Decoding 2.05 K 488.00 μs ±92.28% 419.93 μs 3105.03 μs
Comparison:
Decoding (2019-11-6--17-48-15-utc) 2.10 K
Decoding 2.05 K - 1.03x slower +12.55 μs
Closes #2 (closed)