| commit | 0315074fc670cd81e9be6775ca918abaa2f3729f | |
| author | Eric Blake <eblake@redhat.com> | |
| Mon, 18 Dec 2023 14:04:30 +0000 (18 08:04 -0600) | ||
| committer | Eric Blake <eblake@redhat.com> | |
| Mon, 18 Dec 2023 15:24:50 +0000 (18 09:24 -0600) | ||
| tree | a904564296ca7c958fbed92b89fc49f94abdcd8b | treesnapshot (tar.gz zip) |
| parent | 68fe4fe230cbaf7886783a5739539def55f157b0 | commitdiff |
day 15 golf more, 710 effective bytes
727 bytes here; remove 9 newlines plus these 8 bytes: "$1,a,97,".
Adds more time; now >26m.
For grins and giggles, I tested how bad the O(n^3) effects are.
Comparing the sample input (52 bytes, 11 elements):
$ m4 -dae --trace=_ -DI=file -Dfile=tmp.15 day15.golfm4 2>&1 | wc
10195 17974 1841213
with my input (22969 bytes, 4000 elements):
$ m4 -dae --trace=_ -DI=file -Dfile=day15.input day15.golfm4 2>&1 | wc
58172577 2885010715 625012937407
Yes, that really is 58 million calls, and making m4 parse over half a
terrabyte of data JUST on the arguments and expansion of my macro (not
to mention the parsing of the two index() and ifelse() per _() call).
727 bytes here; remove 9 newlines plus these 8 bytes: "$1,a,97,".
Adds more time; now >26m.
For grins and giggles, I tested how bad the O(n^3) effects are.
Comparing the sample input (52 bytes, 11 elements):
$ m4 -dae --trace=_ -DI=file -Dfile=tmp.15 day15.golfm4 2>&1 | wc
10195 17974 1841213
with my input (22969 bytes, 4000 elements):
$ m4 -dae --trace=_ -DI=file -Dfile=day15.input day15.golfm4 2>&1 | wc
58172577 2885010715 625012937407
Yes, that really is 58 million calls, and making m4 parse over half a
terrabyte of data JUST on the arguments and expansion of my macro (not
to mention the parsing of the two index() and ifelse() per _() call).
| 2023/day15.golfm4 | diffblobblamehistory |