Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
正規表現改善報告する回 / Regexp memoization progress report
Search
Sponsored
·
Ship Features Fearlessly
Turn features on and off without deploys. Used by thousands of Ruby developers.
→
TSUYUSATO Kitsune
August 19, 2023
Programming
710
2
Share
正規表現改善報告する回 / Regexp memoization progress report
RubyKaigi 2023 follow up(
https://rhc.connpass.com/event/288535/
) での発表資料です。
TSUYUSATO Kitsune
August 19, 2023
More Decks by TSUYUSATO Kitsune
See All by TSUYUSATO Kitsune
(Re)make Regexp in Ruby: Democratizing internals for the JIT
makenowjust
3
570
「正規表現をつくる」をつくる / make "make regex"
makenowjust
1
1.6k
最近の研究とか、RubyへのContributionとか / Recent My Study and Ruby Contributions
makenowjust
2
290
Make Parsers Compatible Using Automata Learning
makenowjust
3
13k
YAPC::Japan::Online 2022で発表して WEB+DB PRESSに記事を寄稿した話
makenowjust
0
64
Regular Expressions, REXML, Automata Learning
makenowjust
0
450
オートマトン学習しろ / Do automata learning
makenowjust
3
580
#kaigieffect LT 2024 - rexml-css_selector: A REXML extension for supporting CSS selector
makenowjust
1
460
RubyKaigi 2024 - Make Your Own Regex Engine!
makenowjust
1
2.2k
Other Decks in Programming
See All in Programming
Lightning-Fast Method Calls with Ruby 4.1 ZJIT / RubyKaigi 2026
k0kubun
3
1.1k
運転動画を検索可能にする〜Cosmos-Embed1とDatabricks Vector Searchで〜/cosmos-embed1-databricks-vector-search
studio_graph
0
410
Vibe하게 만드는 Flutter GenUI App With ADK , 박제창, BWAI Incheon 2026
itsmedreamwalker
0
550
Swift Concurrency Type System
inamiy
1
540
感情を設計する
ichimichi
5
1.5k
Programming with a DJ Controller — not vibe coding
m_seki
3
140
VueエンジニアがReactを触って感じた_設計の違い
koukimiura
0
180
Agentic Elixir
whatyouhide
0
380
t *testing.T は どこからやってくるの?
otakakot
1
700
tRPCの概要と少しだけパフォーマンス
misoton665
2
220
ハーネスエンジニアリングとは?
kinopeee
12
5.9k
年間50登壇、単著出版、雑誌寄稿、Podcast出演、YouTube、CM、カンファレンス主催……全部やってみたので面白さ等を比較してみよう / I’ve tried them all, so let’s compare how interesting they are.
nrslib
4
800
Featured
See All Featured
How to Ace a Technical Interview
jacobian
281
24k
Breaking role norms: Why Content Design is so much more than writing copy - Taylor Woolridge
uxyall
0
260
CoffeeScript is Beautiful & I Never Want to Write Plain JavaScript Again
sstephenson
162
16k
Docker and Python
trallard
47
3.8k
A designer walks into a library…
pauljervisheath
211
24k
Thoughts on Productivity
jonyablonski
76
5.1k
The World Runs on Bad Software
bkeepers
PRO
72
12k
Chrome DevTools: State of the Union 2024 - Debugging React & Beyond
addyosmani
10
1.1k
Max Prin - Stacking Signals: How International SEO Comes Together (And Falls Apart)
techseoconnect
PRO
0
150
Templates, Plugins, & Blocks: Oh My! Creating the theme that thinks of everything
marktimemedia
31
2.8k
コードの90%をAIが書く世界で何が待っているのか / What awaits us in a world where 90% of the code is written by AI
rkaga
61
43k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
31
3.2k
Transcript
౻࿘େ !NBLF@OPX@KVTU !3VCZ,BJHJGPMMPXVQ ਖ਼نදݱվળใࠂ͢Δճ
લճͷ͋Β͢͡ IUUQTSVCZLBJHJPSHQSFTFOUBUJPOTNBLFOPXKVTUIUNMEBZ
લճͷ͋Β͢͡ ਖ਼نදݱϚονϯάΛΊͬͪΌͨ͘͠
લճͷ͋Β͢͡ ͔͠͠ɺਖ਼نදݱ͕͘ͳΔͨΊʹ੍͍͔͕ͭ͘
લճͷ͋Β͢͡ ੍ϑϦʔͳੈքΛΊͯ͟͠ ό Ϧ Ξ
੍#FGPSF w ͜ΕΒͷػೳΛ͏ͱɺਖ਼نදݱϚονϯά͕࠷దԽ͞Εͳ͍ ʮઌಡΈɾޙಡΈʯ(?=foo) (?<=bar)ʮΞτϛοΫάϧʔϓʯ(?>foo) ʮ݅ذʯ(?(<x>)yes|no)ʮඇแؚԋࢉࢠʯ(?~foo) ʮޙํࢀরʯ(foo)\1
ʮ෦ࣜͷݺͼग़͠ʯ(?<x>foo)\g<x> w ͦͷଞɺಾͷ੍ͨͪ ճࢦఆͷ܁Γฦ͕͠ωετͰ͖ͳ͍(fo{1,23}){42} ۭจࣈྻʹϚον͢Δ͔͠Εͳ͍܁Γฦ͕͠ωετͰ͖ͳ͍
w ͜ΕΒͷػೳΛ͏ͱɺਖ਼نදݱϚονϯά͕࠷దԽ͞Εͳ͍ ʮઌಡΈɾޙಡΈʯ(?=foo) (?<=bar)ʮΞτϛοΫάϧʔϓʯ(?>foo) ʮ݅ذʯ(?(<x>)yes|no)ʮඇแؚԋࢉࢠʯ(?~foo) ʮޙํࢀরʯ(foo)\1 ʮ෦ࣜͷݺͼग़͠ʯ(?<x>foo)\g<x>
w ͦͷଞɺಾͷ੍ͨͪ ճࢦఆͷ܁Γฦ͕͠ωετͰ͖ͳ͍(fo{1,23}){42} ۭจࣈྻʹϚον͢Δ͔͠Εͳ͍܁Γฦ͕͠ωετͰ͖ͳ͍ ੍"GUFS
w ͜ΕΒͷػೳΛ͏ͱɺਖ਼نදݱϚονϯά͕࠷దԽ͞Εͳ͍ ʮઌಡΈɾޙಡΈʯ(?=foo) (?<=bar)ʮΞτϛοΫάϧʔϓʯ(?>foo) ʮ݅ذʯ(?(<x>)yes|no)ʮඇแؚԋࢉࢠʯ(?~foo) ʮޙํࢀরʯ(foo)\1 ʮ෦ࣜͷݺͼग़͠ʯ(?<x>foo)\g<x>
w ͦͷଞɺಾͷ੍ͨͪ ճࢦఆͷ܁Γฦ͕͠ωετͰ͖ͳ͍(fo{1,23}){42} ۭจࣈྻʹϚον͢Δ͔͠Εͳ͍܁Γฦ͕͠ωετͰ͖ͳ͍ ੍"GUFS ඍົʁ
w ͜ΕΒͷػೳΛ͏ͱɺਖ਼نදݱϚονϯά͕࠷దԽ͞Εͳ͍ ʮઌಡΈɾޙಡΈʯ(?=foo) (?<=bar)ʮΞτϛοΫάϧʔϓʯ(?>foo) ʮ݅ذʯ(?(<x>)yes|no)ʮඇแؚԋࢉࢠʯ(?~foo) ʮޙํࢀরʯ(foo)\1 ʮ෦ࣜͷݺͼग़͠ʯ(?<x>foo)\g<x>
w ͦͷଞɺಾͷ੍ͨͪ ճࢦఆͷ܁Γฦ͕͠ωετͰ͖ͳ͍(fo{1,23}){42} ۭจࣈྻʹϚον͢Δ͔͠Εͳ͍܁Γฦ͕͠ωετͰ͖ͳ͍ ੍"GUFS ཧతʹແཧ ͋·ΓΘΕͯͳ͍
w ͜ΕΒͷػೳΛ͏ͱɺਖ਼نදݱϚονϯά͕࠷దԽ͞Εͳ͍ ʮઌಡΈɾޙಡΈʯ(?=foo) (?<=bar)ʮΞτϛοΫάϧʔϓʯ(?>foo) ʮ݅ذʯ(?(<x>)yes|no)ʮඇแؚԋࢉࢠʯ(?~foo) ʮޙํࢀরʯ(foo)\1 ʮ෦ࣜͷݺͼग़͠ʯ(?<x>foo)\g<x>
w ͦͷଞɺಾͷ੍ͨͪ ճࢦఆͷ܁Γฦ͕͠ωετͰ͖ͳ͍(fo{1,23}){42} ۭจࣈྻʹϚον͢Δ͔͠Εͳ͍܁Γฦ͕͠ωετͰ͖ͳ͍ ੍"GUFS ཧతʹແཧ ͋·ΓΘΕͯͳ͍ ͦΜͳʹ ѱ͘ͳ͍ʁ
͜Ε·ͰͷϝϞԽ w ී௨ͷਖ਼نදݱͷ߹ɺ ʮ͋Δঢ়ଶʹ͋ΔҐஔ͔ΒͷϚονͰࣦഊͨ͠ʯͱ͍͏ใΛه͢Ε0, w ϝϞԽςʔϒϧͷܕ memo: (State,
Int) -> (NoMemo | Failure) w ઌಡΈɾޙಡΈ͕͋Δ߹ɺΞτϛοΫάϧʔϓ͕͋Δ߹ʁ
ઌಡΈɾޙಡΈͷϝϞԽ w ઌಡΈɾޙಡΈ෦Ϛονϯάࣦഊ͍ͯ͠ͳͯ͘ɺ ʮ͋Δঢ়ଶɾ͋ΔҐஔʯʹ͏Ұ౸ୡ͢ΔՄೳੑ͕͋Δ ྫ/a*?(?=a*)z/ w ʮઌಡΈɾޙಡΈ෦ͷϚονϯάʹޭͨ͠ʯͱ͍͏ใΛ
ϝϞԽςʔϒϧʹه͢Δඞཁ͕͋Δ w ϝϞԽςʔϒϧͷܕ memo: (State, Int) -> (NoMemo | Success | Failure)
ΞτϛοΫά ϧʔϓͷϝϞԽ w ΞτϛοΫά ϧʔϓͷ߹ɺ ΞτϛοΫά ϧʔϓͷதͰͷࣦഊͱɺ֎ଆͰͷࣦഊΛ۠ผ͠ͳ͚Ε͍͚ͳ͍ ֎ଆͰࣦഊͨ͠߹ɺΞτϛοΫά ϧʔϓͷதͷόοΫτϥοΫ
লུ͠ͳ͚Ε͍͚ͳ͍ w ϝϞԽςʔϒϧͷܕ memo: (State, Int) -> (NoMemo | Success | AtomicFailure | Failure)
࣮ͷमਖ਼ w ࣮ࡍͷϝϞԽςʔϒϧCJUྻͳͷͰɺ ઌಡΈɾޙಡΈɾΞτϛοΫά ϧʔϓͷঢ়ଶʹରͯ͠ CJUͬͯϝϞԽ͢ΔΑ͏ʹͨ͠ w ͜Ε·Ͱʮ͋Δঢ়ଶʹ͋ΔҐஔͰ౸ୡͨ͠ʯͱ͍͏ใΛه͍͕ͯͨ͠ɺ
ʮ͋Δঢ়ଶʹ͋ΔҐஔ͔ΒͷϚονʹࣦഊͨ͠ʯͱ͍͏ใʹ͢ΔͨΊɺ όοΫτϥοΫதʹϝϞԽςʔϒϧΛߋ৽͢ΔΑ͏ʹͨ͠ IUUQTHJUIVCDPNSVCZSVCZQVMM
ݱࡏͷ੍ w ͜ΕΒͷػೳΛ͏ͱɺਖ਼نදݱϚονϯά͕࠷దԽ͞Εͳ͍ ʮ݅ذʯ(?(<x>)yes|no)ʮඇแؚԋࢉࢠʯ(?~foo) ʮޙํࢀরʯ(foo)\1 ʮ෦ࣜͷݺͼग़͠ʯ(?<x>foo)\g<x> w ͦͷଞɺಾͷ੍ͨͪ
ճࢦఆͷ܁Γฦ͕͠ωετͰ͖ͳ͍(fo{1,23}){42} ઌಡΈɾޙಡΈͱΞτϛοΫά ϧʔϓ͕ωετͰ͖ͳ͍ ઌಡΈɾޙಡΈɾΞτϛοΫά ϧʔϓͷதͰΩϟϓνϟ͑ͳ͍
ΏΔ΅ w ੍ʹ͔͔Βͳ͍͔Ͳ͏͔ΛRegexp.linear_time?ͰνΣοΫͰ͖·͢ w ͜ΕΛͬͯɺϓϩάϥϜதͷਖ਼نදݱ੍͕ʹ͔͔Βͳ͍͔νΣοΫ͢Δ 3VCPDPQϓϥάΠϯΛ࡞ͬͯɺϝϯςφϯεͯ͘͠ΕΔਓΛืू͍ͯ͠·͢ SVCZKQ4MBDLͷSFHFYQνϟϯωϧʹͯ