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
Contributing to the Ruby Parser
Search
S.H.
May 16, 2024
Technology
2
1.2k
Contributing to the Ruby Parser
In RubyKaigi 2024 Lightning Talks Slide
S.H.
May 16, 2024
Tweet
Share
More Decks by S.H.
See All by S.H.
ゆるゆるMastodon 鯖缶生活
gamelinks007
0
660
1週間で作るActivityPubリレーサーバ
gamelinks007
0
110
Rails 8で作るActivityPub リレーサーバ
gamelinks007
0
450
Developing an ActivityPub Relay with Rails 8
gamelinks007
0
61
Mastodon on Ruby master
gamelinks007
0
120
Trying to Make Ruby's Parser Available as a Gem
gamelinks007
1
340
ユーザーから見たLrama
gamelinks007
0
220
Other Decks in Technology
See All in Technology
絶対に失敗できないキャンペーンページの高速かつ安全な開発、WINTICKET × microCMS の開発事例
microcms
0
170
退屈なことはDevinにやらせよう〜〜Devin APIを使ったVisual Regression Testの自動追加〜
kawamataryo
4
810
つくって納得、つかって実感! 大規模言語モデルことはじめ
recruitengineers
PRO
25
8.2k
Goss: Faiss向けの新しい本番環境対応 Goバインディング #coefl_go_jp
bengo4com
0
1.4k
実践データベース設計 ①データベース設計概論
recruitengineers
PRO
4
960
実践AIガバナンス
asei
3
150
DeNA での思い出 / Memories at DeNA
orgachem
PRO
3
1.8k
どこで動かすか、誰が動かすか 〜 kintoneのインフラ基盤刷新と運用体制のシフト 〜
ueokande
0
200
新卒(ほぼ)専業Kagglerという選択肢
nocchi1
1
2.4k
AIとTDDによるNext.js「隙間ツール」開発の実践
makotot
6
750
Preferred Networks (PFN) とLLM Post-Training チームの紹介 / 第4回 関東Kaggler会 スポンサーセッション
pfn
PRO
1
260
実践アプリケーション設計 ①データモデルとドメインモデル
recruitengineers
PRO
4
830
Featured
See All Featured
GitHub's CSS Performance
jonrohan
1031
460k
Thoughts on Productivity
jonyablonski
69
4.8k
How to Create Impact in a Changing Tech Landscape [PerfNow 2023]
tammyeverts
53
2.9k
Building Applications with DynamoDB
mza
96
6.6k
Automating Front-end Workflow
addyosmani
1370
200k
Balancing Empowerment & Direction
lara
2
590
The Cult of Friendly URLs
andyhume
79
6.6k
GraphQLの誤解/rethinking-graphql
sonatard
71
11k
Navigating Team Friction
lara
189
15k
Principles of Awesome APIs and How to Build Them.
keavy
126
17k
Rails Girls Zürich Keynote
gr2m
95
14k
Bootstrapping a Software Product
garrettdimon
PRO
307
110k
Transcript
Contributing to the Ruby Parser S.H.
About Me • S.H. • Software engineer at ESM, Inc.
• Hamada.rb Organizer • I’m hosting Ruby Hacking Challenge in Hamada.rb
None
ruby/ruby contributors
ruby/ruby contributors I’m here!
parse.y contributors
parse.y contributors I’m here!
Today Topic is… • Why I got interested in the
Ruby Parser • How I started contributing to the Ruby Parser • How I started contributing to the Universal Parser
In RubyKaigi 2021 Takeout
In RubyKaigi 2021 Takeout • parse.y is difficult to maintain
◦ No one but nobu can maintain parse.y • parse.y seems hard to understand
Is parse.y really that hard to maintain?
I’m worried
I started to reading parse.y
My first impression of parse.y • The codebase is large
• It’s hard to figure out what to begin reading • I have a no clue about BNF
What I read to study to parse.y • Ruby Under
the Microscope • Ruby Hacking Guide • ruby trunk changes
I’ve started to understand parse.y a bit
Difficulties I noticed while reading • Ruby Parser and Ripper
are interwind • The overall codebase is large
Difficulties I noticed while reading • Ruby Parser and Ripper
are interwind • The overall codebase is large It’s too hard
Difficulties I noticed while reading • Ruby Parser and Ripper
are interwind • The overall codebase is large May be able to do something here?
Let’s do it!
Action • Reusing BNF • Introducing some macros • Introducing
Ruby C API
Reusing BNF (1)
Reusing BNF (1)
Reusing BNF (2)
Reusing BNF (2)
Reusing BNF (3)
Reusing BNF (3)
Introducing macro (1)
Introducing macro (2)
Introducing Ruby C API
Results • Cut down codebase • Added dependencies by introducing
macros and Ruby C API
In RubyKaigi 2023
In RubyKaigi 2023 • I feel the future of the
Ruby Parser in “The Future of the Ruby Parser” at spikeolaf • I realized that added macros and Ruby C API are make dependencies for Universal Parser
After RubyKaigi 2023
After RubyKaigi 2023 • I read “Started parse.y Refactoring Challenge”
◦ https://yui-knk.hatenablog.com/entry/202 3/06/18/162100 • It was written want to remove the generation of literal objects in parse.y
Why? • parse.y needs Ruby C API for generation of
literal objects • Will obtain a simpler AST ◦ e.g. Fixnum, Bignum, and Float…
Let’s do it!
Introducing new NODE (1)
Introducing new NODE (2)
Reducing C API dependencies (1)
Results • Obtained a simpler AST for some literal ◦
e.g. Numeric and __ENCODING__ • Reducing Ruby C API dependencies for Universal Parser
KPT Keep motivation Should more disscuss Share my experience Increase
Parser Contributor Keep Problem Try
parse.y is NOT Hell! We can hack parse.y!