New blog post! If you are a Kotlin or Java Android developer, this one should make you go 🤯.
https://www.romainguy.dev/posts/2024/naming-is-hard/
I'm not even sorry.
=> More informations about this toot | More toots from romainguy@androiddev.social
@romainguy Over the holidays you can rewrite the obfuscator to analyze which fields are used together to pick obfuscated names based on the L1 cache implications. Should be a quick and easy project 😉
=> More informations about this toot | More toots from iangclifton@androiddev.social
@iangclifton I thought about it many times 😅
=> More informations about this toot | More toots from romainguy@androiddev.social
@romainguy This hurt me
=> More informations about this toot | More toots from jw@jakewharton.com
@jw Yeah but you loved it too
=> More informations about this toot | More toots from romainguy@androiddev.social
@romainguy I very much did 🫤
=> More informations about this toot | More toots from jw@jakewharton.com
@romainguy Alphabetizing is such a strange choice. Especially given the use of ProGuard/R8. The declaration order would seem to be preferable in most cases (barring false sharing of course).
=> More informations about this toot | More toots from bondolo@mastodon.social
@bondolo I would indeed prefer declaration order as well. I should ask the ART team about the reasoning behind this choice. I'm sure there's a great reason.
=> More informations about this toot | More toots from romainguy@androiddev.social
@romainguy 🤯
=> More informations about this toot | More toots from vitriolix@mastodon.social
@romainguy That's...interesting.
=> More informations about this toot | More toots from ryan@harter.us
@romainguy WTF 😒
=> More informations about this toot | More toots from Aurimas@androiddev.social
@romainguy Well, that's unexpected 🤯 From now on I'm going to prepend a_
to all important fields. No, of course I'm just joking 😉
=> More informations about this toot | More toots from svenjacobs@androiddev.social
@romainguy i1, i2, i3, ... are returned back to the game :androidJetpack:
=> More informations about this toot | More toots from terrakok@androiddev.social
@romainguy 64 bytes cache line seems like it should be increased by now. I wonder how often phones get an L1 cache hit these days.
(not a magic wand, just an observation. changing the behaviour of ART would also be good)
=> More informations about this toot | More toots from michaelaubert@androiddev.social
😲 how much is R8 messing with micro optimizations? I imagine this would be one case but it cannot be the only
=> More informations about this toot | More toots from igokoro@androiddev.social
@romainguy @jw
=> More informations about this toot | More toots from designatednerd@mastodon.social
@romainguy, how did you even know this was a thing?
=> More informations about this toot | More toots from humazed@mastodon.social
@humazed I always wondered what the memory layouts of objects was so a few years I read through ART's source code. Also validated by testing when attempting to do some data oriented programming. Then procrastinated on a blog post for years 😂
=> More informations about this toot | More toots from romainguy@androiddev.social This content has been proxied by September (3851b).Proxy Information
text/gemini