replace rehydration from CSSOM with progressive regex

[quantizor]

fixes #2812

[quantizor]

@kitten how can I benchmark this to compare against the existing v5 changes?

[kitten]

Not a big fan of the blanket eslint-disable and the content.split(‘“‘)[1] in there but logic looks fine and the code is clear 👍

You could just run a large production site (if you have one) and manually compare the loading stack trace. You should be able to observe the rehydration times. If this is as bad as it was before you’ll see a 5-10x regression

[quantizor]

Open to suggestions, the eslint disable is just because a basic suppression wasn’t working for the while assignment

…

On Mon, Nov 11, 2019 at 4:39 PM Phil Plückthun ***@***.***> wrote: ***@***.**** commented on this pull request. Not a big fan of the blanket eslint-disable and the content.split(‘“‘)[1] in there but logic looks fine and the code is clear 👍 You could just run a large production site (if you have one) and manually compare the loading stack trace. You should be able to observe the rehydration times. If this is as bad as it was before you’ll see a 5-10x regression — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2872?email_source=notifications&email_token=AAELFVWMI7LM5SYAPSUYSBLQTHGHVA5CNFSM4JL25RTKYY3PNVWWK3TUL52HS4DFWFIHK3DMKJSXC5LFON2FEZLWNFSXPKTDN5WW2ZLOORPWSZGOCLESJMI#pullrequestreview-315172017>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAELFVSGUSMNY5GBVDILSI3QTHGHVANCNFSM4JL25RTA> .

[quantizor]

I don’t currently have a large prod app to test it on

…

On Mon, Nov 11, 2019 at 4:46 PM Evan Jacobs ***@***.***> wrote: Open to suggestions, the eslint disable is just because a basic suppression wasn’t working for the while assignment On Mon, Nov 11, 2019 at 4:39 PM Phil Plückthun ***@***.***> wrote: > ***@***.**** commented on this pull request. > > Not a big fan of the blanket eslint-disable and the content.split(‘“‘)[1] > in there but logic looks fine and the code is clear 👍 > > You could just run a large production site (if you have one) and manually > compare the loading stack trace. You should be able to observe the > rehydration times. If this is as bad as it was before you’ll see a 5-10x > regression > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub > <#2872?email_source=notifications&email_token=AAELFVWMI7LM5SYAPSUYSBLQTHGHVA5CNFSM4JL25RTKYY3PNVWWK3TUL52HS4DFWFIHK3DMKJSXC5LFON2FEZLWNFSXPKTDN5WW2ZLOORPWSZGOCLESJMI#pullrequestreview-315172017>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAELFVSGUSMNY5GBVDILSI3QTHGHVANCNFSM4JL25RTA> > . >

[quantizor]

Available to test at styled-components@5.0.0-regexrehydrate