Proper handling of max-age and expires for cookies

[watzon]

This PR is based off of #5042. Seeing as it has been sitting there for some time with no movement, and the current Cookie parsing is breaking with large max-age values, I decided to get to ball rolling again.

[watzon]

It looks like the failures are due to a memory leak of some kind, any idea what could be causing it?

[asterite]

I think compiler specs are leaking. I just noticed that codegen specs that don't use the prelude are allocating memory using Pointer.malloc or Foo.new which end up calling LibC.mallocand that memory is never freed.

However, specs usually don't allocate that much memory in tests so maybe there's something else going on... but it's hard to debug. Plus it only happens on linux 32, mostly.

[watzon]

It's weird. I can't even run the full library specs on my machine or it eats up all my RAM. I've had to force shutdown my laptop twice. The cookie specs themselves pass though.

[asterite]

Looks good!

However, the time comparisons seem a bit fragile, for example if there's a hiccup in CI or something. I think we should bring timecop into the standard library to be able to properly test these things. But we can do it later.

[watzon]

Holy moley, the specs actually passed

[watzon]

Looks like we just need to get @straight-shoota in here for review

[Sija]

@watzon Two approvals is what's needed (would gr8 to have @straight-shoota review nevertheless).

[RX14]

I'd like another approval of the changes made. There's only been one review since the last commit.

[ghost]

Does this PR need someone to adopt it? I could really use the change.

[watzon]

@yui-knk I believe so. I just don't have time to work on the requested changes right now.

obsolete

[beta-ziliani]

Closing this one now that we have #10564 to focus on. Thanks for the effort @watzon !