Stream: t-compiler/wg-prioritization/alerts

Topic: I-prioritize #78471 Code bloat from RawVec::grow_amortized


triagebot (Oct 28 2020 at 11:19, on Zulip):

@WG-prioritization/alerts issue #78471 has been requested for prioritization.

Procedure

Hameer Abbasi (Oct 28 2020 at 11:21, on Zulip):

This might be controversial, but I think Vec::push should be inlined. So P-low.

apiraino (Oct 28 2020 at 12:00, on Zulip):

I went through the linked PR #72227 and it seems the discussion focuses on the best tradeoff for performace and memory usage without mentioning the final binary size, so I guess the final binary size is "collateral damage" I'm not sure it can be avoided, to my (little) understanding.

By the way, I looked "non PGO Firefox builds" up. They are "non Profile-Guided Builds":
https://developer.mozilla.org/en-US/docs/Mozilla/Developer_guide/Build_Instructions/Building_with_Profile-Guided_Optimization

apiraino (Oct 28 2020 at 12:00, on Zulip):

So yeah long story short :smile: I'd agree with P-low

lcnr (Oct 28 2020 at 12:24, on Zulip):

Vec::Push should be inlined, but we may not want to inline grow_amortized into Vec push. I still think that this is P-low, though I would also be fine with P-medium here

triagebot (Oct 28 2020 at 18:58, on Zulip):

Issue #78471's prioritization request has been removed.

Santiago Pastorino (Oct 28 2020 at 18:58, on Zulip):

went with P-medium prefer to go with the highest possible prios :)

Last update: Apr 11 2021 at 18:15UTC