Stream: t-compiler/wg-polonius

Topic: meeting 2019.03.28


Albin Stjerna (Mar 28 2019 at 20:15, on Zulip):

I can't remember if we formally decided on not having a meeting today (and I am guessing we aren't), but I'm going to have a meeting with myself anyway and report on what I have done during the week. Hope everyone are enjoying themselves at a) the conference or b) whatever else they are doing!

I have:
- extended rustc-perf to run an extra check round with Polonius, analogous to how NLL benchmarks are run
- used this extended version of the benchmarks to extract relative runtime and memory usage measurements
- plotted the results, and done manual and k-means clustering to figure out the sorts of performance outliers (compared to NLL) we are looking at
- done the work to Polonius and sub-crates to integrate the necessary facts for liveness analysis
- produced a probably slightly to moderately rubbish implementation of the liveness analysis as described in Polonius#104 (PR, concerns described in the PR)
- cleaned up a few Clippy errors (in separate commits) as I went by

For the remainder of the "week" (which may or may not be aligned to the actual week) I will focus on preparing the paperwork for my master's project (I have still not formally registered for the project course and err am no longer formally a student at all, in fact), and, time permitting, on poking around rustc and seeing if I can figure out what to do to extend the fact generation parts to accommodate the new liveness analysis on my own.

Albin Stjerna (Mar 28 2019 at 20:22, on Zulip):

Great meeting. Can I use this opportunity to elect myself to something cool?

Last update: Nov 15 2019 at 20:00UTC