Stream: t-compiler/wg-incr-comp

Topic: incr compile ICE building rustc


view this post on Zulip Aman Arora (Sep 14 2020 at 07:29):

I got following error while working on modifying capture analysis which is part of rustc_typeck

error: internal compiler error: compiler/rustc_middle/src/ty/query/on_disk_cache.rs:422:23: could not decode cached query result: read_option: expected 0 for None or 1 for Some

thread 'rustc' panicked at 'assertion failed: `(left == right)

Complete log is at http://csclub.uwaterloo.ca/~a52arora/rust-builds/log_75e93202-df10-4ea2-bff9-167ff9b968c2

I couldn't find an open issue with similar problem. I haven't built since, what all information (and how) I can collect to provide enough context

view this post on Zulip Aman Arora (Sep 14 2020 at 07:31):

I build using x.py build -j50 -i --stage 1

view this post on Zulip pnkfelix (Sep 14 2020 at 16:29):

lets see, so you haven't done a x.py clean yet, right? Please don't do a clean; or at least, first archive the state of your build/ directory so that we'll have access to the incremental build artifacts

view this post on Zulip Aman Arora (Sep 14 2020 at 17:10):

I haven't done clean/run x.py since.

My build dir is at 39GB, are there smaller subsections that I can backup instead?

view this post on Zulip simulacrum (Sep 14 2020 at 17:12):

It would be best to do everything just to be 100% sure

view this post on Zulip simulacrum (Sep 14 2020 at 17:13):

it looks like the bug is not in tests, though, so you could leave that out probably

view this post on Zulip simulacrum (Sep 14 2020 at 17:14):

do you have enough space to make a local copy? That would be enough to where you can start deleting things after duplicating it without losing state

view this post on Zulip Aman Arora (Sep 14 2020 at 18:18):

http://csclub.uwaterloo.ca/~a52arora/build.tar

view this post on Zulip Aman Arora (Sep 14 2020 at 22:13):

https://github.com/rust-lang/rust/issues/76720

Please let me know if there is anything else that will be needed, otherwise I'll be cleaning by build dir tonight tomorrow

view this post on Zulip Aman Arora (Sep 17 2020 at 06:30):

Not sure if something changed around incr compile, but I was able to trigger this again: http://csclub.uwaterloo.ca/~a52arora/rust-builds/log_30d4a7ae-dc48-41c6-a2c3-39043eee22bc

and only changes since my last x.py clean have been around how closures and generators are being printed


Last updated: Oct 21 2021 at 21:02 UTC