Stream: rustdoc

Topic: review request: PR #84867 (revert of PR 80653)


view this post on Zulip pnkfelix (May 03 2021 at 14:18):

Hi T-rustdoc

view this post on Zulip pnkfelix (May 03 2021 at 14:19):

Unfortunately I wasn’t able to just trivially git revert PR #80653

view this post on Zulip pnkfelix (May 03 2021 at 14:19):

so PR #84867 is my attempt to manually craft a revert

view this post on Zulip pnkfelix (May 03 2021 at 14:20):

It adds a regression test for #82465. I verified that the regressiont test fails without this PR, and passes with this PR.

view this post on Zulip simulacrum (May 03 2021 at 14:20):

@pnkfelix hmm was this meant to go into 1.52? Do we need to cherry-pick this onto the currently in CI promotion?

view this post on Zulip simulacrum (May 03 2021 at 14:21):

seems like likely yes

view this post on Zulip simulacrum (May 03 2021 at 14:22):

unfortunately your PR doesn't land on beta/stable branches (both largely identical at this moment) cleanly

view this post on Zulip simulacrum (May 03 2021 at 14:23):

let me see if the actual PR reverts cleanly

view this post on Zulip simulacrum (May 03 2021 at 14:23):

cc @Pietro Albini

view this post on Zulip simulacrum (May 03 2021 at 14:23):

er, cc @Pietro Albini (now subscribed)

view this post on Zulip simulacrum (May 03 2021 at 14:24):

no, it doesn't

view this post on Zulip simulacrum (May 03 2021 at 14:24):

gah

view this post on Zulip simulacrum (May 03 2021 at 14:24):

@pnkfelix if you could craft a revert that backports onto beta cleanly, that'd be amazing.

view this post on Zulip pnkfelix (May 03 2021 at 14:26):

okay I think I can do that

view this post on Zulip pnkfelix (May 03 2021 at 14:26):

(It was indeed meant to go into 1.52)

view this post on Zulip pnkfelix (May 03 2021 at 14:33):

But I had another question about this

view this post on Zulip pnkfelix (May 03 2021 at 14:33):

the test suite, I cannot tell from skimming it

view this post on Zulip pnkfelix (May 03 2021 at 14:34):

whehter we tend to rely on the rustdoc test suite alone

view this post on Zulip pnkfelix (May 03 2021 at 14:34):

or if we require some human activity as well before a release, to visually check things out?

view this post on Zulip Joshua Nelson (May 03 2021 at 14:46):

pnkfelix said:

whehter we tend to rely on the rustdoc test suite alone

usually yes

view this post on Zulip Joshua Nelson (May 03 2021 at 14:46):

it's not as complete as it should be, but if your new test passes it should be fine

view this post on Zulip pnkfelix (May 03 2021 at 15:02):

Is the htmlck tool something homegrown?

view this post on Zulip Joshua Nelson (May 03 2021 at 15:09):

yes, almost all the test suites are handwritten

view this post on Zulip Joshua Nelson (May 03 2021 at 15:10):

jsondocck too

view this post on Zulip Joshua Nelson (May 03 2021 at 15:10):

I'm hoping to consolidate htmldocck and jsondocck into the same rust tool at some point (python is hard to test and add dependencies for)

view this post on Zulip Joshua Nelson (May 03 2021 at 15:11):

pnkfelix said:

Is the htmlck tool something homegrown?

if you're looking for documentation there's some at https://rustc-dev-guide.rust-lang.org/rustdoc-internals.html#dotting-is-and-crossing-ts

view this post on Zulip pnkfelix (May 03 2021 at 15:27):

Thanks!

view this post on Zulip pnkfelix (May 03 2021 at 15:27):

I was musing about broader queries about the graph-structure represented by the page, i.e. trying to test the semantic content of the web site

view this post on Zulip pnkfelix (May 03 2021 at 15:28):

(without caring about details like the specific id tags)

view this post on Zulip Joshua Nelson (May 03 2021 at 15:29):

pnkfelix said:

I was musing about broader queries about the graph-structure represented by the page, i.e. trying to test the semantic content of the web site

this would be really nice to have an automated check for

view this post on Zulip Joshua Nelson (May 03 2021 at 15:29):

I think the linkchecker does that for libstd docs

view this post on Zulip Joshua Nelson (May 03 2021 at 15:29):

I wonder if we could extend it to tests too

view this post on Zulip simulacrum (May 03 2021 at 16:17):

@pnkfelix any update on the tidy fixes?

view this post on Zulip pnkfelix (May 03 2021 at 16:19):

They are on their way

view this post on Zulip pnkfelix (May 03 2021 at 16:19):

Had to feed twins lunch

view this post on Zulip pnkfelix (May 03 2021 at 16:20):

(Childcare out sick)

view this post on Zulip pnkfelix (May 03 2021 at 16:25):

@simulacrum hurm. When I leave channel = … as the default, I get: error: no rustfmt binary found in /media/pnkfelix/Rust/rust.git/objdir-512cgu/build/x86_64-unknown-linux-gnu/stage0/bin

view this post on Zulip simulacrum (May 03 2021 at 16:26):

yeah, that's "expected"

view this post on Zulip simulacrum (May 03 2021 at 16:26):

we need to find a better ux for it

view this post on Zulip pnkfelix (May 03 2021 at 16:26):

when I change it to say channel = “beta”, I get fatal: ambiguous argument 'refs/remotes/origin/master..HEAD': unknown revision or path not in the working tree.

view this post on Zulip simulacrum (May 03 2021 at 16:26):

right, you're not in a proper checkout I imagine?

view this post on Zulip pnkfelix (May 03 2021 at 16:26):

what’s a “proper” checkout?

view this post on Zulip simulacrum (May 03 2021 at 16:26):

i.e., refs/remotes/origin != rust-lang/rust

view this post on Zulip pnkfelix (May 03 2021 at 16:27):

origin git@github.com:rust-lang/rust.git (fetch) |

view this post on Zulip pnkfelix (May 03 2021 at 16:27):

(that’s from git remote -v

view this post on Zulip pnkfelix (May 03 2021 at 16:27):

which seems to match what you wrote?

view this post on Zulip simulacrum (May 03 2021 at 16:27):

fwiw I also had this error locally but I chalked it up to my not having an origin setup

view this post on Zulip simulacrum (May 03 2021 at 16:28):

does git rev-list refs/remotes/origin/master..HEAD work for you?

view this post on Zulip pnkfelix (May 03 2021 at 16:28):

to be fair, git branch -a doesn’t list anything for remotes/origin

view this post on Zulip pnkfelix (May 03 2021 at 16:29):

so maybe I need to fetch?

view this post on Zulip pnkfelix (May 03 2021 at 16:29):

simulacrum said:

does git rev-list refs/remotes/origin/master..HEAD work for you?

no it doesn't

view this post on Zulip pnkfelix (May 03 2021 at 16:29):

what is tidy doing here?

view this post on Zulip simulacrum (May 03 2021 at 16:29):

hm

view this post on Zulip simulacrum (May 03 2021 at 16:29):

so this isn't tidy

view this post on Zulip pnkfelix (May 03 2021 at 16:29):

(I’ll try git fetch origin)

view this post on Zulip simulacrum (May 03 2021 at 16:29):

this is bootstrap, trying to figure out what beta you want

view this post on Zulip pnkfelix (May 03 2021 at 16:29):

oh

view this post on Zulip simulacrum (May 03 2021 at 16:29):

i.e., beta.X

view this post on Zulip simulacrum (May 03 2021 at 16:30):

what does ls .git/refs/ say?

view this post on Zulip pnkfelix (May 03 2021 at 16:30):

odd that I feel like I only get this error from doing tidy though

view this post on Zulip simulacrum (May 03 2021 at 16:30):

I think you weren't on channel=beta before?

view this post on Zulip pnkfelix (May 03 2021 at 16:30):

oh, I guess I haven’t tried the full build with channel = beta

view this post on Zulip pnkfelix (May 03 2021 at 16:30):

irght

view this post on Zulip pnkfelix (May 03 2021 at 16:30):

okay

view this post on Zulip pnkfelix (May 03 2021 at 16:30):

git fetch origin fixed it

view this post on Zulip pnkfelix (May 03 2021 at 16:31):

(maybe we need to make our error diagnostic there suggest it, not sure…)

view this post on Zulip simulacrum (May 03 2021 at 16:31):

interesting. oh well. not entirely expected

view this post on Zulip simulacrum (May 03 2021 at 16:31):

(I suspect something may have changed in git, as I've only seen it today)

view this post on Zulip pnkfelix (May 03 2021 at 16:45):

simulacrum said:

pnkfelix any update on the tidy fixes?

PR #84867 (nightly) is r+’ed and the (totally different!) fix to PR #84868 (beta) is pushed to GitHub and awaiting CI.

view this post on Zulip simulacrum (May 03 2021 at 16:48):

Sounds good, thanks


Last updated: Oct 21 2021 at 22:01 UTC