Stream: project-error-handling

Topic: Meeting 2021-04-26


view this post on Zulip Jane Lusby (Apr 26 2021 at 17:18):

Hey all, meeting in 1 hour 40 minutes

view this post on Zulip Jane Lusby (Apr 26 2021 at 17:20):

agenda: https://hackmd.io/@rust-libs/rJSuj3WMO

view this post on Zulip Charles Ellis O'Riley Jr. (Apr 26 2021 at 18:06):

Is the meeting starting?

view this post on Zulip Jane Lusby (Apr 26 2021 at 18:07):

no

view this post on Zulip Jane Lusby (Apr 26 2021 at 18:07):

calendar has it starting in 1 hour

view this post on Zulip Charles Ellis O'Riley Jr. (Apr 26 2021 at 18:07):

2pm?

view this post on Zulip Jane Lusby (Apr 26 2021 at 18:07):

uh

view this post on Zulip Jane Lusby (Apr 26 2021 at 18:07):

I think so

view this post on Zulip Jane Lusby (Apr 26 2021 at 18:08):

view this post on Zulip Charles Ellis O'Riley Jr. (Apr 26 2021 at 18:08):

2pm my time. It's 1:07 now.

view this post on Zulip Jane Lusby (Apr 26 2021 at 18:08):

:+1:

view this post on Zulip Jane Lusby (Apr 26 2021 at 18:57):

3 minute warning

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:00):

alright, meeting time everyone

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:00):

:wave:

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:00):

we will start with reviewing the action items

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:00):

https://hackmd.io/@rust-libs/Hyj7kRSld

view this post on Zulip oliver (Apr 26 2021 at 19:01):

whose here?

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:02):

looks like we have 5, based on the wave

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:02):

also, can someone take minutes?

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:02):

I don't think Sean will be able to attend today

view this post on Zulip oliver (Apr 26 2021 at 19:02):

:owl:

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:02):

gonna start with actions items that have been resolved since the last meeting

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:03):

summarize the discussion in zulip about 3rd party libraries and record the guidance in a new tracking issue about the book

view this post on Zulip oliver (Apr 26 2021 at 19:03):

I'm in a web call atm as well

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:03):

I went ahead and created a tracking issue for the book and added a comment summarizing the zulip thread where we discussed 3rd party deps in the book

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:03):

https://github.com/rust-lang/project-error-handling/issues/38#issuecomment-821673122

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:04):

for now though the book is mostly on hold, with the exception of the diagram that @Charles Ellis O'Riley Jr. is working on

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:04):

but we've put a pause on the book meetings for now since we're not actively working on it for the most part

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:04):

[rust-lang/rfcs#2895] Update object-provider API in generic member access RFC to utilize new dyno design (Jane Lusby) include an example of how dyno could be used for executors

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:04):

I went ahead and updated this RFC, though it's still blocked right now

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:05):

the new plan is to split out the type erasure support code into its own RFC, because its complicating the generic member access RFC significantly

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:05):

and the additional benefit to executors is easier to express if we have it in it's own RFC

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:05):

@Plecra has volunteered to work on that RFC, though I think it will be a little while since they're currently doing finals and stuff afaik

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:06):

so for now that's on the back burner but it is progressing

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:06):

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:06):

I went ahead and made issues for all of thse

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:07):

I think i actually did this a while ago, but only recently cleaned up the bullets on the action items doc

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:07):

follow up on “Fix the Error trait” issue to get the todo list updated issue: [DPC]

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:07):

I also went ahead and finished this one

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:08):

I added a bullet point to the top level comment indicating that its primarily blocked on the backtrace incore impl, and made it clear that the formatting representation issue is no longer blocking

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:08):

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:08):

Now onto open action items

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:09):

During my conversation with Niko about the error handling blog post he mentioned that he felt we could probably resolve the specialization issues with Termination and unwrap

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:09):

or just about anything that specializes Error over Debug

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:09):

the gist of how it works is that we'd have a way to indicate that something specializes for all implementations that apply universally

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:10):

I can't recall the example used to explain how this works, so I can't really explain the exact mechanism at the moment

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:11):

but basically theres a subset of impls of traits that could cause issues with soundness, and we can just not specialize those impls, and then add a lint to warn if an impl is not specializable

view this post on Zulip Jakub Duchniewicz (Apr 26 2021 at 19:11):

if you have time to look for it later I would be glad to learn more

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:11):

sure

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:11):

let me write down a reminder

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:11):

okay

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:12):

I need to go back and lookup the status of the internal representation issue

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:12):

but I think if we can resolve that issue we can possibly bring this up to libs as ready for stabilization, and then add the specialization for Error later once the work around is available

view this post on Zulip Jakub Duchniewicz (Apr 26 2021 at 19:13):

AFAIK it is related to different platforms having different types for exit status code

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:13):

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:14):

so I've had the conversation with niko and am actively working on the rewrite

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:14):

I hope to have it done today or tomorrow

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:14):

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:15):

In the last book meeting on Thursday @oliver @Charles Ellis O'Riley Jr. and I discussed next steps for the diagram and decided that the next step is to port the diagram to graphviz

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:15):

so we can store it in the git repo

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:15):

@Charles Ellis O'Riley Jr. is currently working on that

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:16):

last two action items

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:16):

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:16):

still no progress on either of these items

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:17):

@triagebot end-topic

view this post on Zulip triagebot (Apr 26 2021 at 19:17):

Does anyone have something to add on the current topic?
React with :working_on_it: if you have something to say.
React with :all_good: if not.

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:18):

alright

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:18):

so there are no issues to triage this week

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:19):

so next up I think we should review the project board, which we haven't done in a while

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:19):

https://github.com/rust-lang/project-error-handling/projects/1

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:19):

starting with the proposed column

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:20):

https://github.com/rust-lang/rust/pull/79692

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:20):

we need to reopen this PR which Ashley was working on before he stepped down from libs

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:20):

actually

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:20):

before we get too far

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:20):

is anyone looking for something to work on right now?

view this post on Zulip Jakub Duchniewicz (Apr 26 2021 at 19:21):

well, would be nice to get something to work on while other things are blocked :smile:

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:21):

do you want to take the PR above?

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:21):

I think the implementation is complete, it just needs to be re-opened and pushed through review

view this post on Zulip Jakub Duchniewicz (Apr 26 2021 at 19:22):

yes, sure

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:22):

alright

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:23):

next up is https://github.com/rust-lang/rust/issues/46871

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:23):

this is an issue asking why NoneError doesn't implement Error

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:23):

I think the next steps are that we need to review the discussion so far and leave an explanation of why it doesn't / won't

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:24):

my current understanding is that we don't even want NoneError to exist, because it doesn't represent an error, and we'd prefer to push people towards using ok_or when they want to convert an Option into a Result

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:25):

we might want to touch base with @scottmcm who I think has some opinions on this and mentioned some interaction with the Try trait

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:26):

is anyone interested in following up on this Issue?

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:26):

I will put it as unassigned for now

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:28):

skipping ahead a few items

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:28):

https://github.com/rust-lang/backtrace-rs/issues/397

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:29):

I think this might be a good one to follow up on, basically I'd like to update documentation where we can on how panic = abort interacts with backtraces, likely in the std::backtrace documentation.

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:29):

gonna add this to the action items as unassigned, lmk if you're interested in working on it

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:30):

hmm

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:31):

there's quite a few of these that I should probably update to be unclaimed action items that people can grab if they're looking for work

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:31):

but I'm not sure its an effective use of meeting time to work on this, for now I think I'm gonna stop the project board review, and ill continue this async and post a summary of the work that is available after I'm done

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:31):

@triagebot end-topic

view this post on Zulip triagebot (Apr 26 2021 at 19:31):

Does anyone have something to add on the current topic?
React with :working_on_it: if you have something to say.
React with :all_good: if not.

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:32):

what's up @Charles Ellis O'Riley Jr. ?

view this post on Zulip Charles Ellis O'Riley Jr. (Apr 26 2021 at 19:33):

I didn’t have anything to add. Did I hit a wrong button

view this post on Zulip scottmcm (Apr 26 2021 at 19:34):

Jane Lusby said:

my current understanding is that we don't even want NoneError to exist, because it doesn't represent an error, and we'd prefer to push people towards using ok_or when they want to convert an Option into a Result

With the try_trait_v2 PR I'm currently working on:
image.png
The compiler will stop mentioning NoneError, which will hopefully make these questions go away :smile:

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:35):

Charles Ellis O'Riley Jr. said:

I didn’t have anything to add. Did I hit a wrong button

hehe, yes but thats okay

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:36):

scottmcm said:

Jane Lusby said:

my current understanding is that we don't even want NoneError to exist, because it doesn't represent an error, and we'd prefer to push people towards using ok_or when they want to convert an Option into a Result

With the try_trait_v2 PR I'm currently working on:
image.png
The compiler will stop mentioning NoneError, which will hopefully make these questions go away :smile:

omg that looks sooo good

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:36):

:heart: :heart: :heart:

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:37):

Alright, last section of the meeting is status updates if anyone has anything they'd like to discuss

view this post on Zulip Senyo Simpson (Apr 26 2021 at 19:41):

Question from much earlier but why is the error handling book being put on pause?

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:41):

Senyo Simpson said:

Question from much earlier but why is the error handling book being put on pause?

lack of capacity, nobody has the time to work on it currently

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:41):

I'm focused on the blog post and moving the error trait into core

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:42):

sean has not had time to work on the project group work recently

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:43):

charles is currently working on the diagram, so that's still in progress

view this post on Zulip Senyo Simpson (Apr 26 2021 at 19:43):

Ahh of course. Happy to help where I can there. I'm a bit under the water for the coming month but can dedicate some time

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:43):

and nobody else has expressed interest in writing sections of the book afaik

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:44):

Senyo Simpson said:

Ahh of course. Happy to help where I can there. I'm a bit under the water for the coming month but can dedicate some time

sure! let me know when you have time and we can schedule some time to talk about the plans so far and coordinate the next steps

view this post on Zulip Senyo Simpson (Apr 26 2021 at 19:44):

Perfect :)

view this post on Zulip scottmcm (Apr 26 2021 at 19:45):

Jane Lusby said:

omg that looks sooo good

Send kudos to the folks who invented rustc_on_unimplemented -- it's actually done entirely by a (very, very long) attribute in libcore:
image.png

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:45):

wooow

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:45):

this seems suspiciously like something esteban would have done

view this post on Zulip scottmcm (Apr 26 2021 at 19:46):

That'd be my bet too :)

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:46):

lol

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:46):

alright well I think that's everything for now

view this post on Zulip Jane Lusby (Apr 26 2021 at 19:46):

@triagebot end-meeting

view this post on Zulip triagebot (Apr 26 2021 at 19:46):

Does anyone have something to bring up?
React with :working_on_it: if you have something to say.
React with :all_good: if you're ready to end the meeting.

view this post on Zulip Charles Ellis O'Riley Jr. (Apr 26 2021 at 19:47):

For the sake of expediency, I noticed that svg files were used for The Book, which draw.io can produce but I also just went through the book and noticed a dot directory in which the had code from graphviz. So, to keep things the same, I'll continue a deep dive into graphviz :smile:

view this post on Zulip scottmcm (Apr 26 2021 at 19:57):

Jane Lusby said:

is anyone interested in following up on this Issue?

I replied, FWIW.

view this post on Zulip Jane Lusby (Apr 26 2021 at 23:03):

you're the best, thank you


Last updated: Jan 26 2022 at 14:20 UTC