Vote on primordials in error path #1158

aduh95 · 2022-01-26T20:59:42Z

I've found the subject of voting in TSC comes quite often, and we often have recurrent discussions on what tool to use. I've decided to take upon me the creation of a tool to help us vote using CLI and GitHub UI. It might end up being a terrible idea, or a genius one, but at least we would have try. If it's successful, we can talk about officially adopt this tool for Node.js related votes.

The vote options were presented in #1104 (comment).

Use one of the following methods to case a vote (or abstain):

Vote using npx+gh (requires Node.js 16+, and git)

Requires gh (GitHub CLI tool) to be installed, and logged in to your GitHub account.

npx --package=@aduh95/caritat voteOnGitHub https://github.com/nodejs/TSC/pull/1158

If you want to abstain, you can use the --abstain flag (it will create a ballot with all the scores set to 0):

npx --package=@aduh95/caritat voteOnGitHub https://github.com/nodejs/TSC/pull/1158 --abstain

Vote using npx (requires Node.js 16+, and git)

# Cloning and pushing using SSH
npx --package=@aduh95/caritat voteUsingGit -r git@github.com:nodejs/TSC.git \
  -b vote-primordials-error-path -p votes/primordials-error-path

# Cloning and pushing using HTTPS
USERNAME=<your-github-handle> \
npx --package=@aduh95/caritat voteUsingGit -r https://$USERNAME@github.com/nodejs/TSC.git \
  -b vote-primordials-error-path -p votes/primordials-error-path --handle "$USERNAME"

# To abstain, use the flag --abstain
npx --package=@aduh95/caritat voteUsingGit -r git@github.com:nodejs/TSC.git \
  -b vote-primordials-error-path -p votes/primordials-error-path --abstain

Vote using sh (requires git, and OpenSSL CLI)

Caution: LibreSSL CLI is not compatible, please use OpenSSL 1.1.1+.

curl -LOJ https://github.com/stduhpf/caritat/raw/HEAD/sh/voteUsingGit.sh
curl -LOJ https://github.com/stduhpf/caritat/raw/HEAD/sh/encryptBallot.sh
chmod +x encryptBallot.sh
sh ./voteUsingGit.sh \
  <your-github-handle> \
  git@github.com:nodejs/TSC.git \
  vote-primordials-error-path \
  votes/primordials-error-path
rm encryptBallot.sh
rm voteUsingGit.sh

Vote using PowerShell (requires git, and OpenSSL CLI)

Caution: LibreSSL CLI is not compatible, please use OpenSSL 1.1.1+.

Set-ExecutionPolicy Bypass -Scope Process -Force
[System.Net.ServicePointManager]::SecurityProtocol = [System.Net.ServicePointManager]::SecurityProtocol -bor 3072
$tmpFile = $(New-TemporaryFile | Rename-Item -NewName { $_ -replace 'tmp$', 'ps1'} -PassThru)
$encryptBallot = $tmpFile -replace '\\[^\\]+$','\encryptBallot.ps1'
((New-Object System.Net.WebClient).DownloadString('https://raw.githubusercontent.com/stduhpf/caritat/HEAD/sh/voteUsingGit.ps1')) | Out-File -FilePath $tmpFile
((New-Object System.Net.WebClient).DownloadString('https://raw.githubusercontent.com/stduhpf/caritat/HEAD/sh/encryptBallot.ps1')) | Out-File -FilePath $encryptBallot
& $tmpFile `
  <your-github-handle> `
  git@github.com:nodejs/TSC.git `
  vote-primordials-error-path `
  votes/primordials-error-path
Remove-Item $encryptBallot
Remove-Item $tmpFile

Vote using the web

Fill in the ballot at https://stduhpf.github.io/caritat/#https://github.com/nodejs/TSC/pull/1158. It will copy the encrypted ballot to your clipboard.
Paste the encrypted ballot in https://github.com/nodejs/TSC/new/vote-primordials-error-path/votes/primordials-error-path. It's recommended to use <your-github-handle>.json for the file name.
Commit directly to the vote-primordials-error-path branch.

To abstain, set all scores to 0 before encrypting the ballot.

This will encrypt a file containing your preferences, file that will be decrypted when doing the counting in two weeks or so.

The counting will be done using the Condorcet method.

Refs: #1104 (comment)

Bugs should be sent to: https://github.com/stduhpf/caritat

Trott · 2022-01-26T21:38:39Z

@aduh95 Is there a public source code repository for the tool? It seems to be https://github.com/stduhpf/caritat but with some additional files. Maybe there's a tag/branch/fork I should be looking at? I did the tsc compilation step, but I still don't have (for example) app.js.

aduh95 · 2022-01-26T22:28:16Z

@aduh95 Is there a public source code repository for the tool? It seems to be https://github.com/stduhpf/caritat but with some additional files. Maybe there's a tag/branch/fork I should be looking at? I did the tsc compilation step, but I still don't have (for example) app.js.

Yes, it is https://github.com/stduhpf/caritat. app.js has been deleted, it was a left-over from a not-so-clean local repo. I've published v0.2.0 that fixes that, repo and npm package should be in sync now.

Trott · 2022-01-26T22:30:41Z

Yes, it is https://github.com/stduhpf/caritat. app.js has been deleted, it was a left-over from a not-so-clean local repo. I've published v0.2.0 that fixes that, repo and npm package should be in sync now.

Awesome, thanks.

MylesBorins · 2022-01-27T17:39:15Z

this is very very cool. Will dig more into implementation in a bit

mhdawson · 2022-02-02T17:41:41Z

@aduh95 its not immediately obvious if we put higher scores for options we want my guess is that if I wanted to rank them from highest preference to lowest I could use 0-5 for the scores where 5 is the one I prefer the most? Can you confirm that is correct?

aduh95 · 2022-02-02T18:54:01Z

@aduh95 its not immediately obvious if we put higher scores for options we want my guess is that if I wanted to rank them from highest preference to lowest I could use 0-5 for the scores where 5 is the one I prefer the most? Can you confirm that is correct?

@mhdawson that would works, absolutely. As long you put integers, you can use whichever scale that's more convenient for you (negative numbers also works btw): the largest score for the option(s) you prefer the most (5 in your example), the lowest score for the option(s) you prefer the less (0 in your example).

mhdawson · 2022-02-02T21:34:27Z

@aduh95 thanks

aduh95 · 2022-02-03T18:42:31Z

@nodejs/tsc Reminder that there's an ongoing vote on this PR, you are invited to participate; see the OP for instructions, don't hesitate to ask questions if you have any.
Vote is closing next week if we have quorum.

aduh95 · 2022-02-03T18:48:56Z

@jasnell you mentioned on the issue you don't have a preference but there's one option you don't like. FWIW you can cast a vote that set all the options to 0 (i.e. abstain) except for the issue you don't like that you can set to -1, that will be taken into account when counting the ballots.

RaisinTen · 2022-02-05T13:20:55Z

Running into an EACCES when I'm trying to vote using sh - it commits an empty file for me. Reported to nodejs/caritat#3.

RaisinTen · 2022-02-05T15:18:52Z

I used the web ui to generate the encrypted ballot this time and after pushing I realize that it doesn't include the author info (reported in nodejs/caritat#5). Adding it manually.

tniessen · 2022-02-16T12:42:04Z

Use primordials in the error path only if there are no significant perf regression.

I am confused by the winning option. Does it explicitly support primordials? In other words, could this option be rephrased as Use primordials in the error path unless there are significant perf regressions but not as Do not use primordials in the error path if there are significant perf regressions?

RaisinTen · 2022-02-16T13:17:14Z

To my understanding, it means - if someone volunteers to add primordials to error paths, we are likely to accept it if and only if there are no significant perf regressions.

aduh95 · 2022-02-16T13:57:20Z

And the reciprocal would be – if someone opens a PR to remove primordials from error paths, it would be rejected, unless said primordials was causing a measurable perf regression.
If we do other votes related to primordials in the future, I agree we should try to make that wording less confusing.

tniessen · 2022-02-16T14:42:05Z

we are likely to accept it if and only if there are no significant perf regressions.

Does that mean writing benchmarks for error paths?

Trott · 2022-02-16T15:00:54Z

Given that this vote has been completed, should this be moved out of draft mode and landed?

aduh95 · 2022-02-16T15:13:48Z

Given that this vote has been completed, should this be moved out of draft mode and landed?

The next step would be for me (or anyone that wants to beat me to it) to decrypt all the ballots, then I think it can be merged. Maybe we should also create a .json file similar to 2021-09-29-0.json so participation in this vote can be counted by the automation.

RaisinTen · 2022-02-16T15:21:59Z

Does that mean writing benchmarks for error paths?

I don't think there is any other way to make sure that the performance is not degraded. So IIUC, PRs like nodejs/node#41702 that attempt to add primordials to an error path would be blocked because we still don't have benchmarks testing the error path.

ljharb · 2022-02-16T15:22:30Z

I’m confused; what about code not in the error path?

tniessen · 2022-02-16T15:25:35Z

Does that mean writing benchmarks for error paths?

I don't think there is any other way to make sure that the performance is not degraded.

That's... interesting.

I’m confused; what about code not in the error path?

That was out of scope for this vote... I believe the intention is for more votes to follow.

targos · 2022-02-16T15:55:58Z

Does that mean writing benchmarks for error paths?

I don't think there is any other way to make sure that the performance is not degraded.

That's... interesting.

Interesting indeed. I thought that we were voting on the error paths because performance regressions that only happen there are not problematic.
My understanding was that we would only block a change in an error path if that somehow affects the performance of non-error paths.

Trott · 2022-02-17T04:09:43Z

Concurring with @targos and @tniessen in that I don't think we've ever been terribly concerned about error-path performance. I could be mistaken, but it seems like an odd place to put in extra effort on performance.

votes/2022-02-21-0.json

mhdawson

LGTM

mhdawson · 2022-02-21T21:50:54Z

Concurring with @targos and @tniessen in that I don't think we've ever been terribly concerned about error-path performance. I could be mistaken, but it seems like an odd place to put in extra effort on performance.

For me I found it a bit difficult to be voting about the error path without having discussed if we should use for the main paths. As such I voted more conservatively than I might otherwise have.

votes/2022-02-21-0.json

RaisinTen

LGTM

BridgeAR · 2022-02-23T21:39:52Z

I agree that we have not been concerned about the error cases performance a lot and as such it's probably fine. I just wonder why we added that option to our vote if we do not plan to actually handle it different from "Make the code in the error path full tamper proof by using primordials" as it seems like the overall vote would have been different in that case.

Trott

Tested this with the tooling and it worked great. Thanks.

aduh95 added 2 commits January 26, 2022 18:57

vote on primordials in the error path

3864fa0

vote from aduh95

acbbd8b

This comment was marked as off-topic.

Sign in to view

targos and others added 2 commits January 27, 2022 15:17

vote from targos

8c05246

vote from MylesBorins

62d53c6

Trott and others added 4 commits January 28, 2022 14:29

vote from Rich Trott

4ddea50

vote from Mary Marchini

5dbb431

vote from Matteo Collina

74b8ffa

vote from Gireesh Punathil

5bfb7a8

This comment was marked as off-topic.

Sign in to view

vote from cjihrig

4391dfc

aduh95 mentioned this pull request Feb 3, 2022

vote: future of primordials #1104

Closed

jasnell and others added 4 commits February 3, 2022 11:22

vote from jasnell

d0ada1f

vote from Michael Dawson

28fb209

vote from Robert Nagy

a619c98

vote from RaisinTen

1466e68

vote from RaisinTen

d39d109

RaisinTen and others added 2 commits February 5, 2022 20:52

vote from RaisinTen

a93f03e

vote from Tobias Nießen

3931bcc

close vote and aggregate results

3e99994

aduh95 commented Feb 21, 2022

View reviewed changes

votes/2022-02-21-0.json Outdated Show resolved Hide resolved

Add missing EOL

3b980f9

aduh95 marked this pull request as ready for review February 21, 2022 21:49

mhdawson approved these changes Feb 21, 2022

View reviewed changes

gireeshpunathil reviewed Feb 22, 2022

View reviewed changes

votes/2022-02-21-0.json Show resolved Hide resolved

aduh95 commented Feb 22, 2022

View reviewed changes

votes/2022-02-21-0.json Show resolved Hide resolved

Add back Gireesh vote data

42b7c02

gireeshpunathil approved these changes Feb 22, 2022

View reviewed changes

RaisinTen approved these changes Feb 22, 2022

View reviewed changes

Trott approved these changes Feb 23, 2022

View reviewed changes

Trott merged commit 8fcaf34 into main Feb 23, 2022

Trott deleted the vote-primordials-error-path branch February 23, 2022 22:12

aduh95 mentioned this pull request Feb 24, 2022

doc: add primordials guidelines nodejs/node#38635

Closed

tniessen mentioned this pull request Mar 10, 2022

[meta] settle on a voting solution #1165

Closed

This was referenced Apr 12, 2024

V8 private symbols nodejs/primordials-use-cases#4

Open

doc: specify to use primordials when coding nodejs/node#52558

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Vote on primordials in error path #1158

Vote on primordials in error path #1158

aduh95 commented Jan 26, 2022 •

edited

Loading

Trott commented Jan 26, 2022

aduh95 commented Jan 26, 2022

Trott commented Jan 26, 2022

This comment was marked as off-topic.

MylesBorins commented Jan 27, 2022

This comment was marked as off-topic.

mhdawson commented Feb 2, 2022

aduh95 commented Feb 2, 2022

mhdawson commented Feb 2, 2022

aduh95 commented Feb 3, 2022

aduh95 commented Feb 3, 2022

RaisinTen commented Feb 5, 2022 •

edited

Loading

RaisinTen commented Feb 5, 2022 •

edited

Loading

tniessen commented Feb 16, 2022

RaisinTen commented Feb 16, 2022

aduh95 commented Feb 16, 2022

tniessen commented Feb 16, 2022

Trott commented Feb 16, 2022

aduh95 commented Feb 16, 2022

RaisinTen commented Feb 16, 2022

ljharb commented Feb 16, 2022

tniessen commented Feb 16, 2022

targos commented Feb 16, 2022

Trott commented Feb 17, 2022 •

edited

Loading

mhdawson left a comment

mhdawson commented Feb 21, 2022 •

edited

Loading

RaisinTen left a comment

BridgeAR commented Feb 23, 2022

Trott left a comment

Vote on primordials in error path #1158

Vote on primordials in error path #1158

Conversation

aduh95 commented Jan 26, 2022 • edited Loading

Trott commented Jan 26, 2022

aduh95 commented Jan 26, 2022

Trott commented Jan 26, 2022

This comment was marked as off-topic.

MylesBorins commented Jan 27, 2022

This comment was marked as off-topic.

mhdawson commented Feb 2, 2022

aduh95 commented Feb 2, 2022

mhdawson commented Feb 2, 2022

aduh95 commented Feb 3, 2022

aduh95 commented Feb 3, 2022

RaisinTen commented Feb 5, 2022 • edited Loading

RaisinTen commented Feb 5, 2022 • edited Loading

tniessen commented Feb 16, 2022

RaisinTen commented Feb 16, 2022

aduh95 commented Feb 16, 2022

tniessen commented Feb 16, 2022

Trott commented Feb 16, 2022

aduh95 commented Feb 16, 2022

RaisinTen commented Feb 16, 2022

ljharb commented Feb 16, 2022

tniessen commented Feb 16, 2022

targos commented Feb 16, 2022

Trott commented Feb 17, 2022 • edited Loading

mhdawson left a comment

Choose a reason for hiding this comment

mhdawson commented Feb 21, 2022 • edited Loading

RaisinTen left a comment

Choose a reason for hiding this comment

BridgeAR commented Feb 23, 2022

Trott left a comment

Choose a reason for hiding this comment

aduh95 commented Jan 26, 2022 •

edited

Loading

RaisinTen commented Feb 5, 2022 •

edited

Loading

RaisinTen commented Feb 5, 2022 •

edited

Loading

Trott commented Feb 17, 2022 •

edited

Loading

mhdawson commented Feb 21, 2022 •

edited

Loading