Home
Contribute Contact Us Browse all meetings
Home Contribute Contact Us Browse all meetings
CHAOSS / Evolution Working Group / 22 Feb 2022
CHAOSS / Evolution Working Group / 22 Feb 2022
Previous Meeting Next Meeting
⏯
Add meeting Rate page Subscribe
youtube image
►
From YouTube: CHAOSS Evolution Working Group 2/22/22

Description

Links to minutes from this meeting are on https://chaoss.community/participate.

A

But I'm hitting the record button, and this is the evolution working group for february 22nd um 2022 as it turns out. I don't know about the rest of y'all, but we're gonna get snow again tomorrow. All the snow from last week was melted.

B

We are enjoying the wind, it's like 20 miles.

A

An hour yeah yeah, well nebraska, you don't have any real hills in front of omaha to protect you from that. I just posted the meeting minutes in there. So there was a meeting january 25th, so we nobody met last time. In my absence.

C

I think vanada and I were the only ones to show up right.

A

Yeah.

C

And we just said.

D

Okay, fair enough.

C

We weren't gonna do work so.

B

We started working on our own stuff.

A

All right.

D

So.

A

Today is, uh and we have a deadline coming up, so I think our first order of business is, um I don't see kevin today.

A

uh Let's see we had started working on change, request commits and we had planned to then start working on change, request reviews, but if my memory serves, we might be in much better position to try to finish change request commits than we are to start with change, request reviews.

A

Let me just check here: okay, and actually they both they all. They both have oops change, requests, reviews change, request, commits.

A

I think we have do we want, let's I'm going to suggest that we start with change, request, commits and see if we can finish that and then move on to change, request reviews. If we have time.

B

Seems uh uh what I recall there was just image missing, which you have already added, I think yeah, maybe just we can go through it and.

A

Close it yeah, let's uh just add a chat. Why don't we all take a minute to read through it? um I will pause the recording while we're reading through it, just because nobody wants to watch that oops, not the sharing the recording and since we're starting to discuss it back to recording.

A

Oh.

A

I'm going through and accepting changes, others have provided.

A

Did you find the bitter chilling? No? Okay? No, it's fine! I think.

C

I have.

A

It I'll I'll look, I don't think they have um commits for pull requests or average quits per pull request, but I'm not positive.

A

I think this visualization will be able to be made bigger when we put it natural.

B

Mark down as well yep, we can directly save from the api and put it on the github, then will be much better visible.

A

All right, if anybody dislikes my changes, reject them or accept them, but I think that we are pretty darn close to a metric.

C

This.

C

I pulled that from that sigils page sean does that, would that be better than or do you want the dashboard.

A

um uh The sigils pages, if it's clearly um there that I mean whatever you think, is the best representation. These.

C

Are the things that are included in that panel that they have like the number of author there you go, which seems to number.

A

Of modified files yeah I mean it commits yeah. This panel seems to be exactly the right link to me.

C

We should I put that up under tools providing instead of references, yeah.

A

I think that's tools providing yes, yes, that I do think that.

C

I think this looks good.

D

Yep yeah.

A

Okay, that's wrong. One.

A

So what is uh have we created the issue or do we need to create the issue.

C

I don't think we created it yet.

A

Okay,.

A

See we did somebody want to take an action item to pull this together and submit it as a candidate metric. I can do that. Okay, so before march 1st, I think yep I'll do it.

B

It's not.

D

Oh, thank you. Bernard.

A

Oops, ah I can't seem to operate all right so now, moving on to change, request, reviews.

A

You.

A

You.

A

So I was editing instead of suggesting sorry about that.

A

I just made this sentence here separate from the one.

A

You.

A

Okay,.

A

Okay,.

A

I seriously don't understand how it decides, something is url and I put the mark down I'll fix it.

B

What is time period in the filters just that um time period for the review.

A

Well, for example, I I might be interested in knowing over the course of like for a project- that's been around a while. I might be interested only in the last two years of data to understand the change in practice. Yeah.

B

So maybe.

A

Yeah.

B

I got your point yeah. Just time period was not clear to me like time period for review or something more elaborate time period. Okay,.

B

So change the question before.

A

Oh yeah, that's.

A

I think we both made the same change at the same time. That's like catch that one. Oh yeah,.

C

John, I have a question um it seems like I could be wrong, but it seems like this actually might be a filter on the one we were just doing.

A

Change request commits.

C

Like you would you would see the commit or the whole change request, and then that would be a filter, as did it also go through this formal process.

A

No, I don't think you're completely wrong.

C

Maybe not that metric, maybe a different metric, but.

A

It's I mean the kind of the way that it ends up. Let me see we can find out.

A

Yeah, I got a bunch of these things. I got to close okay.

C

I guess it depends on how the api provides that information.

A

Right, yeah, let me well let me uh let's see.

A

All right, here's a bigger one. So here this particular change request has 133 commits and, and so this many files are changed and so like for the change request commits I'm actually the way the review happens is I'll like go here and make a comment, and it won't relate to a commit so much as it will to a file, because because any one commit could include more than one file.

B

Yep, so it's a two layer of commits you are commenting on the request, like pull request, entire pull, request that discussion and on the each individual file. You are commenting on each individual line. You are commenting, so it maps more closely.

A

To files than it does to commits.

C

So this this metric we're working on if you go back to the original page.

A

Oh.

C

No sorry of the of your pr when it listed the review.

A

Oh yes,.

C

Like so wherever it listed, the review yeah yeah.

D

Right here, yeah, what we're.

C

Looking for in this metric is that there are specific reviewers that have been requested to review the commit right.

A

Correct and and then what is the product of those those reviews? Okay,.

C

So.

A

Do they go through and make comments and suggest changes, or do they accept the pull request, as it is? Okay, um okay, yeah? So it depends a great deal on the practices of a project most projects when they accept commits or except for change, requests from outside their core.

A

Do a review and ask for some changes just to con create. You know, maintain some sort of consistency. It's hard to know the consistency patterns if you're new to a project but.

C

Oh, go.

A

Ahead yeah, so there it'll it could go through several review cycles, which is the other thing it's like. I could review this pull request and ask for some changes: okay and some somebody they would change their fork or their branch and then submit those changes, and I would be prompted to then review this pull request again with the changes that were made at my request, so it doesn't yeah it just it does not. It definitely doesn't really map to commits, because these commits can touch many of the same files.

C

So it's more than just a toggle on and off a toggle of. Did you formally request a review, or did you not.

A

It's uh did was there a review? Really? Okay, if you requested it, but there wasn't a review. There's you there's like no review data.

C

Okay,.

A

That would be an event, but it wouldn't be part of like this metric, but you can see there's 22 files changed, but 133 commits so yeah like I said, the review takes place at the file level. So when we get to change request files, I think there's a relationship to change, request, reviews, okay, but I don't think so for commits.

C

I was just thinking of that first aggregator. We have that says what percentage of change requests are formally reviewed using platform features, so you would so in essence, it sounds like from that aggregator you're collecting all all of everything and then pulling that out.

A

um We're collecting everything about the the change request review, which includes some metadata about the review as a whole. Okay,.

C

Okay, which.

A

Is in here- and it also includes metadata about like when I made a comment back here- it'll include this comment in the review.

C

Okay,.

A

Like any comments, I make about specific pieces of code get incorporated into the analysis of a pull request review and then, whether or not it was approved, this is the next stage.

C

Okay, so this metric is kind of got a lot to it. It's kind of complicated.

A

A little bit yeah yeah, but github and gitlab do have apis that let you gather all of the data that you see here so but pull request yeah so but pull request commits is like a piece of data. That's different that than the review. It's not really linked to the review at all, but the files are when we do pull request files. I think there is a direct connection, because every file could get a comment on it or not.

C

Do we make that more explicit in this metric, or do people generally know that.

A

I don't know um where would we.

B

Make it what would we say here? So it's a file change lines added line deleted there. We can add a review.

A

Okay, um so I think up in the description.

A

I think does that help to make things a little bit more clear.

B

Last point in this paragraph is confusing, like the reasoning behind the change request being closed without getting merged that will be already covered in the comments. If there are comments on it.

A

Yes,.

A

Yes and each I mean whether it's accepted change is requested or it's declined um that may be. That may be it's probably made clear in the top level comments, but sometimes when things aren't merged, there are details about why it wasn't merged that are in the file level comments, okay, um so, for example, if if somebody decided to change the persistence paradigm for an open source project and that probably covered 35 files, there may be specific comments about why the why this doesn't work for each file.

A

That would help a new contributor understand why their change request was rejected. If you were trying to develop community, you would you would comment on the pieces specifically um so that you like, when you just close it without very thorough comment, of course, that's more discouraging to newcomers if you close it with several comments on the files, I think then, at least as a new contributor. I understand why you didn't merge my pull requests.

A

And really one of the things I look at is you know the amount of file level commenting because I think it's indicative of a more genuine process than just saying nothing.

B

It's more a thorough review than just yeah.

A

Maybe yeah and and sometimes there's pull requests that are just off the wall and you do make just a general comment.

A

But we're not going to do this because.

B

We have four minutes left. I think both the metrics are good to go.

A

Yeah I wish we had visualizations, but I don't think we do.

A

I mean, I know, auger doesn't have any visualizations and I.

A

I can accept my own comments, but it's probably better practice. If someone.

B

Else.

A

Accepts or rejects them.

B

I've gone through here so you can accept it or I can accept that. So it's gonna speed things along.

A

uh

A

Burning so uh so, I think we have two that are ready for release. That's great, especially since we didn't meet last night kind of like fire lit a feet under our fire with a fire under our feet, something like that. Yep um yeah.

D

Yeah, I don't even that means and since you're going to do one, do you mind just do it yep yep I'll? Do it no worries I'll? Do it.

B

It's thank you doing one or two is the same effort exactly.

D

It's like.

A

You banan, we appreciate it, okay! Well, I uh excellent meeting everyone be proud of your meeting selves and um are there any other? I don't think we have time to cover really any of the other things on the agenda. So I think I will bid you ado prior to the community.

B

Meeting sean, do you have a minute after this meeting? I just need to talk one thing with you all right: yeah yeah see ya. Okay,.

A

Bye, let me stop the recording okay.