GitLab Static Site Editor, 9 Nov 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Static Site Editor Weekly - Engineering Focused - 2020-11-09

Description

Weekly sync call of the Static Site Editor group focused on engineering efforts.

A

uh Good morning, everyone, uh this is the static site, editor group engineering, synchronous meeting for november 9th and we'll jump into the agenda.

A

As I find the right tab and start with uh nav and settings updates, so I had a question here for oh actually, first, let's highlight something really important so, prior to hitting record, we had a little bit of admin stuff we were talking about and I forgot to go back up to the highlights from last week. um Darva you dropped this. So do you want to do a voiceover of this.

B

Yes, thank you for giving me the honor congratulations, chad for your huge accomplishment uh on uh having a average duration pipeline, ten minutes uh and meeting under our target under our two minutes target my mistake. So um that is really great to hear and to see eric. Did you want to add on to that.

A

Yeah, I I just for for history's sake to just kind of let it settle in when I joined the company. um The the handbook group was tasked with um improving pipeline durations because it was uh so long that it was prohibiting collaboration and causing merge, merge conflicts just in the process of waiting for pipelines. To finish and stuff like you know, there was a mess, it was over 30 minutes most of the time and um we had a goal that seemed ambitious but achievable, and this is a really just a really big moment.

A

We've we've reached that goal. We've exceeded that goal and we've done really everything we can within the structure of our current pipeline to to make it a better experience for um for everybody to contribute to the handbook. So it's a very, very big milestone and um given what we know about middleman, it's it's not likely to get much faster than this, so we can really hang our hat on that and say this is um this: is this is the best pipe? This is the most optimized pipeline we can have and yeah.

C

Like I said there, the really the the actual pipeline run is now not even that much uh percentage-wise that we can do about it. It's just a giant repo and the best thing we could do is move all of the big files and images out of the repo to somewhere else, and then you know prune the repo size down and those are all really disruptive things that would. We would need a lot of people on board and a lot of changes to processes and stuff like that. So that's really not viable. At this point, yeah.

A

And- and it's by no means we're, you know as a company we're not going to stop working on the handbook in general, but um in light of the recent consolidation of editor groups, we no longer will have that as a category that we dedicate time to so this is a great way to kind of wrap up our work on it and we'll continue to iterate, as we can we'll continue to be able to consult as people as subject matter, experts in the handbook uh and help people as needed, but this is really going to make it so that we have a much more.

A

I don't know optimized foundation for people to to build on, and then you know, ultimately, we can still investigate the transition to front man or we can work on removing those um those large files, large assets and then get this thing under five minutes someday. But for now we should definitely celebrate this achievement.

C

And darva, like I mentioned in a 101, I'm still anxious to hear who is going to get responsibility for this, because I I don't think that answer could be nobody. If it's nobody is. I think defect on me cause I've rewritten it all. Over the past year.

B

I will say that in the cases, if anything, you know if they're any bugs or defects or any challenges there um and on the other teams, we basically said you know if the team that is now focusing on the new feature, if they don't have the ability or the skills, yet the backup would be the original uh engineer.

B

But I think, as you know, the weeks progress we'll get more clarity on some of these things. I wouldn't.

D

B

Things are exactly final. I I have, it seems like there's been a change almost every week, so I think that maybe in about three weeks or so just a little bit patience, but if something happens and something breaks or something's needed, you'll probably be called.

C

And that's fine and the the absence of information is perfect. That's what we know right now, just wondering.

A

Cool, so um moving on to tactical updates. uh We only have a couple members of the engineering team here today, but um we'll see what we can get here. um Vasily you had been working on or had been assigned to the settings in nav work to disable features on the left-hand nav. So I was just curious what you've uncovered there? I I see your update that um you had a prototype that disabled, so I'm just curious.

A

If you could give me a sense of like the complexity there and what work is remaining, how how hard is it going to be to do the rest of them? I think.

D

That the first step is the hardest one, because the rest of the these blocks, the side menus- should be done kind of similarly to the first like tasks that I'm working on right now so, but I found some it's like not the problem, but something that I'm needs. Somebody's help its permissions, because gitlab has many permissions and we don't. We not only need to disable this uh tab. We also need to prevent users from accessing the pages that this tab represents.

D

So that's I like in my prototype, it works, but I want to find like a person who knows more about it and ask this person to review this code so yeah and after it's done, I think the other other like items in the menu can be removed.

D

Quite similarly, so we can use this merge request as an example for the future changes.

A

Sounds great thanks? um Do you need any help tracking down somebody to help review that or you kind of know where to head? I.

D

Think I'll check like who reviewed the previous uh mrs, like that you know that were adding these new items and I'll pick somebody from there and see if it was a right person or not great.

A

uh Chad, you've got the next one on analytics.

C

Sure this is one of the ones I didn't make as much progress on last week that I would have liked, but that's mainly because I'm trying to make sure I do this. The right way, there's emerging standards around how these metrics should be named and in ways that they're captured and show up automatically, maybe in reports along with other sorts of metrics like that, and so a lot of the weaker spit.

C

Like reading docs, I set up one-on-ones with two people that I knew had knowledge with it and uh got a lot of good info there and took notes on it, and then I have, but they didn't have the answers to some of this very specific questions I had so. I have a slight question pending. I think I found maybe the right people to ask now so the actual implementation of it. I know eric you wanted. You asked if we could get in for this release.

C

The actual implementation is going to be really straightforward, but since this data you know it like can lag weeks, you have to get the metrics in and wait for it to show up. I want to make sure I have the right naming and stuff and not have to revisit it or have misnamed metrics get out there. So that's.

A

Right totally understand yep yeah, very understandable, great um jacques, not on a call, but I saw he left some updates on um outstanding, mrs for the image upload and previewing in the static site editor. um So we can just follow up async. On that I know it's it's getting close he's got some unless chad, maybe you've got insight into it from following along, but it seems like it was getting there. I.

C

I've seen him doing work on them and uh yeah. He didn't ask me for a review yet so I was not looking at it. No problem.

A

And um I guess the other thing uh is the an update from my side on the proposal for the wysiwyg editor. The rich text editor architecture that we're working on enrique um continued to refine the technical proposal last week, and I think our goal is to let the dust settle on this editor group realignment and then uh bring this in the form of a technical review, a peer review of some kind.

A

You know circulate for feedback among engineers, as well as I'm going to be presenting an opportunity canvas just to summarize the business case and the product justification to the product team. um Probably in the next uh I'd like to schedule that next week, maybe and from there we can move forward with confidence.

A

uh Obviously the editor group would be a great one to present this to.

A

I was um speaking with uh kai about um about all this and I think it might be helpful to schedule uh maybe a deep dive on the architecture and and everything related to the web ide, so that our group can get up to speed on on what goes on under the hood there, as well as uh have enrique, do a deep dive on on his proposal and present that to the current editor group and have those engineers really understand where, where he's been, what he's been thinking about and where he's landed so I'll, try and get those on the calendar.

A

I'll chat with roman about that and um and see what we can do in in the next couple weeks to get those scheduled.

C

On that topic, uh when I attended the editor groups meeting last week, I brought this up and uh said that I was real happy with the the direction of the architecture and all of the thought that we've put into it, and there were some similar sentiments uh like from people like paul that have reviewed some of the architecture we've done in the past, and you know it's.

C

We put a lot of good thought into everything. We've done so uh I think it will go well. I'm looking forward to us working together on it.

A

Yeah I mean I, I couldn't be more happy with how how thorough and and I don't know how complementary this is to just everything that we've we've been tasked with now like. I think it fits really well into uh an editor group's strategy. It fits well into the gitlab ecosystem and the product itself and obviously will help the static site editor grow as soon as we're done implementing it.

A

So that's the update there and I don't have like concrete issues yet to for implementation, but I'm hopeful that we get these reviews underway and maybe by sometime within 13.7, we can start actually dedicating some engineering time to building out the foundations.

A

uh Moving on to handbook and growth marketing chad's got, you've got the only um agenda item there.

C

uh Yes, so towards the end of last week, marcel van rivenden- I don't know if I'm saying his name right, he uh pulled me on an in on an mr saying. I tried to get the www repo running under the pod. Initially I was like, I haven't even used get pod, yet I I really don't know.

C

I told him to do something uh which was actually the wrong advice, because it had been so long since I used rvm. But after I looked at it, I was like this is pretty cool, um so I ended up working on friday and a little bit this weekend getting the www working under get pod, which I think is going to be really cool and I'm going to also follow up by adding some documentation in the handbook and in under the dot folder it's kind of vague, where it actually belongs.

C

So I'm going to put it in one and cross reference it, but this should be a really easy way for people to work on the about site, both marketing and handbook, without having to set up their machine. It's really cool.

A

That's great, uh you know I had been talking with um marcel about this, uh maybe three four weeks ago and he was showing me some of the get pod work and I was like I was sharing some challenges.

A

We had had with um a release post manager who had an old version of ruby installed or something like that, and it caused a bunch of conflicts, and I was like you know for that specific use case it'd be really cool to have good pod for the www repo and um he went and tried to do this and then it sounds like you helped bring it over the finish line, which is very exciting.

C

Yeah there are a couple of things like I was like. Oh, how am I going to make both it's actually a mono repo, it's two sites how's that going to work and they actually had features to do exactly what I wanted. So I I'm pretty impressed with it. It's a cool product, nice.

A

uh When this is all uh buttoned up, uh look just keep me in the loop with this and when it's buttoned up we'll make an announcement uh or you can make an announcement, it doesn't need to come from me, but we should announce this widely so that people know that the integration is is available and how to use it. I think it's great.

C

Yep- and I think it fits in well with maybe wanting to have less effort expended towards helping people work in the www yeah.

A

Cool well excellent, um moving on to get lab docs uh craig thanks for sitting with us for 30 minutes. While we got to.

B

A

B

A

uh We probably should have started with this, so you could drop, but thank you. uh We can uh get an update on the offline documentation stuff on vasily. If you want to run that.

D

uh So yeah last week I had a conversation with palu and he suggested so we provided him a proposal that uh created for the offline documentation, so we checked it together and it seems to be a good way to move forward.

D

However, one of the challenges we have right now is the size of the generated docs, because, like on my local machine, when I try to see how much space it consumes, it was like 600 megabytes for text files with images which is too much and we want to reduce it to sub 100 megabytes like fewer is better and that's what I'm working on right now I created an issue where I added the steps that we can introduce to reduce the size and yeah. That's what I was working on and there are some questions from craig.

D

Can you verbalize them.

E

Sure the first one was um one of the items that you mentioned was changing the sidebar, the I'm guessing the um global nav on the left hand, yeah right make it so that it's not loading everything uh and um you know by reducing the size of the page.

E

If that were a change that were to be made, is that literally only on the offline version of this? Or would this be something that we would need to roll across? All of the code.

D

uh Right now we can choose, because we can uh control what kind of docks we can produce. We can produce it's the same way as we do right now for docsgitlab.com, but for omnibus package we can use a different settings and produce the different uh types of dogs. So in this case it will be still global navigation, but it will have some restrictions, so you will access like everything inside of the section you are in, but for accessing other section. You need to click and reload the page.

E

And I personally, I think that that's actually okay, it's already a step up from what they have, which is nothing yes, if they have to do a little bit of click and load. I'm I'm okay with that in that environment, if it were to change the actual overall experience, just like you, click that and then you're doing another http refresh to get something. I think people would scream.

E

What do you mean it's not actually being responsive um so great. Thank you very much for that uh and I'll put that in the issue. You know kind of set up conversations whatever, um and then um I saw your other item which is I had asked. Do we pick one? Do we pick all, and I see your recommendations, you just take them off.

D

Yes, it's better to take them all, because each step like, for example, reducing the global navigation size. It's like saves, like 50 percent, then uh making siblings for images it's like around 30 megabytes or so so. Each step introduces some savings for the total space. So so far I made it as small as 26 megabytes. It should be good enough, like to convince baloo like to include this uh to omit this package.

E

One other quick question and I'll put it in the thing and quick when, uh when this actually builds everything and does stuff, is it only uh building and distributing those items that are actually used within the documentation? Or is it actually pulling some things from the from some images from the repo that may not yet may not be used anymore, but are still just actually listed in the repo, basically old images that we're not using anymore.

D

I think it uses only the current images, so it takes the content of the docs folder from the gitlab repository, so everything that's there and it should be like some actual documentation pages. It's should be the same content. Basically, we convert markdown files to html and uh copy all of the images.

E

Okay, so we're not actually building a package that has images that we're no longer using, but just happen to be sitting in the repo but nobody's touching them.

D

I think that can happen, but probably it's not the case. Okay, I I don't know like if we track the not longer used images in the docs folder like when we update documentation for the gitlab repository. Do we also check if there is no images that are still in the wrapper but has no links to them?

D

If you do that, then we should not have any kind of like extra images. Okay,.

C

I wouldn't guess that unless there's a process or something automated to explicitly find them, the answer is yes, there are some. Oh they're. Definitely.

E

In there I could, I could swear, I think.

D

Okay, so yeah in this case, it might be also like one of the ways to improve. We can simply remove these images and it will affect the total size as well.

E

um Yeah uh we uh uh we do chores on a regular basis, I'll see if we can do some things to kind of clean things up. But um it's thank you for putting the letting me know about.

D

That sure it will be useful thanks.

C

Just a random question on that that came up and maybe you've already thought about it, but is there any concerns about the technical debt? That's going to be introduced by having the nav behave, two different ways for different people or increased chance of regressions in one versus the other? If it's less frequently used like the omnibus one, the need for automated tests or something.

D

Yeah, that's a good question because yeah maybe it will be like two separate ways how to build it, but I think uh when it's ready for testing we can see. Maybe the second approach that produces a less html code is can be used for docs as well. I don't know how much user experience it affects because I tried it like on my local machine. It's kind of you. If you really use these tabs to open like without reloading the page, you will notice the difference, but I was actually before.

D

I was not aware that you can open them, so I just clicked and reload the page like explicitly so. For me it was like no significant change, but yeah. We can test.

D

A

Cool, so we're we're a little over time. um I don't have a lot of updates other than I'm working on planning, um specifically the settings in nav, I'm trying to consolidate that work and um make it more clear, whoa. We are actually working on in the near future.

A

What the scope of work for our engagement on the settings and network is, and um that will likely take the form of an epic similar to our category maturity, epic, with the static site editor, I'm trying to build some hierarchy there, working on that this week and planning for 13.7 with a consolidated, editor group, so more on that in the product and design sync this evening, um my time and obviously any questions you have related to that you can send them in slack, I'm always available we'll have um a planning issue shortly and some uh put some thoughts on what work is carrying over from 136 who's, going to pick that up as we consolidate editor groups and hand, hand-off work and try and tie up some loose ends before focusing on the you know, the other work that we have on our plate.

A

So that's that'll be my focus this week and uh probably have more detail on that by midweek. If not friday,.

C

I have a quick question: what is the day I would have to have the the project navigation metrics merged in order for it to be in this release.

A

So the event would be something that we would need to have in the or the whole point is to get it in the packaged release. So it would be the I believe the 17th is when they cut that build might be good to have it merged by the 16th, because I don't know what time of day and what time zone.

A

I think you're muted, but.

C

Sorry I'll I'll try to make sure to have it ready and I should get the answers I need by then.

A

Cool appreciate it all right. Well I'll, stop the recording thanks, everyone for joining and we'll chat soon.

E

D