[2026-11-19] eagle: work/zoom-transcripts/2026-10-19 Michal Smaga's Personal Meeting Room QvxllpCuTTSLGEZp-R6jQw/transcript.json work/zoom-transcripts/2026-10-19 Michal Smaga's Personal Meeting Room QvxllpCuTTSLGEZp-R6jQw/transcript.md work/zoom-transcripts/2026-10-27 Apple Weekly Team Meeting - Zoom ID -I3aVYDWTkKZ_VTRLdjGnA/transcript.json work/zoom-transcripts/2026-10-27 Apple Weekly Team Meeting - Zoom ID -I3aVYDWTkKZ_VTRLdjGnA/transcript.md work/zoom-transcripts/index.json work/zoom-transcripts/sync-status.json
This commit is contained in:
1 parent
8802ea61a1
commit
bc1c74b0e5
6 files changed
+5
-1061
No files matched your search
-650
@@ -23,656 +23,6 @@
|
||||
"time": "18:10:09",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Uh, I think I'm, I'm like… done reviewing the first one about the pixel. I had the issue that I wasn't able to reproduce it, like, I was trying to kill the WebKit process."
|
||||
},
|
||||
{
|
||||
"time": "18:10:22",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Mm-hmm."
|
||||
},
|
||||
{
|
||||
"time": "18:10:23",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Do you have any, like, extra steps?"
|
||||
},
|
||||
{
|
||||
"time": "18:10:27",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Uh, so generally it fails from the second try on this exact test page that is attached there."
|
||||
},
|
||||
{
|
||||
"time": "18:10:37",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "So after, after I, I killed the, the, the…"
|
||||
},
|
||||
{
|
||||
"time": "18:10:37",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Okay."
|
||||
},
|
||||
{
|
||||
"time": "18:10:41",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Web extension, web process. I should reload the page?"
|
||||
},
|
||||
{
|
||||
"time": "18:10:47",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yes, in general."
|
||||
},
|
||||
{
|
||||
"time": "18:10:49",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Do I need to do it, like, quickly, or do I need to, like…"
|
||||
},
|
||||
{
|
||||
"time": "18:10:53",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Um…"
|
||||
},
|
||||
{
|
||||
"time": "18:10:53",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Wait a bit."
|
||||
},
|
||||
{
|
||||
"time": "18:10:55",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Usually… yeah. I don't know, within one second, just kill the process, then reload the page."
|
||||
},
|
||||
{
|
||||
"time": "18:11:02",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:11:03",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "If it doesn't work from the first try, you just give it and reload again, and usually from second attempt, it does, uh… Uh, hang the… the extension. So it's not… it's not the genuine issue, so it's the…"
|
||||
},
|
||||
{
|
||||
"time": "18:11:20",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, artificial."
|
||||
},
|
||||
{
|
||||
"time": "18:11:22",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "artificial reproducing step that we didn't enable with the feature flag of the fix that went before, but it should be still here, and we… and allow to… To produce it. Uh, that's… So generally, yeah, I think you've seen the comment I posted to… to the… Project communication task. And the data, the data analysis."
|
||||
},
|
||||
{
|
||||
"time": "18:11:52",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, yeah, yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:11:52",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "So… so generally there is, uh… yeah, it's kind of heavy codecs and cloud work, and, um… I cannot… Oh. I cannot tell that I understand, like, 100% what is happening beneath, uh, under the hood, and… Um… Generally, we have reproduced most of the… of the stall issues, uh, with some… Thank you. my display on black. Wow. I just need to switch off something, I guess."
|
||||
},
|
||||
{
|
||||
"time": "18:12:35",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "I can hear you, I cannot see you."
|
||||
},
|
||||
{
|
||||
"time": "18:12:36",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Hello? Yeah, I cannot see you, because… Give it a minute. Yeah. No. Fuck. And… Good morning, Mark. Sorry. One moment, Amy. I need to burn off something."
|
||||
},
|
||||
{
|
||||
"time": "18:13:10",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Did your laptop display died?"
|
||||
},
|
||||
{
|
||||
"time": "18:13:13",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Not yet, I guess. I hope. Uh, here… Here we are. Uh-huh. So, we do. Yeah. All right, I'm back. And… Oh, yeah, I just got some. Testing. Fairness around here. Uh, where was I? Yeah, the issues. Uh, so generally, there's four issues in the work, and uh… One is behind the disabled feature flag. Because it reloads the page, and I don't think it's optimal. And probably a couple of more go next, and we'll leave. the smallest groups for the next iteration, I guess. So here we have to, uh… To expand the diagnostics. So generally, to fulfill the privacy triage, uh, I was requested to… To reduce what we… actually submit, and the decision was to merge the events into a single event. It should not generally harm our, uh… our data. We can lose something in case the browser crashes, or if it's reloaded. But I guess it's acceptable at this point, since we get these isolated groups already. And we just need some proofs. for the remaining and the pixels. Prior to this."
|
||||
},
|
||||
{
|
||||
"time": "18:15:02",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yes. So, so the first…"
|
||||
},
|
||||
{
|
||||
"time": "18:15:04",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Pixel fire for the pixel, yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:15:06",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "First PR is the pixels, yeah? So the one from the PR."
|
||||
},
|
||||
{
|
||||
"time": "18:15:09",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yes, so the suggestion is to improve the diagnostics first, and then the stacked ones go onto it."
|
||||
},
|
||||
{
|
||||
"time": "18:15:18",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, and what is the status of the other, like, PRs, the follow-up ones, with the fixes?"
|
||||
},
|
||||
{
|
||||
"time": "18:15:24",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Uh, so we have the… the recovery. And… First one is InitLoss. And InitLoss includes P1 and P2 as inputs. As I remember. Yeah, and P4, that's not enabled at the moment. It's, uh, 7379. It should be… Should be generated already. For review. And another one is 7459. It's also open. And… Well, generally, should also be pretty much ready, so the… There is integration tests, integration test harness that's supposed to only be run once to reproduce the issue."
|
||||
},
|
||||
{
|
||||
"time": "18:16:38",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Mm-hmm."
|
||||
},
|
||||
{
|
||||
"time": "18:16:39",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "In, like, lab condition."
|
||||
},
|
||||
{
|
||||
"time": "18:16:40",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Can you say. Which one was the last PR?"
|
||||
},
|
||||
{
|
||||
"time": "18:16:45",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "The last one, uh, 7459. This one."
|
||||
},
|
||||
{
|
||||
"time": "18:17:00",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, yeah, yeah, I got it, yeah, the network one, yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:17:11",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "So the first, that is 7379, it should… So basically, we guess that it may be not an issue. And it's maybe an issue of slow. Slow loading and the fix itself is triggering the background will be load with the navigation itself. Because, uh, the initial… initial startup, or return from background, maybe not, uh, warming up the web view. Or, um… Or our extension. And, uh… The situation with P4 is that we may need to reload the page after, you know, after the extension is loaded, so it is applied to the page. And, uh, I was planning to investigate it further. Is there a way to run the… Plug in after the page is already loaded, maybe send another page load event. And see if it… if it just works. Uh, but the reload is there. And the feature flag."
|
||||
},
|
||||
{
|
||||
"time": "18:18:36",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "And do you think, like, uh… Are we able to detect the fourth one without the feature flag? Just to get the pixels about, like, how…"
|
||||
},
|
||||
{
|
||||
"time": "18:18:48",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Just to get the pixels… yeah, I guess so. I guess so we can make the pixel."
|
||||
},
|
||||
{
|
||||
"time": "18:18:59",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, because it would be get, like. It's good to get the signal, like, how many of those cases they potentially would be fixed by the future flag."
|
||||
},
|
||||
{
|
||||
"time": "18:19:11",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yeah, I agree."
|
||||
},
|
||||
{
|
||||
"time": "18:19:12",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, but I mean, like, reloading the page is kind of, like, a risky thing, yeah, it can be disruptive, and… The question is, like, how frequently it would be."
|
||||
},
|
||||
{
|
||||
"time": "18:19:22",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Mm-hmm."
|
||||
},
|
||||
{
|
||||
"time": "18:19:23",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, and probably we would also have, like, the question is, like, how much confidence that they are actually, like, experiencing the problem. Just not to do the the, like, the reload. When there's no… not an issue for that."
|
||||
},
|
||||
{
|
||||
"time": "18:19:57",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Okay, and the next one is the network. Network process crash. Uh, it's a bit trickier. I… I didn't finish going through it. Actually… But, uh… So general state of it, it's kind of reproduced on the integration test harness. And the fix… Seems… to fix it. So, again, it's kinda… not… non-proved concepts and kind of speculative fixes. And we are hoping to… So first of all, those should be feature flagged, just in case anything goes wrong with it. You can private selectors or whatever are disturbing. The overall extension flow, so we could disable it remotely, and uh… Yeah, so again, I'm… Oh… Not 100% sure on those, but how it is reproduced in the test harness and by the automation. It seems we could use those fixes. Another one pending, it's not opened yet, is the base URL thing. And, uh… And the speculative steps to reproduce is, uh, if we reload the, um… if we update the web extension. it changes the base URL. Because we don't have this table base irrelevant. Those… It was flagged in the early research phase. Uh, what I'm not sure with it, uh, does it harm any… Can it bring any privacy or security concerns with it if we…"
|
||||
},
|
||||
{
|
||||
"time": "18:22:09",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "But which base? I, I don't, I don't get it, like, fully, like, with me."
|
||||
},
|
||||
{
|
||||
"time": "18:22:13",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "The extension was 0. So it's UUID that's generated for the extension every time it's launched. And when it changes, it brings some… Disturbance into the web extension."
|
||||
},
|
||||
{
|
||||
"time": "18:22:28",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Okay."
|
||||
},
|
||||
{
|
||||
"time": "18:22:29",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "And if we set it fixed as a fixed UID. For our web extension, it generally follows the WebKit rules as I understand it and may fix another part of the failures."
|
||||
},
|
||||
{
|
||||
"time": "18:22:44",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Mhmm. But it's only only on update, like, when we uninstall and install again, right?"
|
||||
},
|
||||
{
|
||||
"time": "18:22:45",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "And, uh… Yeah, but I, uh… I can guess it may be some significant portion of users that receive the service related to the app updates. As, uh, as it was, uh, as it was… described and issued initially."
|
||||
},
|
||||
{
|
||||
"time": "18:23:14",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Okay, so it's like, uh… 3… Three PRs with fixes. and potentially four."
|
||||
},
|
||||
{
|
||||
"time": "18:23:24",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Mmhm. network, the P1 one, uh, base URL. One for the loading. And, uh, yeah, potentially the fifth one. So, yeah, currently there is four, uh, four fixes in… in my… On my… work."
|
||||
},
|
||||
{
|
||||
"time": "18:23:53",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "So, do you have any, like, doubts, like, that, like, some of the fixes are, like, too risky, and, like, for example, the reload one? you would be not too sure about, like, shipping that one? Because I'm just, like, thinking…"
|
||||
},
|
||||
{
|
||||
"time": "18:24:05",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Uh, Zip?"
|
||||
},
|
||||
{
|
||||
"time": "18:24:07",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "thinking about, like, probably we can have the questions from Russell or someone, like. How much more time we can spend here, like, just to, like, focus on the, like, bare minimum, because probably, like, just to, like… better calculate the effort that we want to put into it."
|
||||
},
|
||||
{
|
||||
"time": "18:24:27",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yeah, generally, we already had this question by Russell, as I mentioned, and I tried to answer it today."
|
||||
},
|
||||
{
|
||||
"time": "18:24:30",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, yeah, yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:24:37",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "No, we've seen it. Uh, so the issue here is, uh, that we cannot distribute. the failures. You cannot tell there is one main… ah, yeah, there is another. Another one… And… Great deeds. Yeah, another genuine one that is kind of proved is window git close. So the original. Window and tab merging management, uh, fix it, uh… It didn't cover every… Every issue with the pin tab, and there is still one more remaining, uh, for the window closing, and, uh, yeah, should be a short one. Uh, it should cover the genuine issue of closing in tab using the web extension."
|
||||
},
|
||||
{
|
||||
"time": "18:25:31",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "But it's only McCoy, right?"
|
||||
},
|
||||
{
|
||||
"time": "18:25:32",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "I don't know. Yeah, my question. So, yeah, that is, uh, for… Four, one, two, three, one, two, three, four main groups that are targeted here, one requiring the page reload. And the P3 that, uh, without… Context without view, it may be the pin tab. And we have one iOS of 0.03%. macOS at around 6%, but it may be wrong. And we have the other group that cannot be considered at the moment. Maybe it gets fixed with the four fixes, or maybe we receive some more data. So I try to express this point in the conversation task. But if the hypothesis are correct, we should cover most of the cases."
|
||||
},
|
||||
{
|
||||
"time": "18:26:37",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "So, do you estimate to get them done this week?"
|
||||
},
|
||||
{
|
||||
"time": "18:26:37",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "So, yeah. Yeah, so the point here is to timebox it to…"
|
||||
},
|
||||
{
|
||||
"time": "18:26:45",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "To merge it this week, and so we can, like, get it internally next week, right?"
|
||||
},
|
||||
{
|
||||
"time": "18:26:45",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Two currents. Yeah. Yeah, so if we… if we get it to the release… I mean, the data analysis, the data attribution extension, and those fixes, if we get it to internal. Today and tomorrow. Uh, we can reason about it on… Next week."
|
||||
},
|
||||
{
|
||||
"time": "18:27:11",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Okay, so you would mean to merge it to… Release branches."
|
||||
},
|
||||
{
|
||||
"time": "18:27:17",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yes, if it's possible."
|
||||
},
|
||||
{
|
||||
"time": "18:27:22",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "So, the two first PRs, like, one for the telemetry for the pixels, and the second with the P1 and P2, right?"
|
||||
},
|
||||
{
|
||||
"time": "18:27:31",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Um… It depends, I guess, if we can do it, if it's possible, I would give a go for all of them. Again, behind the feature flags."
|
||||
},
|
||||
{
|
||||
"time": "18:27:43",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, let's try. So, like, I'm… I'm… I'm… I'm done, like… reading and commenting for the first PRs, there are a couple of, like. leftovers for the dead code, so please check the comments. I'll try to, like, try forcing to reproduce it."
|
||||
},
|
||||
{
|
||||
"time": "18:27:55",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Mmhm."
|
||||
},
|
||||
{
|
||||
"time": "18:28:01",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "So so I get it and then I can move to the next PR. I can focus today on that. So we have the the the chart."
|
||||
},
|
||||
{
|
||||
"time": "18:28:08",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Okay, good. And when we have those reviewed, we can open the exception and try to push it to the internal."
|
||||
},
|
||||
{
|
||||
"time": "18:28:18",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "I mean, like, if it's, like, Monday, Tuesday, you can do it probably, like, for free."
|
||||
},
|
||||
{
|
||||
"time": "18:28:24",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Mm-hmm, okay."
|
||||
},
|
||||
{
|
||||
"time": "18:28:24",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "If it's later than tomorrow. Yeah. Okay. Okay, so let's try to do that."
|
||||
},
|
||||
{
|
||||
"time": "18:28:30",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Okay, I'll… I'll keep you pinged, yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:28:31",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "And… Please ping me when you have the following one, because I want to, like…"
|
||||
},
|
||||
{
|
||||
"time": "18:28:37",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Mm-hmm."
|
||||
},
|
||||
{
|
||||
"time": "18:28:38",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Schedule the time for them, uh, like, in between my, my work, just to, just to not, like… Delay you. Uh, yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:28:47",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yeah, I guess it's… I guess it may take some time, maybe the rest of your day today, and those two PRs. Could feel it."
|
||||
},
|
||||
{
|
||||
"time": "18:28:58",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, so I'll focus on that, not to block you, so let's push for the first two ones, and yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:29:04",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Mmhm."
|
||||
},
|
||||
{
|
||||
"time": "18:29:04",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "And the others, like, later in the week."
|
||||
},
|
||||
{
|
||||
"time": "18:29:07",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Mm-hmm."
|
||||
},
|
||||
{
|
||||
"time": "18:29:08",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, I agree."
|
||||
},
|
||||
{
|
||||
"time": "18:29:09",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "So, and yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:29:10",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Just to try to get them out as soon as possible to get some results."
|
||||
},
|
||||
{
|
||||
"time": "18:29:16",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "And yes, since it's highly speculative, I would give it another mid-mortem. So yeah, time box it to this week. I'd say to the first half of this week. Because there is not much work really to be done about it. Just, uh… Just so you know, I have, like, a huge testing harness work tree, and I'm trying to split the parts from it that reproduce one another issue and make it into…"
|
||||
},
|
||||
{
|
||||
"time": "18:29:48",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "No doubt."
|
||||
},
|
||||
{
|
||||
"time": "18:29:53",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "into digestible pull requests. But generally, most of them are kind of reproduced in the lab. But we have no any proofs that they actually match the production issues because they are based on the event sequence that we get."
|
||||
},
|
||||
{
|
||||
"time": "18:30:08",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:30:14",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "We got to produce the event sequence that we are receiving, but we still cannot reason about the real issues leading to those. So, yeah, that's the state. All right, thank you. I won't take more of your time."
|
||||
},
|
||||
{
|
||||
"time": "18:30:31",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "So, after this week, you would put the, like, you would have the mid-mortem this week, or, like, next week after the…"
|
||||
},
|
||||
{
|
||||
"time": "18:30:41",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Next week, so let's try to get it out and have some data. The meeting is on Tuesday, right? So we could have the data analyzed and, uh, have a mid-mortem. and decide on the next steps whether we need. Keep on going, or we are satisfied with those, or we just give up and…"
|
||||
},
|
||||
{
|
||||
"time": "18:31:06",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah. But do you think, like, on Tuesday, you will have, like, any meaningful data? It will be, like, just out for two days. Right?"
|
||||
},
|
||||
{
|
||||
"time": "18:31:13",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Uh, yes, we should, uh, since we… we received most of the current data by Tuesday. And it didn't change significantly throughout the week. I think there is a lot of issues are reported."
|
||||
},
|
||||
{
|
||||
"time": "18:31:36",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Okay."
|
||||
},
|
||||
{
|
||||
"time": "18:31:37",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "And there's more than enough data to analyze. So, yeah, Tuesday should be should be fine to have a big more time. I'll probably update the comments. Highlighting this."
|
||||
},
|
||||
{
|
||||
"time": "18:31:51",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Okay, sounds good. Okay, so I'm back to reviewing it, and, like, let's… Probably tomorrow you can, like, will you be on the OE weekly?"
|
||||
},
|
||||
{
|
||||
"time": "18:32:05",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yes, tomorrow, tomorrow, tomorrow, tomorrow is Tuesday, yes, I think so."
|
||||
},
|
||||
{
|
||||
"time": "18:32:11",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, so we can recap the plan and see. So this is, like, your top priority, right? You're focusing on…"
|
||||
},
|
||||
{
|
||||
"time": "18:32:20",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yeah, yeah. So, I mean, there is, uh, there is enough work for me outside of the top priority. Like where you are and pushing other stuff forward. But, yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:32:39",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Makes sense. Cool. Are you blocked by anything, or just, like, this is, like… The general, like, work is enough for you."
|
||||
},
|
||||
{
|
||||
"time": "18:32:49",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Alright. Blocked with the data missing. And, uh, so… Only the pull requests are on the plate. Bottle making it. And, uh, yeah, the gaps between the releases, so we… keep perceiving it and. Analyzing how it goes."
|
||||
},
|
||||
{
|
||||
"time": "18:33:11",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Have you assigned me to all of the PRs? Because, uh, I see them…"
|
||||
},
|
||||
{
|
||||
"time": "18:33:15",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Not yet, I guess. Let me assign the network process. I didn't go through it myself yet, but yeah, I've assigned it to you."
|
||||
},
|
||||
{
|
||||
"time": "18:33:29",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Yeah, I think, like, something is broken with the…"
|
||||
},
|
||||
{
|
||||
"time": "18:33:33",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yeah, yeah, I'll… I'm on it."
|
||||
},
|
||||
{
|
||||
"time": "18:33:36",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "No, no, like, I mean, like, uh, I'm browsing through the, the, this PR, these are the PRs, and, like, I'm filtering by the assigned to me. And I only picked it up."
|
||||
},
|
||||
{
|
||||
"time": "18:33:47",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Ah, assignees. I need to assign them. We generally don't assign the PRs, just set a reviewer."
|
||||
},
|
||||
{
|
||||
"time": "18:33:54",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "I don't know, like, you have the filter, like, review requested."
|
||||
},
|
||||
{
|
||||
"time": "18:34:00",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Mm-hmm."
|
||||
},
|
||||
{
|
||||
"time": "18:34:01",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "They can…"
|
||||
},
|
||||
{
|
||||
"time": "18:34:01",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yeah, let me just send. Yes, I need. and requests. Yeah, so there is three pull requests."
|
||||
},
|
||||
{
|
||||
"time": "18:34:07",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Can you… can you see? So those are, like, open PRs from you all, but you have the filter for review, and it's, like, awaiting review from…"
|
||||
},
|
||||
{
|
||||
"time": "18:34:14",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yep."
|
||||
},
|
||||
{
|
||||
"time": "18:34:18",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Me? I only get two of them."
|
||||
},
|
||||
{
|
||||
"time": "18:34:24",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Okay. 7, 4, 5, 5, 7."
|
||||
},
|
||||
{
|
||||
"time": "18:34:26",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Which is weird."
|
||||
},
|
||||
{
|
||||
"time": "18:34:29",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Oh, 7376 is missing for some reason. Probably you gave a review for it already. And it's waiting for…"
|
||||
},
|
||||
{
|
||||
"time": "18:34:38",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Okay."
|
||||
},
|
||||
{
|
||||
"time": "18:34:40",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "For the resolutions. Angus."
|
||||
},
|
||||
{
|
||||
"time": "18:34:43",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Okay, so maybe it's like, yeah."
|
||||
},
|
||||
{
|
||||
"time": "18:34:46",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yeah, so currently two more in stock."
|
||||
},
|
||||
{
|
||||
"time": "18:34:50",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Okay. So it's it's in the stack, so I can get them from from the list. Right?"
|
||||
},
|
||||
{
|
||||
"time": "18:34:55",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yep."
|
||||
},
|
||||
{
|
||||
"time": "18:34:56",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Okay, cool. Can I help you in any other way?"
|
||||
},
|
||||
{
|
||||
"time": "18:34:59",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Review stuck. Yes, stuck, stuck, stuck. Create stuck. Create stuck. And now those are in the stacks. Okay, good."
|
||||
},
|
||||
{
|
||||
"time": "18:35:15",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Do you need any other help?"
|
||||
},
|
||||
{
|
||||
"time": "18:35:18",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Not at the moment. I'm going to update the first one, telemetry one. And give you a ping, and uh… And when I complete the third one, I'll ping you as well, and probably… Probably the window did close, should be an easy one for today. As well. If it sounds good."
|
||||
},
|
||||
{
|
||||
"time": "18:35:46",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Okay. Okay, so I'm on the PRs. Please check the comments for the first one. I'm all moving for the second one."
|
||||
},
|
||||
{
|
||||
"time": "18:35:58",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Sure. Okay, thank you."
|
||||
},
|
||||
{
|
||||
"time": "18:36:03",
|
||||
"speaker": "Michal Smaga",
|
||||
"text": "Thank you."
|
||||
}
|
||||
]
|
||||
}
|
||||
-130
@@ -15,133 +15,3 @@ source: "https://us01mynotes.zoom.us/QvxllpCuTTSLGEZp-R6jQw"
|
||||
- **18:10:01 · Michal Smaga:** Good.
|
||||
- **18:10:01 · Alex Martemyanov:** I'm pretty good. Jumping between windows. Yes.
|
||||
- **18:10:09 · Michal Smaga:** Uh, I think I'm, I'm like… done reviewing the first one about the pixel. I had the issue that I wasn't able to reproduce it, like, I was trying to kill the WebKit process.
|
||||
- **18:10:22 · Alex Martemyanov:** Mm-hmm.
|
||||
- **18:10:23 · Michal Smaga:** Do you have any, like, extra steps?
|
||||
- **18:10:27 · Alex Martemyanov:** Uh, so generally it fails from the second try on this exact test page that is attached there.
|
||||
- **18:10:37 · Michal Smaga:** So after, after I, I killed the, the, the…
|
||||
- **18:10:37 · Alex Martemyanov:** Okay.
|
||||
- **18:10:41 · Michal Smaga:** Web extension, web process. I should reload the page?
|
||||
- **18:10:47 · Alex Martemyanov:** Yes, in general.
|
||||
- **18:10:49 · Michal Smaga:** Do I need to do it, like, quickly, or do I need to, like…
|
||||
- **18:10:53 · Alex Martemyanov:** Um…
|
||||
- **18:10:53 · Michal Smaga:** Wait a bit.
|
||||
- **18:10:55 · Alex Martemyanov:** Usually… yeah. I don't know, within one second, just kill the process, then reload the page.
|
||||
- **18:11:02 · Michal Smaga:** Yeah.
|
||||
- **18:11:03 · Alex Martemyanov:** If it doesn't work from the first try, you just give it and reload again, and usually from second attempt, it does, uh… Uh, hang the… the extension. So it's not… it's not the genuine issue, so it's the…
|
||||
- **18:11:20 · Michal Smaga:** Yeah, artificial.
|
||||
- **18:11:22 · Alex Martemyanov:** artificial reproducing step that we didn't enable with the feature flag of the fix that went before, but it should be still here, and we… and allow to… To produce it. Uh, that's… So generally, yeah, I think you've seen the comment I posted to… to the… Project communication task. And the data, the data analysis.
|
||||
- **18:11:52 · Michal Smaga:** Yeah, yeah, yeah.
|
||||
- **18:11:52 · Alex Martemyanov:** So… so generally there is, uh… yeah, it's kind of heavy codecs and cloud work, and, um… I cannot… Oh. I cannot tell that I understand, like, 100% what is happening beneath, uh, under the hood, and… Um… Generally, we have reproduced most of the… of the stall issues, uh, with some… Thank you. my display on black. Wow. I just need to switch off something, I guess.
|
||||
- **18:12:35 · Michal Smaga:** I can hear you, I cannot see you.
|
||||
- **18:12:36 · Alex Martemyanov:** Hello? Yeah, I cannot see you, because… Give it a minute. Yeah. No. Fuck. And… Good morning, Mark. Sorry. One moment, Amy. I need to burn off something.
|
||||
- **18:13:10 · Michal Smaga:** Did your laptop display died?
|
||||
- **18:13:13 · Alex Martemyanov:** Not yet, I guess. I hope. Uh, here… Here we are. Uh-huh. So, we do. Yeah. All right, I'm back. And… Oh, yeah, I just got some. Testing. Fairness around here. Uh, where was I? Yeah, the issues. Uh, so generally, there's four issues in the work, and uh… One is behind the disabled feature flag. Because it reloads the page, and I don't think it's optimal. And probably a couple of more go next, and we'll leave. the smallest groups for the next iteration, I guess. So here we have to, uh… To expand the diagnostics. So generally, to fulfill the privacy triage, uh, I was requested to… To reduce what we… actually submit, and the decision was to merge the events into a single event. It should not generally harm our, uh… our data. We can lose something in case the browser crashes, or if it's reloaded. But I guess it's acceptable at this point, since we get these isolated groups already. And we just need some proofs. for the remaining and the pixels. Prior to this.
|
||||
- **18:15:02 · Michal Smaga:** Yes. So, so the first…
|
||||
- **18:15:04 · Alex Martemyanov:** Pixel fire for the pixel, yeah.
|
||||
- **18:15:06 · Michal Smaga:** First PR is the pixels, yeah? So the one from the PR.
|
||||
- **18:15:09 · Alex Martemyanov:** Yes, so the suggestion is to improve the diagnostics first, and then the stacked ones go onto it.
|
||||
- **18:15:18 · Michal Smaga:** Yeah, and what is the status of the other, like, PRs, the follow-up ones, with the fixes?
|
||||
- **18:15:24 · Alex Martemyanov:** Uh, so we have the… the recovery. And… First one is InitLoss. And InitLoss includes P1 and P2 as inputs. As I remember. Yeah, and P4, that's not enabled at the moment. It's, uh, 7379. It should be… Should be generated already. For review. And another one is 7459. It's also open. And… Well, generally, should also be pretty much ready, so the… There is integration tests, integration test harness that's supposed to only be run once to reproduce the issue.
|
||||
- **18:16:38 · Michal Smaga:** Mm-hmm.
|
||||
- **18:16:39 · Alex Martemyanov:** In, like, lab condition.
|
||||
- **18:16:40 · Michal Smaga:** Can you say. Which one was the last PR?
|
||||
- **18:16:45 · Alex Martemyanov:** The last one, uh, 7459. This one.
|
||||
- **18:17:00 · Michal Smaga:** Yeah, yeah, yeah, I got it, yeah, the network one, yeah.
|
||||
- **18:17:11 · Alex Martemyanov:** So the first, that is 7379, it should… So basically, we guess that it may be not an issue. And it's maybe an issue of slow. Slow loading and the fix itself is triggering the background will be load with the navigation itself. Because, uh, the initial… initial startup, or return from background, maybe not, uh, warming up the web view. Or, um… Or our extension. And, uh… The situation with P4 is that we may need to reload the page after, you know, after the extension is loaded, so it is applied to the page. And, uh, I was planning to investigate it further. Is there a way to run the… Plug in after the page is already loaded, maybe send another page load event. And see if it… if it just works. Uh, but the reload is there. And the feature flag.
|
||||
- **18:18:36 · Michal Smaga:** And do you think, like, uh… Are we able to detect the fourth one without the feature flag? Just to get the pixels about, like, how…
|
||||
- **18:18:48 · Alex Martemyanov:** Just to get the pixels… yeah, I guess so. I guess so we can make the pixel.
|
||||
- **18:18:59 · Michal Smaga:** Yeah, because it would be get, like. It's good to get the signal, like, how many of those cases they potentially would be fixed by the future flag.
|
||||
- **18:19:11 · Alex Martemyanov:** Yeah, I agree.
|
||||
- **18:19:12 · Michal Smaga:** Yeah, but I mean, like, reloading the page is kind of, like, a risky thing, yeah, it can be disruptive, and… The question is, like, how frequently it would be.
|
||||
- **18:19:22 · Alex Martemyanov:** Mm-hmm.
|
||||
- **18:19:23 · Michal Smaga:** Yeah, and probably we would also have, like, the question is, like, how much confidence that they are actually, like, experiencing the problem. Just not to do the the, like, the reload. When there's no… not an issue for that.
|
||||
- **18:19:57 · Alex Martemyanov:** Okay, and the next one is the network. Network process crash. Uh, it's a bit trickier. I… I didn't finish going through it. Actually… But, uh… So general state of it, it's kind of reproduced on the integration test harness. And the fix… Seems… to fix it. So, again, it's kinda… not… non-proved concepts and kind of speculative fixes. And we are hoping to… So first of all, those should be feature flagged, just in case anything goes wrong with it. You can private selectors or whatever are disturbing. The overall extension flow, so we could disable it remotely, and uh… Yeah, so again, I'm… Oh… Not 100% sure on those, but how it is reproduced in the test harness and by the automation. It seems we could use those fixes. Another one pending, it's not opened yet, is the base URL thing. And, uh… And the speculative steps to reproduce is, uh, if we reload the, um… if we update the web extension. it changes the base URL. Because we don't have this table base irrelevant. Those… It was flagged in the early research phase. Uh, what I'm not sure with it, uh, does it harm any… Can it bring any privacy or security concerns with it if we…
|
||||
- **18:22:09 · Michal Smaga:** But which base? I, I don't, I don't get it, like, fully, like, with me.
|
||||
- **18:22:13 · Alex Martemyanov:** The extension was 0. So it's UUID that's generated for the extension every time it's launched. And when it changes, it brings some… Disturbance into the web extension.
|
||||
- **18:22:28 · Michal Smaga:** Okay.
|
||||
- **18:22:29 · Alex Martemyanov:** And if we set it fixed as a fixed UID. For our web extension, it generally follows the WebKit rules as I understand it and may fix another part of the failures.
|
||||
- **18:22:44 · Michal Smaga:** Mhmm. But it's only only on update, like, when we uninstall and install again, right?
|
||||
- **18:22:45 · Alex Martemyanov:** And, uh… Yeah, but I, uh… I can guess it may be some significant portion of users that receive the service related to the app updates. As, uh, as it was, uh, as it was… described and issued initially.
|
||||
- **18:23:14 · Michal Smaga:** Okay, so it's like, uh… 3… Three PRs with fixes. and potentially four.
|
||||
- **18:23:24 · Alex Martemyanov:** Mmhm. network, the P1 one, uh, base URL. One for the loading. And, uh, yeah, potentially the fifth one. So, yeah, currently there is four, uh, four fixes in… in my… On my… work.
|
||||
- **18:23:53 · Michal Smaga:** So, do you have any, like, doubts, like, that, like, some of the fixes are, like, too risky, and, like, for example, the reload one? you would be not too sure about, like, shipping that one? Because I'm just, like, thinking…
|
||||
- **18:24:05 · Alex Martemyanov:** Uh, Zip?
|
||||
- **18:24:07 · Michal Smaga:** thinking about, like, probably we can have the questions from Russell or someone, like. How much more time we can spend here, like, just to, like, focus on the, like, bare minimum, because probably, like, just to, like… better calculate the effort that we want to put into it.
|
||||
- **18:24:27 · Alex Martemyanov:** Yeah, generally, we already had this question by Russell, as I mentioned, and I tried to answer it today.
|
||||
- **18:24:30 · Michal Smaga:** Yeah, yeah, yeah.
|
||||
- **18:24:37 · Alex Martemyanov:** No, we've seen it. Uh, so the issue here is, uh, that we cannot distribute. the failures. You cannot tell there is one main… ah, yeah, there is another. Another one… And… Great deeds. Yeah, another genuine one that is kind of proved is window git close. So the original. Window and tab merging management, uh, fix it, uh… It didn't cover every… Every issue with the pin tab, and there is still one more remaining, uh, for the window closing, and, uh, yeah, should be a short one. Uh, it should cover the genuine issue of closing in tab using the web extension.
|
||||
- **18:25:31 · Michal Smaga:** But it's only McCoy, right?
|
||||
- **18:25:32 · Alex Martemyanov:** I don't know. Yeah, my question. So, yeah, that is, uh, for… Four, one, two, three, one, two, three, four main groups that are targeted here, one requiring the page reload. And the P3 that, uh, without… Context without view, it may be the pin tab. And we have one iOS of 0.03%. macOS at around 6%, but it may be wrong. And we have the other group that cannot be considered at the moment. Maybe it gets fixed with the four fixes, or maybe we receive some more data. So I try to express this point in the conversation task. But if the hypothesis are correct, we should cover most of the cases.
|
||||
- **18:26:37 · Michal Smaga:** So, do you estimate to get them done this week?
|
||||
- **18:26:37 · Alex Martemyanov:** So, yeah. Yeah, so the point here is to timebox it to…
|
||||
- **18:26:45 · Michal Smaga:** To merge it this week, and so we can, like, get it internally next week, right?
|
||||
- **18:26:45 · Alex Martemyanov:** Two currents. Yeah. Yeah, so if we… if we get it to the release… I mean, the data analysis, the data attribution extension, and those fixes, if we get it to internal. Today and tomorrow. Uh, we can reason about it on… Next week.
|
||||
- **18:27:11 · Michal Smaga:** Okay, so you would mean to merge it to… Release branches.
|
||||
- **18:27:17 · Alex Martemyanov:** Yes, if it's possible.
|
||||
- **18:27:22 · Michal Smaga:** So, the two first PRs, like, one for the telemetry for the pixels, and the second with the P1 and P2, right?
|
||||
- **18:27:31 · Alex Martemyanov:** Um… It depends, I guess, if we can do it, if it's possible, I would give a go for all of them. Again, behind the feature flags.
|
||||
- **18:27:43 · Michal Smaga:** Yeah, let's try. So, like, I'm… I'm… I'm… I'm done, like… reading and commenting for the first PRs, there are a couple of, like. leftovers for the dead code, so please check the comments. I'll try to, like, try forcing to reproduce it.
|
||||
- **18:27:55 · Alex Martemyanov:** Mmhm.
|
||||
- **18:28:01 · Michal Smaga:** So so I get it and then I can move to the next PR. I can focus today on that. So we have the the the chart.
|
||||
- **18:28:08 · Alex Martemyanov:** Okay, good. And when we have those reviewed, we can open the exception and try to push it to the internal.
|
||||
- **18:28:18 · Michal Smaga:** I mean, like, if it's, like, Monday, Tuesday, you can do it probably, like, for free.
|
||||
- **18:28:24 · Alex Martemyanov:** Mm-hmm, okay.
|
||||
- **18:28:24 · Michal Smaga:** If it's later than tomorrow. Yeah. Okay. Okay, so let's try to do that.
|
||||
- **18:28:30 · Alex Martemyanov:** Okay, I'll… I'll keep you pinged, yeah.
|
||||
- **18:28:31 · Michal Smaga:** And… Please ping me when you have the following one, because I want to, like…
|
||||
- **18:28:37 · Alex Martemyanov:** Mm-hmm.
|
||||
- **18:28:38 · Michal Smaga:** Schedule the time for them, uh, like, in between my, my work, just to, just to not, like… Delay you. Uh, yeah.
|
||||
- **18:28:47 · Alex Martemyanov:** Yeah, I guess it's… I guess it may take some time, maybe the rest of your day today, and those two PRs. Could feel it.
|
||||
- **18:28:58 · Michal Smaga:** Yeah, so I'll focus on that, not to block you, so let's push for the first two ones, and yeah.
|
||||
- **18:29:04 · Alex Martemyanov:** Mmhm.
|
||||
- **18:29:04 · Michal Smaga:** And the others, like, later in the week.
|
||||
- **18:29:07 · Alex Martemyanov:** Mm-hmm.
|
||||
- **18:29:08 · Michal Smaga:** Yeah, I agree.
|
||||
- **18:29:09 · Alex Martemyanov:** So, and yeah.
|
||||
- **18:29:10 · Michal Smaga:** Just to try to get them out as soon as possible to get some results.
|
||||
- **18:29:16 · Alex Martemyanov:** And yes, since it's highly speculative, I would give it another mid-mortem. So yeah, time box it to this week. I'd say to the first half of this week. Because there is not much work really to be done about it. Just, uh… Just so you know, I have, like, a huge testing harness work tree, and I'm trying to split the parts from it that reproduce one another issue and make it into…
|
||||
- **18:29:48 · Michal Smaga:** No doubt.
|
||||
- **18:29:53 · Alex Martemyanov:** into digestible pull requests. But generally, most of them are kind of reproduced in the lab. But we have no any proofs that they actually match the production issues because they are based on the event sequence that we get.
|
||||
- **18:30:08 · Michal Smaga:** Yeah.
|
||||
- **18:30:14 · Alex Martemyanov:** We got to produce the event sequence that we are receiving, but we still cannot reason about the real issues leading to those. So, yeah, that's the state. All right, thank you. I won't take more of your time.
|
||||
- **18:30:31 · Michal Smaga:** So, after this week, you would put the, like, you would have the mid-mortem this week, or, like, next week after the…
|
||||
- **18:30:41 · Alex Martemyanov:** Next week, so let's try to get it out and have some data. The meeting is on Tuesday, right? So we could have the data analyzed and, uh, have a mid-mortem. and decide on the next steps whether we need. Keep on going, or we are satisfied with those, or we just give up and…
|
||||
- **18:31:06 · Michal Smaga:** Yeah. But do you think, like, on Tuesday, you will have, like, any meaningful data? It will be, like, just out for two days. Right?
|
||||
- **18:31:13 · Alex Martemyanov:** Uh, yes, we should, uh, since we… we received most of the current data by Tuesday. And it didn't change significantly throughout the week. I think there is a lot of issues are reported.
|
||||
- **18:31:36 · Michal Smaga:** Okay.
|
||||
- **18:31:37 · Alex Martemyanov:** And there's more than enough data to analyze. So, yeah, Tuesday should be should be fine to have a big more time. I'll probably update the comments. Highlighting this.
|
||||
- **18:31:51 · Michal Smaga:** Okay, sounds good. Okay, so I'm back to reviewing it, and, like, let's… Probably tomorrow you can, like, will you be on the OE weekly?
|
||||
- **18:32:05 · Alex Martemyanov:** Yes, tomorrow, tomorrow, tomorrow, tomorrow is Tuesday, yes, I think so.
|
||||
- **18:32:11 · Michal Smaga:** Yeah, so we can recap the plan and see. So this is, like, your top priority, right? You're focusing on…
|
||||
- **18:32:20 · Alex Martemyanov:** Yeah, yeah. So, I mean, there is, uh, there is enough work for me outside of the top priority. Like where you are and pushing other stuff forward. But, yeah.
|
||||
- **18:32:39 · Michal Smaga:** Makes sense. Cool. Are you blocked by anything, or just, like, this is, like… The general, like, work is enough for you.
|
||||
- **18:32:49 · Alex Martemyanov:** Alright. Blocked with the data missing. And, uh, so… Only the pull requests are on the plate. Bottle making it. And, uh, yeah, the gaps between the releases, so we… keep perceiving it and. Analyzing how it goes.
|
||||
- **18:33:11 · Michal Smaga:** Have you assigned me to all of the PRs? Because, uh, I see them…
|
||||
- **18:33:15 · Alex Martemyanov:** Not yet, I guess. Let me assign the network process. I didn't go through it myself yet, but yeah, I've assigned it to you.
|
||||
- **18:33:29 · Michal Smaga:** Yeah, I think, like, something is broken with the…
|
||||
- **18:33:33 · Alex Martemyanov:** Yeah, yeah, I'll… I'm on it.
|
||||
- **18:33:36 · Michal Smaga:** No, no, like, I mean, like, uh, I'm browsing through the, the, this PR, these are the PRs, and, like, I'm filtering by the assigned to me. And I only picked it up.
|
||||
- **18:33:47 · Alex Martemyanov:** Ah, assignees. I need to assign them. We generally don't assign the PRs, just set a reviewer.
|
||||
- **18:33:54 · Michal Smaga:** I don't know, like, you have the filter, like, review requested.
|
||||
- **18:34:00 · Alex Martemyanov:** Mm-hmm.
|
||||
- **18:34:01 · Michal Smaga:** They can…
|
||||
- **18:34:01 · Alex Martemyanov:** Yeah, let me just send. Yes, I need. and requests. Yeah, so there is three pull requests.
|
||||
- **18:34:07 · Michal Smaga:** Can you… can you see? So those are, like, open PRs from you all, but you have the filter for review, and it's, like, awaiting review from…
|
||||
- **18:34:14 · Alex Martemyanov:** Yep.
|
||||
- **18:34:18 · Michal Smaga:** Me? I only get two of them.
|
||||
- **18:34:24 · Alex Martemyanov:** Okay. 7, 4, 5, 5, 7.
|
||||
- **18:34:26 · Michal Smaga:** Which is weird.
|
||||
- **18:34:29 · Alex Martemyanov:** Oh, 7376 is missing for some reason. Probably you gave a review for it already. And it's waiting for…
|
||||
- **18:34:38 · Michal Smaga:** Okay.
|
||||
- **18:34:40 · Alex Martemyanov:** For the resolutions. Angus.
|
||||
- **18:34:43 · Michal Smaga:** Okay, so maybe it's like, yeah.
|
||||
- **18:34:46 · Alex Martemyanov:** Yeah, so currently two more in stock.
|
||||
- **18:34:50 · Michal Smaga:** Okay. So it's it's in the stack, so I can get them from from the list. Right?
|
||||
- **18:34:55 · Alex Martemyanov:** Yep.
|
||||
- **18:34:56 · Michal Smaga:** Okay, cool. Can I help you in any other way?
|
||||
- **18:34:59 · Alex Martemyanov:** Review stuck. Yes, stuck, stuck, stuck. Create stuck. Create stuck. And now those are in the stacks. Okay, good.
|
||||
- **18:35:15 · Michal Smaga:** Do you need any other help?
|
||||
- **18:35:18 · Alex Martemyanov:** Not at the moment. I'm going to update the first one, telemetry one. And give you a ping, and uh… And when I complete the third one, I'll ping you as well, and probably… Probably the window did close, should be an easy one for today. As well. If it sounds good.
|
||||
- **18:35:46 · Michal Smaga:** Okay. Okay, so I'm on the PRs. Please check the comments for the first one. I'm all moving for the second one.
|
||||
- **18:35:58 · Alex Martemyanov:** Sure. Okay, thank you.
|
||||
- **18:36:03 · Michal Smaga:** Thank you.
|
||||
-230
@@ -23,236 +23,6 @@
|
||||
"time": "20:03:30",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "The meeting is at the wrong time in Europe this week, so some people may be missing because of that."
|
||||
},
|
||||
{
|
||||
"time": "20:03:38",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "I totally lost track of which one's which, that's fine, I was just pulling up rotations so we can get started anyway. Hello, everyone. Nice to see you all. Quick run through our rotations as usual. Bartek is on iOS drives this week, followed by Brindy and then Hassan. Echoist triages Jorge this week, followed by Rachel, and then Alessandro. For release duties, iOS this week, Bartosz, followed by Mihao, and then Pete. And macOS is Tom this week, followed by Juan, and then Joseph. Any issues, go ahead and flag those in Asana as usual, please. Sam, you have an announcement to share with us. I think you're, yes."
|
||||
},
|
||||
{
|
||||
"time": "20:04:27",
|
||||
"speaker": "SS",
|
||||
"text": "Right, this is the AOR change. Sorry, I don't have the agenda pulled up. Uh, yes, uh…"
|
||||
},
|
||||
{
|
||||
"time": "20:04:29",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Thank you. Yes, sorry."
|
||||
},
|
||||
{
|
||||
"time": "20:04:34",
|
||||
"speaker": "SS",
|
||||
"text": "Is George here? Yes, he is. George has taken over secondary of the VPN AOR from me after running some projects to get us better observability into the VPN, so… Thank you. Nice work. Yes, also that."
|
||||
},
|
||||
{
|
||||
"time": "20:04:48",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Excellent. Thank you and congratulations."
|
||||
},
|
||||
{
|
||||
"time": "20:04:50",
|
||||
"speaker": "SS",
|
||||
"text": "So also also. Also, what Jacek said in the chat is accurate."
|
||||
},
|
||||
{
|
||||
"time": "20:04:58",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Yeah, I mean, I guess this is as good a time as any way to remind people that AOR, DRI ships are not written in stone. They are changeable. If you want one and don't have one, want a particular one, want to get out of one. We can talk about it. You don't even have to talk to me. You can make a proposal. They are changeable. Thank you. Dominic, CI usage update from you."
|
||||
},
|
||||
{
|
||||
"time": "20:05:25",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "Uh, yes. Um… It's pretty early in the agenda here. But okay, anyway. I have this… Artifact. Here, that I can share with you. Or not. Doesn't work. One second, I'm sorry. Um… Oh, yes. So anyway, I was looking at… I was looking at, um… And CI time. In the recent months and just to… just to have a look, just to get a feeling of how… Bitrise works or doesn't work. Here's the link, but let me also share it with you. Show my screen, I guess. I'm, uh, I mean, uh, so, uh… I have this… job hours and runs per month since February. Uh, it was going up and down, and at some point, uh, job runs… Pretty steady at… 5,500 a month, whereas job hours are going down. Quite drastically, actually, because it's at 1700 for September. And it has never been below 2000 earlier this year. Not sure about January, but anyway. Yeah, which is interesting. And also… Average… Job duration. It goes… it went down to 18.9 minutes from over 30. This is only for the PR checks, so VSK, DBP, iOS, and macOS PR checks only. in the monorepo, uh, which… Looks like… Uh, confirming that Bitrise is doing the job. Uh, but my initial… uh, motivation to… to run this was to see if we should consider getting more machines from Bitrise, because maybe we're running more jobs. Which was my assumption. Because… Code is cheap these days, relatively cheap. cheaper than it used to be, and uh… because of that, we might be opening more pull requests, running more workflows, and so on and so on, so maybe we need more machines, because we would have paid more for GitHub anyway. not sure if it's the case yet. October is not here, even though it's 27th, and I did yesterday, but in October, we enabled… Required checks. in the repository, which means all the checks run on all pull requests now, so naturally there's higher volume of checks, and it wasn't fitting this graph nicely. So I guess we'll have a new baseline in October, and then. We'll be comparing against that in the following months. That's what I have. Tommy has any thoughts?"
|
||||
},
|
||||
{
|
||||
"time": "20:08:50",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "When did we actually fully switch to Bitrise?"
|
||||
},
|
||||
{
|
||||
"time": "20:08:53",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "Um… September 14th?"
|
||||
},
|
||||
{
|
||||
"time": "20:08:59",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "So that would, because they're faster, that would drive down our run times while jobs remain steady potentially, which it looked like that actually showed in the data you were just showing, right?"
|
||||
},
|
||||
{
|
||||
"time": "20:08:59",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "Yes. So, one correction. September 14th was when we got the final deal in place, but… around 10th of August, I think it was 11th of August when we started the POC, and we were already on Bitrise, uh, with a significant, uh, amount of our jobs."
|
||||
},
|
||||
{
|
||||
"time": "20:09:25",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Right. Right, so things will go faster, which means we would see less time per job overall, right?"
|
||||
},
|
||||
{
|
||||
"time": "20:09:27",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "Well, yeah. Yes, yes."
|
||||
},
|
||||
{
|
||||
"time": "20:09:33",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Perfect."
|
||||
},
|
||||
{
|
||||
"time": "20:09:34",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "And let's see if we see more jobs overall. I mean, like I said, in October, there are already more jobs, but let's see about November, December, and so on."
|
||||
},
|
||||
{
|
||||
"time": "20:09:39",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Yep. Yeah."
|
||||
},
|
||||
{
|
||||
"time": "20:09:45",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "Jeff, from Partnerships, has been on a 3-week vacation, but he's coming back tomorrow, or Thursday, and we might catch up with Bitrise. I'm… I'm on calls with Beatrice anyway, once a month, so I might reach out to them. Amen. Time we want, uh, to… to discuss that. Yeah, I think we should probably wait, still wait a bit. But yeah, it should be possible. Well, it is possible. He told me it's possible, of course."
|
||||
},
|
||||
{
|
||||
"time": "20:10:22",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Cool. Well, thank you. That's an interesting update."
|
||||
},
|
||||
{
|
||||
"time": "20:10:22",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "That's it. Thank you very much."
|
||||
},
|
||||
{
|
||||
"time": "20:10:26",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Diego, two items from you."
|
||||
},
|
||||
{
|
||||
"time": "20:10:31",
|
||||
"speaker": "DR",
|
||||
"text": "Really? I'm not prepared. Um, is that the demos from last week?"
|
||||
},
|
||||
{
|
||||
"time": "20:10:36",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "I think so, yeah, the passkeys issue and the Bitwarden extension."
|
||||
},
|
||||
{
|
||||
"time": "20:10:40",
|
||||
"speaker": "DR",
|
||||
"text": "Okay, I don't have a passkeys issue ready demo, but I'm going to explain what happened briefly and then just so everybody knows what's going on with that. So we had an issue with passkeys in macOS. Um… The browser stops asking for passkey authorization. You don't see the prompt anymore after a while. Uh, it wasn't clear what was going on. Um, apparently what happens is that, uh, Safari… doesn't suffer from this because they're allowed to make as many requests to the authentication service as they want for Passkey. So if they make. Two requests, the… The service is going to honor both requests and show the UI for both. But in the case of our browser, we can only make one request at a time. And WebKit isn't, um, handling that properly. They're not cleaning up, uh… They're not cleaning up the requests. When you do certain actions with the windows, like, for example, if you click on the button. In one window, and then open another window, and have another request, the second one is not gonna show. And at some point, these interactions were breaking passkeys completely to the point where they didn't work anymore for the user. So what we did? We implemented a workaround where we're, again, hooking into private APIs for WebKit to make. The APIs release the passkey requests correctly. And I have pending tasks that I haven't been able to get to yet to open, fix. a PR to fix the issue in WebKit initially, because there's an issue in WebKit where it's not canceling the request, right? And another issue where we're gonna be updating our communications with Apple, because previously, I think Dominic was talking with them about this at some point, and we didn't have a good, uh… minimum example of how… what the issue was to show them. So we're gonna provide something to them as well. Uh, this is also what triggered the idea, I don't know if you saw that apple that I reposed, to… uh, start encouraging more our interactions with the WebKit, uh, team, and proposing changes, and… You know, even submitting PRs, if that makes sense. Um… So yeah, that's it. And the second you saw on Friday, if you were there, which is that we have now Bitwarden extensions and 1Password extensions in the browser. which you can use. I think Dominic posted about it today, I think, or yesterday, I can't remember."
|
||||
},
|
||||
{
|
||||
"time": "20:13:35",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "Yesterday."
|
||||
},
|
||||
{
|
||||
"time": "20:13:35",
|
||||
"speaker": "DR",
|
||||
"text": "Um, yeah. And if you want to give it a go, it would be great. You need to be an internal user. Let me see if I have it installed here. Doc. Mm-hmm. It's here. All right. No, I don't have it installed. Sorry. I wasn't ready for the demos today. But yeah, that's it for me. Thanks, everybody."
|
||||
},
|
||||
{
|
||||
"time": "20:14:08",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "I wanted to ask. Because we, uh, run this entire thing about passkeys on macOS, but… We didn't check iOS at all, right? But I think… Does it work the same on iOS? I don't recall anymore. We probably also had that entitlement requested for iOS or not."
|
||||
},
|
||||
{
|
||||
"time": "20:14:30",
|
||||
"speaker": "DR",
|
||||
"text": "I don't know, honestly. That's a good question. I'm going to take note now."
|
||||
},
|
||||
{
|
||||
"time": "20:14:33",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "Okay."
|
||||
},
|
||||
{
|
||||
"time": "20:14:35",
|
||||
"speaker": "DR",
|
||||
"text": "To look into that. But it's possible we're going to be seeing the same issue. I don't know, because the entitlement that we would need for that not to be an issue is… I don't think Apple is giving that away, and I don't think we have that on iOS, so it may be an issue there as well. Uh… But yeah, I mean, the request to make the WebKit submission to fix this. Would fix it for all. browsers. But it would be interesting to know if it's breaking in iOS. I agree."
|
||||
},
|
||||
{
|
||||
"time": "20:15:09",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "Yeah, of course. Now that I think of it, I… I'm not looking at iOS feedback that much, but probably we don't have? feedback about passkeys not working in the iOS app?"
|
||||
},
|
||||
{
|
||||
"time": "20:15:24",
|
||||
"speaker": "DR",
|
||||
"text": "I did look at some point and I didn't find much from iOS, but then I didn't follow up to check if it was causing trouble."
|
||||
},
|
||||
{
|
||||
"time": "20:15:26",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "There's only one recorder."
|
||||
},
|
||||
{
|
||||
"time": "20:15:33",
|
||||
"speaker": "DR",
|
||||
"text": "Which is the interesting bit."
|
||||
},
|
||||
{
|
||||
"time": "20:15:35",
|
||||
"speaker": "Dominik Kapusta",
|
||||
"text": "Mm-hmm. Cool, thanks. One last thing about extensions is that, and this hasn't been announced anywhere, but we also support Dark Reader. If any one of you fancies Dark Reader, you can install it from Chrome Web Store. Just go to Chrome Web Store, search for Dark Reader and install it and it works. It works much better than whatever we had for Dark Reader, uh, integrated behind the scenes, and that thing is going away at some point. But yeah. We also support iCloud Passwords, which is a terrible extension, but it's the only way to connect a browser that's not Safari. with passwords up. Thank you."
|
||||
},
|
||||
{
|
||||
"time": "20:16:23",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Awesome, thank you. And thank you, Diego, for rolling with that. If you have tasks or anything that are related to those that you could pop into the agenda at some point, that would help people find some more of this history, too. That'd be great. Uh, if not, that's fine. Very handy, that'd be…"
|
||||
},
|
||||
{
|
||||
"time": "20:16:36",
|
||||
"speaker": "DR",
|
||||
"text": "Yeah, I'm gonna… no, I'm gonna look them up now, yeah, and then I'll…"
|
||||
},
|
||||
{
|
||||
"time": "20:16:40",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Cool. Thank you. Feel free to drop them in the agenda. That'd be fine. And then the last thing on the agenda, Alex, is a postmortem from you."
|
||||
},
|
||||
{
|
||||
"time": "20:16:50",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yeah, hey, it's the postmortem for VS Code support. So actually the migration to support VS Code and Carstore initially. I'll… So, the… Originally, the project started more than a year ago as a Huggies project. It grew out of… general AUR optimization and refactoring work related to storyboards, removal, modularization, and other. And, uh… Uh, my half-day's effort was to try to integrate it to Cursor in that time, and, uh, it's shown that the approach is actually working. Uh, we could… potentially in, uh, use CorpStore and, uh, integrate with it, having, uh, syntax, uh, highlighting, uh, documentation, and, uh, jump to symbol, builds, and so on, and debugging. Ah, so it's, uh… the project was, uh… formed from the steps remaining to make it work. But first of all, it included all the. AUR work that was already started, like storyboards removal, uh, some of, uh, flags organization and other stuff, and, uh… It had no kickoff as it was kind of running in my background between top priorities, and that's why it took so long to finish. Uh… and uh… Yeah, it had. It had some some issues visibility issues, I'd say so. First of all, did it achieve its success criteria and its purpose? So I think strong. Yes, so. in most of the work. or preparation was finished. Some of it is… It's moved out of the project back to the AUR, but the amount of work needed to make it work was done, and it would have been done in scope of the AUR as the tech debt pending probably someday. Uh, but yeah, this project gave it kind of, um, shape and served as an umbrella for those, uh. I think it's… Vs code. Cursor was completed, at least on my end. It's available in my public repo. If anybody is interested and requires a small github githook to integrate into project. and uses mostly native means, and fully compatible with Xcode, and does not replace or harm Xcode experience. So what went well and I could highlight? Yeah, each of the project steps were shipped as a small and separate refactors. Each one was an improvement on its own. There were several tech designs. First of all, the first tech design of. Modernization and Cursor support was originally discussed and approved when we mostly adopted Cursor, and it seemed like a good target. Uh, and it had a follow-up this year, recently, that was, uh, that was canceled. And… Basically, it was the same tech design, but it was really discussed after… after some… changes and when we build closer to completion. And with EXPO27 release, the integration was much simpler than the original plan, and did not require most controversial parts of the project. So this was discussed with the team, and. Yeah, the decision was made to not to move forward with it, but have a small… Small fixes or small adaptation that allow it to complete. Yeah, so the setup is usable. I use it instead of Xcode, except for release deals or Ui tests. And, uh… Yeah, basically it's moving, it's moving. seemingly moving forward to make… and Apple is moving forward to make, uh, it's more open and more viable with other dev tools, if anybody wants to experiment with it. Uh, and the work, I think, I could also show, highlight that the work is not completed here. Uh, we still have a lot of tech debt to solve, a lot of, uh, modelization efforts to be done there."
|
||||
},
|
||||
{
|
||||
"time": "20:21:58",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "And it continues, basically. What didn't go well? Uh, there was, uh, an incident on macOS 12, uh, because it's generally not tested in our, uh, in our setup, as we already discussed in the incident post-mortem last time. Uh, while we still support, officially support, the SmuckOS version, we are not testing it effectively. And, uh, this, uh, gave me some thoughts of how could it be improved, probably. So, first of all, I… experiment… I experiment locally with, uh, testing… testing harness and agent harness, supporting macOS 12 as well. And, uh… later on, I'll try to… Push it forward to… Okay. Uh, to be able to actually test it. So, first of all, I've made an effort to, uh… to split. with actual UI test builds from macOS runners to one macOS builder, unified between different versions, so runners won't have. I have to build it every time, and this is improvement on its own. It improves build times, basically. And another is testing hardness to support Mac OS 12. So it's not… it's not open as a pull request to review yet, but it's working. Uh, so it's proof that it can work, and we can… could move forward with it, and probably rent some, uh, older machines somewhere in the cloud, and uh… have it tested. The second suggestion here would be to, uh, improve our processes, so I didn't, uh, complete the proposal yet, with all this, uh, performance review stuff, the priority. Uh, but yeah, I was planning to, uh… at a proposal. to require writing UI tests for any modified or added user-facing features and extend coverage overall as it does not require so much effort. We only need some Agentic work to analyze. coverage to suggest some tests and basically push it one by one. Um… Another issue was a tab bar disappearing. It was spotted during the internal testing. Basically, general work. And again, UI testing would support it. Yeah. Next thing is, uh, related to the visibility. Uh, so basically, demand for the VS Code integration and commercial integration was checked too late, as the project was, uh, done as in my background. So in its purpose, its aim was mostly my personal and included most tension, while most of the project was not related to actual VS Code migration, but served as an. Umbrella project to… Granted. Uh, this was, uh… This was only touched on the last stage, and uh… Probably we… While the project was progressing in the days before, we could have discussed it during some of Apple Weekly. But maybe not since. the actual implementation and the actual step that required this discussion was raised just in time, when Xcode 27 was available, and this step was about to be made. So. yeah, there were some issues with visibility, but it's mostly related to my… to me doing this project as my… not as my top priority, but it's generally your work when I have time for it. Okay. Yeah, and another outcome from it is the project goal was from the round one IDE support, as I already told, basically to migrate to VS Code and eat. might separate some tension and questions. Whether do we need to migrate to Vs code while again, 85% of the project was not related to actual Vs code migration and was general refactoring and improvements. It's, yeah, mostly related to the issue number two. of… Basically, basically, that's it, everything we discussed, and uh… I'll…"
|
||||
},
|
||||
{
|
||||
"time": "20:26:58",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "push forward with the Ui testing suggest proposal, and. older version testing, which is a requirement, especially with the more and more usage of Agentic workflows, where we have, like, 95 to 98% of. perfect, there are 8-time code, but there is still this 2% that is not perfect, and it tends to fail in the most critical scenarios. And it becomes a requirement, so… it's not the implementation, uh, that needs to be tested more, but the test… actual testing. And, uh, we need to… Obviously, we need to raise our quality assurance. Versuses. Let's see. Yeah, and the remaining factors will be handles. I have them tracked in the AOR. Not much of priority again, but they are on my plate. That's it. Any comments, any questions?"
|
||||
},
|
||||
{
|
||||
"time": "20:28:14",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Cool. Thank you, Alex. For the other refactors and stuff that are going to happen as part of the AOR. I think we just need to make sure that the value of those are clear in their own steps, right? Rather than, like you said, rather than focusing them on a VS Code support item, we'll just talk about each of those and what they bring to the team in their own right."
|
||||
},
|
||||
{
|
||||
"time": "20:28:35",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "Yes, so they were originally part included in this project, but we should separate cleanup targets. So each of the dependencies is not used by multiple targets, but has its own framework, setting up testing harness."
|
||||
},
|
||||
{
|
||||
"time": "20:28:40",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Yep. Yeah. Yeah, and a lot of that is in line with the modularization desires that we've had anyway, so."
|
||||
},
|
||||
{
|
||||
"time": "20:28:53",
|
||||
"speaker": "Alex Martemyanov",
|
||||
"text": "supporting UI person. Yep, yep."
|
||||
},
|
||||
{
|
||||
"time": "20:28:58",
|
||||
"speaker": "Jonathan Jackson",
|
||||
"text": "Excellent. Cool. Thank you. Appreciate it. Anything else from anyone before we wrap our meeting for the day? Remember that today is the first of three days for career advisors to be writing, uh, assessments for their advisees, so… We will all probably be heads down on that. I guess if you have questions, career advisors, if you have any questions that I can help with, go ahead and let me know. Feel free to reach out. Thanks, everyone. Good to see you all."
|
||||
}
|
||||
]
|
||||
}
|
||||
-46
@@ -15,49 +15,3 @@ source: "https://us01mynotes.zoom.us/-I3aVYDWTkKZ_VTRLdjGnA"
|
||||
- **20:02:36 · Jorge Leandro Perez:** So, Tokyo 2027? Yes!
|
||||
- **20:02:45 · Pikor:** And then we need that cool t-shirt, because someone created a Hug Days project with designs of t-shirts from the past, and the backend meetup had really one nice one with the. Japanese, um… I guess it said Dactago didn't check it, but if we would have that nice t-shirt of… Our meetup, that would be cool.
|
||||
- **20:03:30 · Dominik Kapusta:** The meeting is at the wrong time in Europe this week, so some people may be missing because of that.
|
||||
- **20:03:38 · Jonathan Jackson:** I totally lost track of which one's which, that's fine, I was just pulling up rotations so we can get started anyway. Hello, everyone. Nice to see you all. Quick run through our rotations as usual. Bartek is on iOS drives this week, followed by Brindy and then Hassan. Echoist triages Jorge this week, followed by Rachel, and then Alessandro. For release duties, iOS this week, Bartosz, followed by Mihao, and then Pete. And macOS is Tom this week, followed by Juan, and then Joseph. Any issues, go ahead and flag those in Asana as usual, please. Sam, you have an announcement to share with us. I think you're, yes.
|
||||
- **20:04:27 · SS:** Right, this is the AOR change. Sorry, I don't have the agenda pulled up. Uh, yes, uh…
|
||||
- **20:04:29 · Jonathan Jackson:** Thank you. Yes, sorry.
|
||||
- **20:04:34 · SS:** Is George here? Yes, he is. George has taken over secondary of the VPN AOR from me after running some projects to get us better observability into the VPN, so… Thank you. Nice work. Yes, also that.
|
||||
- **20:04:48 · Jonathan Jackson:** Excellent. Thank you and congratulations.
|
||||
- **20:04:50 · SS:** So also also. Also, what Jacek said in the chat is accurate.
|
||||
- **20:04:58 · Jonathan Jackson:** Yeah, I mean, I guess this is as good a time as any way to remind people that AOR, DRI ships are not written in stone. They are changeable. If you want one and don't have one, want a particular one, want to get out of one. We can talk about it. You don't even have to talk to me. You can make a proposal. They are changeable. Thank you. Dominic, CI usage update from you.
|
||||
- **20:05:25 · Dominik Kapusta:** Uh, yes. Um… It's pretty early in the agenda here. But okay, anyway. I have this… Artifact. Here, that I can share with you. Or not. Doesn't work. One second, I'm sorry. Um… Oh, yes. So anyway, I was looking at… I was looking at, um… And CI time. In the recent months and just to… just to have a look, just to get a feeling of how… Bitrise works or doesn't work. Here's the link, but let me also share it with you. Show my screen, I guess. I'm, uh, I mean, uh, so, uh… I have this… job hours and runs per month since February. Uh, it was going up and down, and at some point, uh, job runs… Pretty steady at… 5,500 a month, whereas job hours are going down. Quite drastically, actually, because it's at 1700 for September. And it has never been below 2000 earlier this year. Not sure about January, but anyway. Yeah, which is interesting. And also… Average… Job duration. It goes… it went down to 18.9 minutes from over 30. This is only for the PR checks, so VSK, DBP, iOS, and macOS PR checks only. in the monorepo, uh, which… Looks like… Uh, confirming that Bitrise is doing the job. Uh, but my initial… uh, motivation to… to run this was to see if we should consider getting more machines from Bitrise, because maybe we're running more jobs. Which was my assumption. Because… Code is cheap these days, relatively cheap. cheaper than it used to be, and uh… because of that, we might be opening more pull requests, running more workflows, and so on and so on, so maybe we need more machines, because we would have paid more for GitHub anyway. not sure if it's the case yet. October is not here, even though it's 27th, and I did yesterday, but in October, we enabled… Required checks. in the repository, which means all the checks run on all pull requests now, so naturally there's higher volume of checks, and it wasn't fitting this graph nicely. So I guess we'll have a new baseline in October, and then. We'll be comparing against that in the following months. That's what I have. Tommy has any thoughts?
|
||||
- **20:08:50 · Jonathan Jackson:** When did we actually fully switch to Bitrise?
|
||||
- **20:08:53 · Dominik Kapusta:** Um… September 14th?
|
||||
- **20:08:59 · Jonathan Jackson:** So that would, because they're faster, that would drive down our run times while jobs remain steady potentially, which it looked like that actually showed in the data you were just showing, right?
|
||||
- **20:08:59 · Dominik Kapusta:** Yes. So, one correction. September 14th was when we got the final deal in place, but… around 10th of August, I think it was 11th of August when we started the POC, and we were already on Bitrise, uh, with a significant, uh, amount of our jobs.
|
||||
- **20:09:25 · Jonathan Jackson:** Right. Right, so things will go faster, which means we would see less time per job overall, right?
|
||||
- **20:09:27 · Dominik Kapusta:** Well, yeah. Yes, yes.
|
||||
- **20:09:33 · Jonathan Jackson:** Perfect.
|
||||
- **20:09:34 · Dominik Kapusta:** And let's see if we see more jobs overall. I mean, like I said, in October, there are already more jobs, but let's see about November, December, and so on.
|
||||
- **20:09:39 · Jonathan Jackson:** Yep. Yeah.
|
||||
- **20:09:45 · Dominik Kapusta:** Jeff, from Partnerships, has been on a 3-week vacation, but he's coming back tomorrow, or Thursday, and we might catch up with Bitrise. I'm… I'm on calls with Beatrice anyway, once a month, so I might reach out to them. Amen. Time we want, uh, to… to discuss that. Yeah, I think we should probably wait, still wait a bit. But yeah, it should be possible. Well, it is possible. He told me it's possible, of course.
|
||||
- **20:10:22 · Jonathan Jackson:** Cool. Well, thank you. That's an interesting update.
|
||||
- **20:10:22 · Dominik Kapusta:** That's it. Thank you very much.
|
||||
- **20:10:26 · Jonathan Jackson:** Diego, two items from you.
|
||||
- **20:10:31 · DR:** Really? I'm not prepared. Um, is that the demos from last week?
|
||||
- **20:10:36 · Jonathan Jackson:** I think so, yeah, the passkeys issue and the Bitwarden extension.
|
||||
- **20:10:40 · DR:** Okay, I don't have a passkeys issue ready demo, but I'm going to explain what happened briefly and then just so everybody knows what's going on with that. So we had an issue with passkeys in macOS. Um… The browser stops asking for passkey authorization. You don't see the prompt anymore after a while. Uh, it wasn't clear what was going on. Um, apparently what happens is that, uh, Safari… doesn't suffer from this because they're allowed to make as many requests to the authentication service as they want for Passkey. So if they make. Two requests, the… The service is going to honor both requests and show the UI for both. But in the case of our browser, we can only make one request at a time. And WebKit isn't, um, handling that properly. They're not cleaning up, uh… They're not cleaning up the requests. When you do certain actions with the windows, like, for example, if you click on the button. In one window, and then open another window, and have another request, the second one is not gonna show. And at some point, these interactions were breaking passkeys completely to the point where they didn't work anymore for the user. So what we did? We implemented a workaround where we're, again, hooking into private APIs for WebKit to make. The APIs release the passkey requests correctly. And I have pending tasks that I haven't been able to get to yet to open, fix. a PR to fix the issue in WebKit initially, because there's an issue in WebKit where it's not canceling the request, right? And another issue where we're gonna be updating our communications with Apple, because previously, I think Dominic was talking with them about this at some point, and we didn't have a good, uh… minimum example of how… what the issue was to show them. So we're gonna provide something to them as well. Uh, this is also what triggered the idea, I don't know if you saw that apple that I reposed, to… uh, start encouraging more our interactions with the WebKit, uh, team, and proposing changes, and… You know, even submitting PRs, if that makes sense. Um… So yeah, that's it. And the second you saw on Friday, if you were there, which is that we have now Bitwarden extensions and 1Password extensions in the browser. which you can use. I think Dominic posted about it today, I think, or yesterday, I can't remember.
|
||||
- **20:13:35 · Dominik Kapusta:** Yesterday.
|
||||
- **20:13:35 · DR:** Um, yeah. And if you want to give it a go, it would be great. You need to be an internal user. Let me see if I have it installed here. Doc. Mm-hmm. It's here. All right. No, I don't have it installed. Sorry. I wasn't ready for the demos today. But yeah, that's it for me. Thanks, everybody.
|
||||
- **20:14:08 · Dominik Kapusta:** I wanted to ask. Because we, uh, run this entire thing about passkeys on macOS, but… We didn't check iOS at all, right? But I think… Does it work the same on iOS? I don't recall anymore. We probably also had that entitlement requested for iOS or not.
|
||||
- **20:14:30 · DR:** I don't know, honestly. That's a good question. I'm going to take note now.
|
||||
- **20:14:33 · Dominik Kapusta:** Okay.
|
||||
- **20:14:35 · DR:** To look into that. But it's possible we're going to be seeing the same issue. I don't know, because the entitlement that we would need for that not to be an issue is… I don't think Apple is giving that away, and I don't think we have that on iOS, so it may be an issue there as well. Uh… But yeah, I mean, the request to make the WebKit submission to fix this. Would fix it for all. browsers. But it would be interesting to know if it's breaking in iOS. I agree.
|
||||
- **20:15:09 · Dominik Kapusta:** Yeah, of course. Now that I think of it, I… I'm not looking at iOS feedback that much, but probably we don't have? feedback about passkeys not working in the iOS app?
|
||||
- **20:15:24 · DR:** I did look at some point and I didn't find much from iOS, but then I didn't follow up to check if it was causing trouble.
|
||||
- **20:15:26 · Dominik Kapusta:** There's only one recorder.
|
||||
- **20:15:33 · DR:** Which is the interesting bit.
|
||||
- **20:15:35 · Dominik Kapusta:** Mm-hmm. Cool, thanks. One last thing about extensions is that, and this hasn't been announced anywhere, but we also support Dark Reader. If any one of you fancies Dark Reader, you can install it from Chrome Web Store. Just go to Chrome Web Store, search for Dark Reader and install it and it works. It works much better than whatever we had for Dark Reader, uh, integrated behind the scenes, and that thing is going away at some point. But yeah. We also support iCloud Passwords, which is a terrible extension, but it's the only way to connect a browser that's not Safari. with passwords up. Thank you.
|
||||
- **20:16:23 · Jonathan Jackson:** Awesome, thank you. And thank you, Diego, for rolling with that. If you have tasks or anything that are related to those that you could pop into the agenda at some point, that would help people find some more of this history, too. That'd be great. Uh, if not, that's fine. Very handy, that'd be…
|
||||
- **20:16:36 · DR:** Yeah, I'm gonna… no, I'm gonna look them up now, yeah, and then I'll…
|
||||
- **20:16:40 · Jonathan Jackson:** Cool. Thank you. Feel free to drop them in the agenda. That'd be fine. And then the last thing on the agenda, Alex, is a postmortem from you.
|
||||
- **20:16:50 · Alex Martemyanov:** Yeah, hey, it's the postmortem for VS Code support. So actually the migration to support VS Code and Carstore initially. I'll… So, the… Originally, the project started more than a year ago as a Huggies project. It grew out of… general AUR optimization and refactoring work related to storyboards, removal, modularization, and other. And, uh… Uh, my half-day's effort was to try to integrate it to Cursor in that time, and, uh, it's shown that the approach is actually working. Uh, we could… potentially in, uh, use CorpStore and, uh, integrate with it, having, uh, syntax, uh, highlighting, uh, documentation, and, uh, jump to symbol, builds, and so on, and debugging. Ah, so it's, uh… the project was, uh… formed from the steps remaining to make it work. But first of all, it included all the. AUR work that was already started, like storyboards removal, uh, some of, uh, flags organization and other stuff, and, uh… It had no kickoff as it was kind of running in my background between top priorities, and that's why it took so long to finish. Uh… and uh… Yeah, it had. It had some some issues visibility issues, I'd say so. First of all, did it achieve its success criteria and its purpose? So I think strong. Yes, so. in most of the work. or preparation was finished. Some of it is… It's moved out of the project back to the AUR, but the amount of work needed to make it work was done, and it would have been done in scope of the AUR as the tech debt pending probably someday. Uh, but yeah, this project gave it kind of, um, shape and served as an umbrella for those, uh. I think it's… Vs code. Cursor was completed, at least on my end. It's available in my public repo. If anybody is interested and requires a small github githook to integrate into project. and uses mostly native means, and fully compatible with Xcode, and does not replace or harm Xcode experience. So what went well and I could highlight? Yeah, each of the project steps were shipped as a small and separate refactors. Each one was an improvement on its own. There were several tech designs. First of all, the first tech design of. Modernization and Cursor support was originally discussed and approved when we mostly adopted Cursor, and it seemed like a good target. Uh, and it had a follow-up this year, recently, that was, uh, that was canceled. And… Basically, it was the same tech design, but it was really discussed after… after some… changes and when we build closer to completion. And with EXPO27 release, the integration was much simpler than the original plan, and did not require most controversial parts of the project. So this was discussed with the team, and. Yeah, the decision was made to not to move forward with it, but have a small… Small fixes or small adaptation that allow it to complete. Yeah, so the setup is usable. I use it instead of Xcode, except for release deals or Ui tests. And, uh… Yeah, basically it's moving, it's moving. seemingly moving forward to make… and Apple is moving forward to make, uh, it's more open and more viable with other dev tools, if anybody wants to experiment with it. Uh, and the work, I think, I could also show, highlight that the work is not completed here. Uh, we still have a lot of tech debt to solve, a lot of, uh, modelization efforts to be done there.
|
||||
- **20:21:58 · Alex Martemyanov:** And it continues, basically. What didn't go well? Uh, there was, uh, an incident on macOS 12, uh, because it's generally not tested in our, uh, in our setup, as we already discussed in the incident post-mortem last time. Uh, while we still support, officially support, the SmuckOS version, we are not testing it effectively. And, uh, this, uh, gave me some thoughts of how could it be improved, probably. So, first of all, I… experiment… I experiment locally with, uh, testing… testing harness and agent harness, supporting macOS 12 as well. And, uh… later on, I'll try to… Push it forward to… Okay. Uh, to be able to actually test it. So, first of all, I've made an effort to, uh… to split. with actual UI test builds from macOS runners to one macOS builder, unified between different versions, so runners won't have. I have to build it every time, and this is improvement on its own. It improves build times, basically. And another is testing hardness to support Mac OS 12. So it's not… it's not open as a pull request to review yet, but it's working. Uh, so it's proof that it can work, and we can… could move forward with it, and probably rent some, uh, older machines somewhere in the cloud, and uh… have it tested. The second suggestion here would be to, uh, improve our processes, so I didn't, uh, complete the proposal yet, with all this, uh, performance review stuff, the priority. Uh, but yeah, I was planning to, uh… at a proposal. to require writing UI tests for any modified or added user-facing features and extend coverage overall as it does not require so much effort. We only need some Agentic work to analyze. coverage to suggest some tests and basically push it one by one. Um… Another issue was a tab bar disappearing. It was spotted during the internal testing. Basically, general work. And again, UI testing would support it. Yeah. Next thing is, uh, related to the visibility. Uh, so basically, demand for the VS Code integration and commercial integration was checked too late, as the project was, uh, done as in my background. So in its purpose, its aim was mostly my personal and included most tension, while most of the project was not related to actual VS Code migration, but served as an. Umbrella project to… Granted. Uh, this was, uh… This was only touched on the last stage, and uh… Probably we… While the project was progressing in the days before, we could have discussed it during some of Apple Weekly. But maybe not since. the actual implementation and the actual step that required this discussion was raised just in time, when Xcode 27 was available, and this step was about to be made. So. yeah, there were some issues with visibility, but it's mostly related to my… to me doing this project as my… not as my top priority, but it's generally your work when I have time for it. Okay. Yeah, and another outcome from it is the project goal was from the round one IDE support, as I already told, basically to migrate to VS Code and eat. might separate some tension and questions. Whether do we need to migrate to Vs code while again, 85% of the project was not related to actual Vs code migration and was general refactoring and improvements. It's, yeah, mostly related to the issue number two. of… Basically, basically, that's it, everything we discussed, and uh… I'll…
|
||||
- **20:26:58 · Alex Martemyanov:** push forward with the Ui testing suggest proposal, and. older version testing, which is a requirement, especially with the more and more usage of Agentic workflows, where we have, like, 95 to 98% of. perfect, there are 8-time code, but there is still this 2% that is not perfect, and it tends to fail in the most critical scenarios. And it becomes a requirement, so… it's not the implementation, uh, that needs to be tested more, but the test… actual testing. And, uh, we need to… Obviously, we need to raise our quality assurance. Versuses. Let's see. Yeah, and the remaining factors will be handles. I have them tracked in the AOR. Not much of priority again, but they are on my plate. That's it. Any comments, any questions?
|
||||
- **20:28:14 · Jonathan Jackson:** Cool. Thank you, Alex. For the other refactors and stuff that are going to happen as part of the AOR. I think we just need to make sure that the value of those are clear in their own steps, right? Rather than, like you said, rather than focusing them on a VS Code support item, we'll just talk about each of those and what they bring to the team in their own right.
|
||||
- **20:28:35 · Alex Martemyanov:** Yes, so they were originally part included in this project, but we should separate cleanup targets. So each of the dependencies is not used by multiple targets, but has its own framework, setting up testing harness.
|
||||
- **20:28:40 · Jonathan Jackson:** Yep. Yeah. Yeah, and a lot of that is in line with the modularization desires that we've had anyway, so.
|
||||
- **20:28:53 · Alex Martemyanov:** supporting UI person. Yep, yep.
|
||||
- **20:28:58 · Jonathan Jackson:** Excellent. Cool. Thank you. Appreciate it. Anything else from anyone before we wrap our meeting for the day? Remember that today is the first of three days for career advisors to be writing, uh, assessments for their advisees, so… We will all probably be heads down on that. I guess if you have questions, career advisors, if you have any questions that I can help with, go ahead and let me know. Feel free to reach out. Thanks, everyone. Good to see you all.
|
||||
@@ -278,8 +278,8 @@
|
||||
"note_modified_at": "2026-10-27T14:02:34.482Z",
|
||||
"source_url": "https://us01mynotes.zoom.us/-I3aVYDWTkKZ_VTRLdjGnA",
|
||||
"folder": "2026-10-27 Apple Weekly Team Meeting - Zoom ID -I3aVYDWTkKZ_VTRLdjGnA",
|
||||
"sha256": "eca3ff1650275faccc26a7d030a207788f36bd194f860b5e232560613956f225",
|
||||
"cues": 49
|
||||
"sha256": "e2679b9fc63a1b4244db265a12dcba88b991d0060ea1fc01b02409bd749ea3e2",
|
||||
"cues": 3
|
||||
},
|
||||
"1T8djM8kQaOWNDk-yjDc9A": {
|
||||
"title": "Alex Martemyanov's Zoom Meeting",
|
||||
@@ -362,8 +362,8 @@
|
||||
"note_modified_at": "2026-10-19T12:37:38.903Z",
|
||||
"source_url": "https://us01mynotes.zoom.us/QvxllpCuTTSLGEZp-R6jQw",
|
||||
"folder": "2026-10-19 Michal Smaga's Personal Meeting Room QvxllpCuTTSLGEZp-R6jQw",
|
||||
"sha256": "a4b3c5ce7fafb99496019591da19f9a103e35cf5f37ec42aa8f154021fdebcae",
|
||||
"cues": 133
|
||||
"sha256": "5cf16938df6ebe0110bec8e8b4311fcbe55a764bd2bb72f64faa049ba1777d9e",
|
||||
"cues": 3
|
||||
},
|
||||
"KcmYOq5_SQGPxdkdbhRx9A": {
|
||||
"title": "Dominik Kapusta's Personal Meeting Room",
|
||||
|
||||
@@ -1 +1 @@
|
||||
{"observed_at":"2026-11-19T11:46:31.849Z","scanned":31,"imported":0,"skipped":31,"withoutTranscript":0,"coverage":"complete","sections":[{"name":"Owned by me","pages":5,"notes":86,"complete":true,"terminal_cursor":true},{"name":"Shared with me","pages":1,"notes":0,"complete":true,"terminal_cursor":true},{"name":"Starred","pages":1,"notes":0,"complete":true,"terminal_cursor":true}],"discovered":86,"gaps":[],"recovered_transcripts":0,"policy":"reread_all_owned_shared_and_starred_notes_since_cutoff","since":"2026-09-30"}
|
||||
{"observed_at":"2026-11-19T12:18:13.202Z","scanned":24,"imported":2,"skipped":21,"withoutTranscript":0,"coverage":"partial","sections":[{"name":"Owned by me","pages":5,"notes":86,"complete":true,"terminal_cursor":true},{"name":"Shared with me","pages":1,"notes":0,"complete":true,"terminal_cursor":true},{"name":"Starred","pages":1,"notes":0,"complete":true,"terminal_cursor":true}],"discovered":86,"gaps":["locator.waitFor: Timeout 20000ms exceeded.\nCall log:\n - waiting for locator('#docs-title')\n"],"recovered_transcripts":0,"policy":"reread_all_owned_shared_and_starred_notes_since_cutoff","since":"2026-09-30"}
|
||||
Reference in new issue
Block a user