diff --git a/personal/projects/personal-os/personal-os-action-plan-2026-10-05.md b/personal/projects/personal-os/personal-os-action-plan-2026-10-05.md index 7929ba0d..b0b2eaf4 100644 --- a/personal/projects/personal-os/personal-os-action-plan-2026-10-05.md +++ b/personal/projects/personal-os/personal-os-action-plan-2026-10-05.md @@ -195,6 +195,8 @@ Eligibility проверяет текущую версию своей работ ### E. Процесс 2 — наблюдать рабочий день и доставлять сигналы +Telegram указан в исходном дизайне Тактика. В текущем контуре существующий личный бот Hermes принимает диалог и команду `/personalos`. Personal OS отдельно отправляет туда готовность актуального дневного плана и сигнал завершения рабочего дня. Уведомление включает агрегированные минуты и число действий. Полный план и сверка открываются на рабочем Mac. Приёмка проверяет квитанцию Telegram, отсутствие повтора и запрос плана из личного чата. [Исходный дизайн](files/personal-os-design.md) + ```text ActivityWatch window/AFK/editor/browser + accepted worktree bindings + run events → day_observer раз в 60 с @@ -213,7 +215,6 @@ ActivityWatch window/AFK/editor/browser + accepted worktree bindings + run event - **`break_due`**. Фильтр/порог первой версии: 50 мин непрерывной наблюдаемой активности без AFK ≥5 мин. Когда доставлять: В момент порога, мягкая локальная карточка. Для чего: Напомнить о перерыве; это таймер активности, не вывод о здоровье/усталости. - **`day_end`**. Фильтр/порог первой версии: `now >= work_end`; один сигнал на дату. Когда доставлять: На границе рабочего дня. Для чего: Сохранить незавершённое и ближайшее продолжение для следующего дня. - Утром «Мои задачи» предлагают добровольно отметить настроение и энергию отдельно. Явная низкая энергия или выбранный облегчённый режим снижает дневную нагрузку. Настроение сохраняется как самооценка и не меняет рабочий бюджет само по себе. Приёмка проверяет первый вход утром, сохранение и переход на новый день, а также пересчёт при явном изменении энергии. Для замечания об ухудшении фокуса сопоставлять наблюдаемые переключения, перерасход времени и сигналы из общения с агентами. Анализ общения включается отдельно для согласованного локального источника. Резкость и ругань могут быть дополнительным наблюдением. Одна реплика или отдельное слово не подтверждают усталость, настроение или ухудшение качества работы. Сигнал показывает наблюдаемые причины и предлагает пользователю выбрать перерыв, более простой остаток дня или продолжение. Проверить полезность на реальных днях и отдельно проверить ложные сигналы. @@ -227,6 +228,12 @@ Signal contract: `signal_id`, `kind`, `work/action/run_id`, `source_revision`, ` **Критические входы для старта:** пригодная привязка наблюдаемого интервала к work/action; частый local observer; политика времени/дедупликации; видимый local inbox с подтверждением доставки. Канал первого пилота — локальный экран. Интеграции чатов/почты в этот запуск не входят. +#### Будущий этап — оценка интервала сна + +- Телефон передаёт время подключения и отключения зарядки. Подключение вечером открывает предполагаемый ночной интервал. Утреннее отключение отмечает возможное пробуждение и вызывает одно сообщение «Доброе утро» в личном Telegram. +- Кнопка «Уже проснулся» или ответ на утреннее сообщение подтверждает, что день начался. В журнале отдельно видны время снятия с зарядки, время ответа и статус оценки. Пользователь может поправить время сна и подъёма или убрать запись. +- При дневной зарядке, повторном подключении, пропущенном событии или отсутствии ответа система показывает неполную оценку либо предлагает ручной ввод. Приёмка проверяет эти случаи, обычную ночь через полночь и отсутствие повторных приветствий. Влияние на планирование оценивается отдельно после пилота трекера. + ### F. Процесс 3 — карьерный путь и подсветка подходящих задач ```text diff --git a/personal/projects/personal-os/personal-os-fixes-2026-11-13.md b/personal/projects/personal-os/personal-os-fixes-2026-11-13.md index 0449db7e..6d473419 100644 --- a/personal/projects/personal-os/personal-os-fixes-2026-11-13.md +++ b/personal/projects/personal-os/personal-os-fixes-2026-11-13.md @@ -7,6 +7,7 @@ - Старый план Тактика включает Telegram. Рабочий Hermes `hermes-whale` уже использовал того же бота и принимал Telegram-обновления. В Personal OS отсутствовали транспорт, адрес назначения и квитанция. Через Bot API подтверждён личный чат. Адаптер повторно использует учётные данные Hermes и добавляет исходящие сообщения без второго процесса приёма. Дневной план читается через тот же локальный HTTP-допуск, что и пользовательский интерфейс. На рабочей границе действует независимый путь через начало E5. Исходящие сообщения содержат агрегированные числа и локальную навигацию. - Рабочая отправка принятого плана в 11:54:53 по Бишкеку вернула Telegram message ID `398`. Команда `notify` во второй раз вернула исходную квитанцию. Проверены наличие `/personalos` в каталоге команд настоящего Hermes и ответ локальной команды чтения плана. Команда от владельца и фактическая вечерняя доставка ещё нужны для сквозной приёмки. Прошли 336 тестов Python и браузерный сценарий отдельной квитанции Telegram. Стенд проверил удержание устаревшего плана, отказ группы и сменённого получателя, один сигнал на дату, неизвестный результат и повтор после указанной Telegram задержки. Шестой слайд обновлён, остальные шесть совпали с предыдущей отрисовкой. Квитанция презентации `.local/first-day-presentation-r3/validation.json`. - Разрешение на публикацию тестированных изменений в существующем Windmill записано в `windmill/workspace/AGENTS.md`. Dry-run ограничен объектом E5. Копия предыдущей версии и квитанция подтверждают переход `191 → 192` и оба новых шага договора в опубликованном определении. Диспетчер запустил один рабочий прогон `83466a1b-8dd2-4544-94cb-cf5095784b91`. Квитанции `windmill-e5-{before,after}-short-request-20261117.json`. +- В 12:03:36 по Бишкеку `83466a1b-8dd2-4544-94cb-cf5095784b91` завершился успешно. Windmill, локальная публикация и рабочий HTTP показали один расчёт `current`. Из 55 минут 10 отведены на срочный ответ Бартека с PR #7784 и прошедшим CI. Запрос о краше находится в ожидании ревью PR #8212 со связанной сессией Claude. Недельный вход проверен опубликованной версией и удержан по `typed_sources_changed`. Новый цикл `48a5fd2b-21f9-4004-8f1c-b9817c41cec3` уже шёл. Следующий шаг по недельному расчёту — получить допуск по совпавшим ревизиям. Повторной отправки Telegram не было. Рабочий API приёмки отдаёт квитанцию №398 отдельно от macOS. ## 17 ноября, после 11:38 по Бишкеку — A01/A06, адресный ответ и доказательство работы по PR diff --git a/personal/projects/personal-os/personal-os-production-acceptance.md b/personal/projects/personal-os/personal-os-production-acceptance.md index a37f3cf7..76fbb551 100644 --- a/personal/projects/personal-os/personal-os-production-acceptance.md +++ b/personal/projects/personal-os/personal-os-production-acceptance.md @@ -8,10 +8,14 @@ - Параллельно закрываются технические проверки A01–A17. Для Outlook требуется реальный ответ free/busy после разблокировки Mac. Для обсуждений и Claims нужны квитанции актуального метода и точного входа, подтверждение сохранения решений и разбор оставшихся удержаний. Регистрируются выбранный подход, причина выбора, состояние до исправления, проверка после и оставшийся пробел. - После выполнения критериев допуска провести полное ревью решения и эксплуатации, устранить замечания, повторить затронутые сценарии и сквозную проверку. Только после этого подготовить презентацию и провести демонстрацию заказчику с нуля. Она показывает назначение системы, путь данных, рабочий день от получения плана до вечерней сверки, действия в интерфейсе, историю, недельный обзор и диагностику сбоев. В конце предъявляются доказательства готовности и известные ограничения. Расписание Codex отключается после сдачи, включая презентацию. +## 17 ноября, 12:06 по Бишкеку — будущий трекер сна + +- В план после текущего пилота добавлена оценка ночного интервала по зарядке телефона с утренним подтверждением в личном Telegram и ручной правкой. Критерии включают пропущенные события, дневную зарядку и повторные подключения. [План трекера сна](personal-os-action-plan-2026-10-05.md#будущий-этап--оценка-интервала-сна) + ## 17 ноября, после 11:55 по Бишкеку — Telegram и публикация E5 - В исходном дизайне Тактика предусмотрены диалог и утренний план в Telegram. Существующий Hermes уже слушал личный чат, Personal OS не передавал ему уведомления. Подключён проверенный личный адрес. Telegram принял сообщение об актуальном плане в 11:54, ID `398`. Повтор использовал ту же квитанцию. Входящую команду `/personalos` видит шлюз Hermes. Ответ владельцу на реальную команду и вечернее сообщение предстоит подтвердить. [Инструкция первого дня](personal-os-first-day.md) -- По постоянному разрешению владельца опубликован только `f/attend/attention_loop_e5`, версия `191 → 192`. Оба прохода содержат новые правила срочных запросов и связь с недельным планом. Рабочий пересчёт `83466a1b-8dd2-4544-94cb-cf5095784b91` запущен через штатный диспетчер. [Журнал исправлений](personal-os-fixes-2026-11-13.md) +- По постоянному разрешению владельца опубликован только `f/attend/attention_loop_e5`, версия `191 → 192`. Оба прохода содержат новые правила срочных запросов и связь с недельным планом. Штатный `83466a1b-8dd2-4544-94cb-cf5095784b91` завершился `current`. В рабочем API ответ Бартека занял 10 минут и помечен срочным, краш находится в ожидании ревью #8212 со ссылкой на Claude. Недельное распределение на этом проходе удержано по `typed_sources_changed`. [Журнал исправлений](personal-os-fixes-2026-11-13.md) ## 17 ноября, после 11:38 по Бишкеку — проверка двух адресных запросов Asana diff --git a/work/zoom-transcripts/index.json b/work/zoom-transcripts/index.json index b1cd0955..016faecc 100644 --- a/work/zoom-transcripts/index.json +++ b/work/zoom-transcripts/index.json @@ -122,8 +122,8 @@ "note_modified_at": "2026-11-12T10:14:13.001Z", "source_url": "https://us01mynotes.zoom.us/rJFn9D3lRJmnlCMSomhPsw", "folder": "undated Michal Smaga's Personal Meeting Room rJFn9D3lRJmnlCMSomhPsw", - "sha256": "55727e14d28c8dfe03f20623c696d1e756da52b8e985ccafe454a70f04c58693", - "cues": 60 + "sha256": "95c17a300c3de74483cc899f8d8cb9b878ee47d0bd952ff62c5f89c507c3c5ca", + "cues": 3 }, "DLwAbUQ_RZ6X_D63xiX_Ww": { "title": "Max Tsoy's Personal Meeting Room", diff --git a/work/zoom-transcripts/undated Michal Smaga's Personal Meeting Room rJFn9D3lRJmnlCMSomhPsw/transcript.json b/work/zoom-transcripts/undated Michal Smaga's Personal Meeting Room rJFn9D3lRJmnlCMSomhPsw/transcript.json index 79921b2b..d4567fcd 100644 --- a/work/zoom-transcripts/undated Michal Smaga's Personal Meeting Room rJFn9D3lRJmnlCMSomhPsw/transcript.json +++ b/work/zoom-transcripts/undated Michal Smaga's Personal Meeting Room rJFn9D3lRJmnlCMSomhPsw/transcript.json @@ -23,291 +23,6 @@ "time": "16:15:28", "speaker": "Michal Smaga", "text": "Okay. So without matching, we may not see all the issues. Because it's, like…" - }, - { - "time": "16:15:36", - "speaker": "Alex Martemyanov", - "text": "I don't… Yes, uh…" - }, - { - "time": "16:15:39", - "speaker": "Michal Smaga", - "text": "Because if you have two tabs for the same domain, one may load, and if one fails, you will get, like, the the error won't be reported. Right?" - }, - { - "time": "16:15:47", - "speaker": "Alex Martemyanov", - "text": "Yes, kind of it. So basically, we could keep the same matching algorithm if we receive it twice. for the same URL, just consider it healthy, and, uh, not report. It would… it would work, basically. So… But yeah, this is an extra mile." - }, - { - "time": "16:16:07", - "speaker": "Michal Smaga", - "text": "And in the matching… And the matching would require, like, the web extension change as well, right?" - }, - { - "time": "16:16:14", - "speaker": "Alex Martemyanov", - "text": "Yes. So, the changes to send the… to send the window ID and all the window IDs." - }, - { - "time": "16:16:16", - "speaker": "Michal Smaga", - "text": "Just to report the…" - }, - { - "time": "16:16:22", - "speaker": "Alex Martemyanov", - "text": "In the sequence that we've got them." - }, - { - "time": "16:16:25", - "speaker": "Michal Smaga", - "text": "Yeah, yeah." - }, - { - "time": "16:16:26", - "speaker": "Alex Martemyanov", - "text": "So we could match to arrays. in how the web extension IDs given in order, and how they were accessed from the upside in the same order, but different IDs." - }, - { - "time": "16:16:35", - "speaker": "Michal Smaga", - "text": "Okay. Yeah, but this is kind of, like, an optional, or do you suggest doing that? Okay." - }, - { - "time": "16:16:44", - "speaker": "Alex Martemyanov", - "text": "Yeah, yeah, this is an optional. I'm gonna comment in the design for matching, and it probably won't be included in the pull request I'm gonna open soon." - }, - { - "time": "16:16:50", - "speaker": "Michal Smaga", - "text": "Okay." - }, - { - "time": "16:16:56", - "speaker": "Alex Martemyanov", - "text": "So, currently, there is two… To ready ones." - }, - { - "time": "16:17:00", - "speaker": "Michal Smaga", - "text": "Two PRs, yeah." - }, - { - "time": "16:17:03", - "speaker": "Alex Martemyanov", - "text": "To be ours mostly or not to. I think it's only one, right? Second one is not there yet." - }, - { - "time": "16:17:13", - "speaker": "Michal Smaga", - "text": "No, there are two." - }, - { - "time": "16:17:15", - "speaker": "Alex Martemyanov", - "text": "There is two. Okay, probably the second one maybe not pushed. I'll just check a bit later." - }, - { - "time": "16:17:22", - "speaker": "Michal Smaga", - "text": "There's remove diagnostics and remove ineffective fixes." - }, - { - "time": "16:17:33", - "speaker": "Alex Martemyanov", - "text": "Yeah, let me just check it quickly. If it's up to date. Yeah, so remove, uh, remove fixes, the first one. It removes the feature flags and the fixes that weren't there, so it's mostly removal. It shouldn't be nothing complicated. Yeah, the second one is also actual. It removes the metrics. So, it's meant to be merged." - }, - { - "time": "16:18:06", - "speaker": "Michal Smaga", - "text": "Okay, so… so there will be, like, a third… there will be a third one for the… for the permanent ones. Okay." - }, - { - "time": "16:18:09", - "speaker": "Alex Martemyanov", - "text": "There will be third one re-adding. Yes, yes. So, and I'm…" - }, - { - "time": "16:18:13", - "speaker": "Michal Smaga", - "text": "I've started setting up the testing for the PR, so I'll be reviewing it shortly, so we have it merged soon. Uh, yeah, other than that, I will catch up on the tech design and to see, like, uh, if I understood them correctly. And yeah." - }, - { - "time": "16:18:31", - "speaker": "Alex Martemyanov", - "text": "Okay, okay, great. Thanks." - }, - { - "time": "16:18:33", - "speaker": "Michal Smaga", - "text": "So do you have anything, like, blocking you for moving forward to the third one, or…" - }, - { - "time": "16:18:40", - "speaker": "Alex Martemyanov", - "text": "Not really. Uh, so I was investigating, uh, how to… how to match it exactly, how far should we go. Probably we shouldn't, uh, go with it on the start PR." - }, - { - "time": "16:18:40", - "speaker": "Michal Smaga", - "text": "We need to…" - }, - { - "time": "16:18:52", - "speaker": "Alex Martemyanov", - "text": "And it could be totally optional." - }, - { - "time": "16:18:56", - "speaker": "Michal Smaga", - "text": "Yeah. I I I would discover the the index thing for, like, it's good that you have, like, research. It's so, like, doing, like, some kind of, like, a brain dump of, like, the approach or, like, having, like, some spec for it. I think, maximum." - }, - { - "time": "16:18:56", - "speaker": "Alex Martemyanov", - "text": "And, yeah." - }, - { - "time": "16:19:13", - "speaker": "Michal Smaga", - "text": "Drop the link to, like, earlier discussion that we had, like, earlier this year, like, uh… We were figuring out that it wasn't, like, easily possible, so we we kind of, like, moved away from it. So maybe it would be worthy to to to, like, follow up it separately." - }, - { - "time": "16:19:28", - "speaker": "Alex Martemyanov", - "text": "Mm-hmm." - }, - { - "time": "16:19:29", - "speaker": "Michal Smaga", - "text": "Uh, on Tuesday, like, Russell was asking about the scope of the project and, like, when we finish up, so I would probably, like. Focus on like, like, like. Basically, focus on wrapping stuff up and, like, probably having a post-mortem, like, next week. So so I think this is, like, of a priority. And, yeah, and setting up the, like, the yeah. Just just wrap up the project, and probably we can do some additional work after that." - }, - { - "time": "16:20:00", - "speaker": "Alex Martemyanov", - "text": "Yeah, okay. Okay, so I will just, my main question or my main point here was to think on the metric and how it works. So basically, we load a page, and uh… we report it after we close the tab, or if we navigate somewhere else. And if… The page was alive, was up for more than 30 seconds. So we had enough time for extension to load, and it didn't receive any CPM messages. It would be reported as broken. you… It should probably give us a general idea if how the trend is going." - }, - { - "time": "16:20:43", - "speaker": "Michal Smaga", - "text": "And if you close the tab or navigate away within the 30 seconds, we don't report it, like, for nothing, right?" - }, - { - "time": "16:20:52", - "speaker": "Alex Martemyanov", - "text": "Oh. I think so. Yeah." - }, - { - "time": "16:20:59", - "speaker": "Michal Smaga", - "text": "I'll check… yeah, I'll try to give you the feedback on the tech design, so probably, like, this is something that we want to…" - }, - { - "time": "16:20:59", - "speaker": "Alex Martemyanov", - "text": "So we can…" - }, - { - "time": "16:21:07", - "speaker": "Michal Smaga", - "text": "like, clarify this week, so, like, probably Monday you can, like, implement the… Sorry. Eat precisely, and then we are good, I think." - }, - { - "time": "16:21:18", - "speaker": "Alex Martemyanov", - "text": "Yeah, so it's most ready here locally, and I think today, later today, I'm going to open. Open it up and ping you and update the tech designs. So, yeah, there is some exception for the reporting, like a page or private pages." - }, - { - "time": "16:21:30", - "speaker": "Michal Smaga", - "text": "Okay." - }, - { - "time": "16:21:37", - "speaker": "Alex Martemyanov", - "text": "Alright. Or session restoration, or back navigation, or history navigation, so basically the ones that, uh, loaded from cache, and we had some… problems with them and race conditions." - }, - { - "time": "16:21:50", - "speaker": "Michal Smaga", - "text": "Okay." - }, - { - "time": "16:21:50", - "speaker": "Alex Martemyanov", - "text": "And then… But for general navigations, and yeah, so there's page lifetime, let's… It should be more than 30 seconds. If we had it two times in a row for different tabs, it means, like, permanent breakage of the extension." - }, - { - "time": "16:22:07", - "speaker": "Michal Smaga", - "text": "And for the page restoration, do you detect it and act differently?" - }, - { - "time": "16:22:07", - "speaker": "Alex Martemyanov", - "text": "Should give… Uh, it's me, uh… So we have this wait time before navigation, so we expect the extension to load. And it's limited for 5 seconds, but if it takes longer, it may still load after the page loads from cache, and it means it won't go through the page deload path and… Won't activate the extension. So, the… CPM won't be fired up and CUTIBRANCE won't be handled sometimes." - }, - { - "time": "16:22:46", - "speaker": "Michal Smaga", - "text": "Mm-hmm. Yeah." - }, - { - "time": "16:22:48", - "speaker": "Alex Martemyanov", - "text": "Thank you. It was, uh, like, our best effort, and it seems unreasonable to keep it… to keep it waiting longer than 5 seconds for users, uh, to extend this type of… this startup time. But there are still some reports. It seems like it went slightly down, but not really fixing all the cases. But those are not permanent, mostly." - }, - { - "time": "16:23:18", - "speaker": "Michal Smaga", - "text": "Okay, yeah. Hmmm…" - }, - { - "time": "16:23:25", - "speaker": "Alex Martemyanov", - "text": "Yeah, that should be it, I guess." - }, - { - "time": "16:23:29", - "speaker": "Michal Smaga", - "text": "Okay. So, I'm off to to testing the PRs and and let you know." - }, - { - "time": "16:23:33", - "speaker": "Alex Martemyanov", - "text": "Mm-hmm." - }, - { - "time": "16:23:34", - "speaker": "Michal Smaga", - "text": "Cool. If you have anything I can help you with, uh, just feel free to ping me. And…" - }, - { - "time": "16:23:35", - "speaker": "Alex Martemyanov", - "text": "Okay, thanks." - }, - { - "time": "16:23:41", - "speaker": "Michal Smaga", - "text": "Maybe if there's, like, we need another discussion with Max or whatever, like, we can, like, chat later or tomorrow morning." - }, - { - "time": "16:23:50", - "speaker": "Alex Martemyanov", - "text": "Sure, I'll ping him the task center. Thank you, Anthony." - }, - { - "time": "16:23:53", - "speaker": "Michal Smaga", - "text": "Okay, cool. Thank you." - }, - { - "time": "16:23:56", - "speaker": "Alex Martemyanov", - "text": "Okay. Yep. Thank you." } ] } diff --git a/work/zoom-transcripts/undated Michal Smaga's Personal Meeting Room rJFn9D3lRJmnlCMSomhPsw/transcript.md b/work/zoom-transcripts/undated Michal Smaga's Personal Meeting Room rJFn9D3lRJmnlCMSomhPsw/transcript.md index e9023fdf..17c56b09 100644 --- a/work/zoom-transcripts/undated Michal Smaga's Personal Meeting Room rJFn9D3lRJmnlCMSomhPsw/transcript.md +++ b/work/zoom-transcripts/undated Michal Smaga's Personal Meeting Room rJFn9D3lRJmnlCMSomhPsw/transcript.md @@ -15,60 +15,3 @@ source: "https://us01mynotes.zoom.us/rJFn9D3lRJmnlCMSomhPsw" - **16:14:12 · Michal Smaga:** read the text. designs yet. So I have, like, maybe a dumb question. Like, why is the the the, like, figuring out the specific top index and, like, matching the tops required for that? - **16:14:23 · Alex Martemyanov:** Basically, I was trying to make it as stable as possible to… To make it, like, live permanently there. And, uh, as Max told, we have, uh, some problem on, or maybe false positives, uh, for, uh, unmatched tab, so the matching is done by URL, basically. And if you open the same tab twice, with the same URL, it would be ambiguous at that point. And uh… so in the previous metrics, it's used to… Um, reports, like, uh, healthy, if, uh, we have… two consequently awaiting tabs. So, one could be unhealthy, but we won't report it anyways if we receive the message twice. But to make it kind of 100% bulletproof, and as well in other places, if we match exactly the web extension tab against the actual tab, it would. - **16:15:28 · Michal Smaga:** Okay. So without matching, we may not see all the issues. Because it's, like… -- **16:15:36 · Alex Martemyanov:** I don't… Yes, uh… -- **16:15:39 · Michal Smaga:** Because if you have two tabs for the same domain, one may load, and if one fails, you will get, like, the the error won't be reported. Right? -- **16:15:47 · Alex Martemyanov:** Yes, kind of it. So basically, we could keep the same matching algorithm if we receive it twice. for the same URL, just consider it healthy, and, uh, not report. It would… it would work, basically. So… But yeah, this is an extra mile. -- **16:16:07 · Michal Smaga:** And in the matching… And the matching would require, like, the web extension change as well, right? -- **16:16:14 · Alex Martemyanov:** Yes. So, the changes to send the… to send the window ID and all the window IDs. -- **16:16:16 · Michal Smaga:** Just to report the… -- **16:16:22 · Alex Martemyanov:** In the sequence that we've got them. -- **16:16:25 · Michal Smaga:** Yeah, yeah. -- **16:16:26 · Alex Martemyanov:** So we could match to arrays. in how the web extension IDs given in order, and how they were accessed from the upside in the same order, but different IDs. -- **16:16:35 · Michal Smaga:** Okay. Yeah, but this is kind of, like, an optional, or do you suggest doing that? Okay. -- **16:16:44 · Alex Martemyanov:** Yeah, yeah, this is an optional. I'm gonna comment in the design for matching, and it probably won't be included in the pull request I'm gonna open soon. -- **16:16:50 · Michal Smaga:** Okay. -- **16:16:56 · Alex Martemyanov:** So, currently, there is two… To ready ones. -- **16:17:00 · Michal Smaga:** Two PRs, yeah. -- **16:17:03 · Alex Martemyanov:** To be ours mostly or not to. I think it's only one, right? Second one is not there yet. -- **16:17:13 · Michal Smaga:** No, there are two. -- **16:17:15 · Alex Martemyanov:** There is two. Okay, probably the second one maybe not pushed. I'll just check a bit later. -- **16:17:22 · Michal Smaga:** There's remove diagnostics and remove ineffective fixes. -- **16:17:33 · Alex Martemyanov:** Yeah, let me just check it quickly. If it's up to date. Yeah, so remove, uh, remove fixes, the first one. It removes the feature flags and the fixes that weren't there, so it's mostly removal. It shouldn't be nothing complicated. Yeah, the second one is also actual. It removes the metrics. So, it's meant to be merged. -- **16:18:06 · Michal Smaga:** Okay, so… so there will be, like, a third… there will be a third one for the… for the permanent ones. Okay. -- **16:18:09 · Alex Martemyanov:** There will be third one re-adding. Yes, yes. So, and I'm… -- **16:18:13 · Michal Smaga:** I've started setting up the testing for the PR, so I'll be reviewing it shortly, so we have it merged soon. Uh, yeah, other than that, I will catch up on the tech design and to see, like, uh, if I understood them correctly. And yeah. -- **16:18:31 · Alex Martemyanov:** Okay, okay, great. Thanks. -- **16:18:33 · Michal Smaga:** So do you have anything, like, blocking you for moving forward to the third one, or… -- **16:18:40 · Alex Martemyanov:** Not really. Uh, so I was investigating, uh, how to… how to match it exactly, how far should we go. Probably we shouldn't, uh, go with it on the start PR. -- **16:18:40 · Michal Smaga:** We need to… -- **16:18:52 · Alex Martemyanov:** And it could be totally optional. -- **16:18:56 · Michal Smaga:** Yeah. I I I would discover the the index thing for, like, it's good that you have, like, research. It's so, like, doing, like, some kind of, like, a brain dump of, like, the approach or, like, having, like, some spec for it. I think, maximum. -- **16:18:56 · Alex Martemyanov:** And, yeah. -- **16:19:13 · Michal Smaga:** Drop the link to, like, earlier discussion that we had, like, earlier this year, like, uh… We were figuring out that it wasn't, like, easily possible, so we we kind of, like, moved away from it. So maybe it would be worthy to to to, like, follow up it separately. -- **16:19:28 · Alex Martemyanov:** Mm-hmm. -- **16:19:29 · Michal Smaga:** Uh, on Tuesday, like, Russell was asking about the scope of the project and, like, when we finish up, so I would probably, like. Focus on like, like, like. Basically, focus on wrapping stuff up and, like, probably having a post-mortem, like, next week. So so I think this is, like, of a priority. And, yeah, and setting up the, like, the yeah. Just just wrap up the project, and probably we can do some additional work after that. -- **16:20:00 · Alex Martemyanov:** Yeah, okay. Okay, so I will just, my main question or my main point here was to think on the metric and how it works. So basically, we load a page, and uh… we report it after we close the tab, or if we navigate somewhere else. And if… The page was alive, was up for more than 30 seconds. So we had enough time for extension to load, and it didn't receive any CPM messages. It would be reported as broken. you… It should probably give us a general idea if how the trend is going. -- **16:20:43 · Michal Smaga:** And if you close the tab or navigate away within the 30 seconds, we don't report it, like, for nothing, right? -- **16:20:52 · Alex Martemyanov:** Oh. I think so. Yeah. -- **16:20:59 · Michal Smaga:** I'll check… yeah, I'll try to give you the feedback on the tech design, so probably, like, this is something that we want to… -- **16:20:59 · Alex Martemyanov:** So we can… -- **16:21:07 · Michal Smaga:** like, clarify this week, so, like, probably Monday you can, like, implement the… Sorry. Eat precisely, and then we are good, I think. -- **16:21:18 · Alex Martemyanov:** Yeah, so it's most ready here locally, and I think today, later today, I'm going to open. Open it up and ping you and update the tech designs. So, yeah, there is some exception for the reporting, like a page or private pages. -- **16:21:30 · Michal Smaga:** Okay. -- **16:21:37 · Alex Martemyanov:** Alright. Or session restoration, or back navigation, or history navigation, so basically the ones that, uh, loaded from cache, and we had some… problems with them and race conditions. -- **16:21:50 · Michal Smaga:** Okay. -- **16:21:50 · Alex Martemyanov:** And then… But for general navigations, and yeah, so there's page lifetime, let's… It should be more than 30 seconds. If we had it two times in a row for different tabs, it means, like, permanent breakage of the extension. -- **16:22:07 · Michal Smaga:** And for the page restoration, do you detect it and act differently? -- **16:22:07 · Alex Martemyanov:** Should give… Uh, it's me, uh… So we have this wait time before navigation, so we expect the extension to load. And it's limited for 5 seconds, but if it takes longer, it may still load after the page loads from cache, and it means it won't go through the page deload path and… Won't activate the extension. So, the… CPM won't be fired up and CUTIBRANCE won't be handled sometimes. -- **16:22:46 · Michal Smaga:** Mm-hmm. Yeah. -- **16:22:48 · Alex Martemyanov:** Thank you. It was, uh, like, our best effort, and it seems unreasonable to keep it… to keep it waiting longer than 5 seconds for users, uh, to extend this type of… this startup time. But there are still some reports. It seems like it went slightly down, but not really fixing all the cases. But those are not permanent, mostly. -- **16:23:18 · Michal Smaga:** Okay, yeah. Hmmm… -- **16:23:25 · Alex Martemyanov:** Yeah, that should be it, I guess. -- **16:23:29 · Michal Smaga:** Okay. So, I'm off to to testing the PRs and and let you know. -- **16:23:33 · Alex Martemyanov:** Mm-hmm. -- **16:23:34 · Michal Smaga:** Cool. If you have anything I can help you with, uh, just feel free to ping me. And… -- **16:23:35 · Alex Martemyanov:** Okay, thanks. -- **16:23:41 · Michal Smaga:** Maybe if there's, like, we need another discussion with Max or whatever, like, we can, like, chat later or tomorrow morning. -- **16:23:50 · Alex Martemyanov:** Sure, I'll ping him the task center. Thank you, Anthony. -- **16:23:53 · Michal Smaga:** Okay, cool. Thank you. -- **16:23:56 · Alex Martemyanov:** Okay. Yep. Thank you.