Repository navigation
«Вибрати всі» там, где отмечали по одному: каталог методик с действиями над отмеченными, админы групп, права ролей, отзыв доступа - #226
Conversation
…ми над отмеченными, админы групп, права ролей, отзыв доступа, наборов и приглашений Каталог методик: отметка строк, «Вибрати всі N» по всему виду каталога, а не странице; над отмеченными — «Призначити», «Перемістити» в группу и папку, «Зняти з використання» (окно со списком и подтверждением числом) и «Повернути в роботу». Отказ по одной методике не отменяет остальных (eachOf в ui/action.ts). Там же, где добавляли и снимали по одному: - группа пациентов: «Додати всіх» найденных одним запросом; - ящик «Призначити»: «Додати всіх знайдених»; - админы группы методик: ящик с галочками и «Вибрати всіх» вместо селекта на одного; - матрица прав: галочка раздела у роли ставит или снимает все его показанные права; - доступ к методике, назначения набора, приглашения: отметка строк и «Відкликати» отмеченным; - нормы: все шкалы, которым хватает выборки; - ключ шкалы: номера перечнем и диапазоном («1-35, 41»); - поиск по записям: «Усі» над авторами и видами; - статистика: «Усі / Жодної» у вариантов ответа. Эталон экрана каталога переснят под колонку отметок (macOS).
…compose.yml; ответы API те же
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: e4f54c396a
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| assignOne={async (userId, due, note) => { | ||
| for (const s of assignable) await api.grant(s.id, userId, note ?? undefined, due); | ||
| }} |
There was a problem hiding this comment.
If one selected survey is retired or becomes inaccessible after selection, its grant request rejects and this loop skips every subsequent survey for that patient. AssignRecipients catches failures per patient, not per survey, so it reports the patient as wholly unsuccessful even when earlier grants succeeded and provides no indication of which surveys were omitted. The group-assignment loop below has the same problem. Track outcomes per survey and continue after individual failures so one rejected methodology does not prevent the remaining assignments.
Useful? React with 👍 / 👎.
| const plan = movePlan(s, groupId, folderId || null); | ||
| if (plan.group) await api.updateSurvey(s.id, { groupId }); | ||
| if (plan.folder) await api.moveSurvey(s.id, folderId || null); |
There was a problem hiding this comment.
When the group PATCH succeeds but the folder request fails, the survey has already moved to the new group's root, yet left and the parent selection retain its original group and folder. Subsequent attempts compute movePlan from that obsolete state. For example, choosing the original group and root to recover can disable submission as an apparent no-op, or report success without actually restoring the group. Record the successful group change in the retained item, or reload failed items before allowing another move.
Useful? React with 👍 / 👎.
| for (const part of raw.split(/[\s,;]+/)) { | ||
| const m = /^(\d+)(?:[-–—](\d+))?$/.exec(part); |
There was a problem hiding this comment.
When a clinician pastes a range containing spaces, such as 1 – 10, splitting on whitespace first turns it into three tokens and silently adds only items 1 and 10. This produces an incomplete scoring key rather than the requested range, and the missing intermediate entries can change calculated scores without making the key structurally invalid. Normalize spaces around supported dashes before tokenization, or reject the malformed range explicitly instead of accepting its endpoints.
Useful? React with 👍 / 👎.
…еноса — от нового места, номера ключа — без молчаливых потерь - «Призначити» отмеченным: отказ по одной методике больше не обрывает остальные для того же человека или группы; не прошедшие называются кодом и причиной. Ящик печатает причину отказа рядом с именем (assignEach отдаёт её по id). - «Перемістити» отмеченные: если группа сменилась, а папка нет, повтор считается от корня новой группы; каталог перечитывается и в этом случае, а строки отмеченных берутся свежими из него. - Ключ шкалы: «1 – 10» с пробелами — диапазон, а не два номера; неразобранное не добавляется вовсе — поле остаётся открытым и называет, что не так; номер за концом методики не отбрасывается, а показывается янтарём, как набранный по одному.
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 0b1cfeff84
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| if (!m || from < 1 || to - from >= 1000) { | ||
| bad.push(token); | ||
| continue; | ||
| } | ||
| for (let n = from; n <= to; n++) out.add(n); |
There was a problem hiding this comment.
If a pasted item number is 9007199254740992 or larger, it can pass the current validation but make this loop run forever: JavaScript cannot reliably increment integers beyond its safe-integer limit, so n++ leaves n unchanged. Since parsing runs synchronously on Enter or blur, this freezes the constructor tab and can lose unsaved edits. Validate both endpoints with Number.isSafeInteger before expanding the range and return unsupported numbers in bad.
Useful? React with 👍 / 👎.
…целых разбор диапазона висел вечно и вешал вкладку конструктора (замечание ревью)
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: e3cf9c150f
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| const assignEachSurvey = async (act: (s: SurveyListItem) => Promise<unknown>) => { | ||
| const { failed } = await eachOf(assignable, act); | ||
| if (!failed.length) return; | ||
| const list = failed.map((f) => `${splitCode(f.item.title).code ?? f.item.title}${f.reason ? `: ${f.reason}` : ""}`).join("; "); | ||
| throw new Error(ut(failed.length < assignable.length ? "mc.assignManyPartly" : "mc.assignManyNone").replace("{list}", list)); |
There was a problem hiding this comment.
When one survey assignment fails after another succeeds, the drawer retains the recipient, but retrying runs every survey in assignable again. These requests are not idempotent even with the same deadline: apps/api/src/lib/grantAccess.ts resets grantedAt and attemptsUsed on every re-grant. If the patient completes the successful survey before the clinician retries the failed one, that retry grants another attempt and moves the existing response outside the current grant's attempt-counting window. Track successful recipient–survey pairs and retry only failed pairs; the group-assignment path needs the same protection.
Useful? React with 👍 / 👎.
…т × методика» — повторная выдача обнуляет счёт попыток, и пройденная между попытками методика получила бы ещё одну (замечание ревью)
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 97de41b8ba
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| const of = assignable.map((s) => s.id).join(","); | ||
| if (assigned.current.of !== of) assigned.current = { of, pairs: new Set(), recipients: new Map() }; | ||
| const { failed } = await assignPending(assignable, target, assigned.current.pairs, act); | ||
| if (!failed.length) return; |
There was a problem hiding this comment.
After a bulk assignment succeeds, AssignRecipients resets and closes the drawer, but this cache retains every successful pair until a different survey-ID sequence is submitted. If the clinician subsequently assigns the same selected surveys to the same patient with the same deadline and comment—for example, to authorize another attempt after completion—assignPending skips every request while the drawer still reports success. Reassigning a patient group also skips newly added members in this situation. Scope the successful-pair cache to an unfinished assignment operation, preserving it for partial-failure retries but clearing completed operations.
Useful? React with 👍 / 👎.
…и — адресат, которому дошло всё, забывается, и следующее назначение тем же людям (ещё одна попытка после прохождения, новые участники группы) снова выдаётся (замечание ревью)
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 97f7553b39
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| onClick={() => { | ||
| /* новая операция: пары прошлой, брошенной на полпути, к ней не относятся */ | ||
| assigned.current = { of: "", pairs: new Set(), recipients: new Map() }; | ||
| setAssignMany(true); |
There was a problem hiding this comment.
After a partially successful assignment, closing the drawer and reopening it through the bulk “Assign” button clears the successful-pair cache, so submitting the retained recipient retries surveys that already succeeded. The fresh evidence is that this handler resets assigned.current, while the mounted AssignRecipients preserves its failed recipients, deadline, and comment across closing. As noted in the earlier retry finding, re-granting resets grantedAt and attemptsUsed, potentially authorizing another attempt if the patient completed a successful survey meanwhile. Preserve the cache when resuming that draft; clear it only when the operation completes or the draft is explicitly discarded.
Useful? React with 👍 / 👎.
…— память о прошедших парах живёт вместе с черновиком ящика и стирается, когда адресату дошло всё или ящик снят со страницы, а не при новом нажатии «Призначити» и не при смене набора отмеченных (замечание ревью)
|
@codex review |
|
Codex Review: Didn't find any major issues. Breezy! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
Там, где приходилось отмечать или добавлять по одному, появились «Вибрати всі» и действия над отмеченными.
Каталог методик
Другие экраны
Чего нет и почему
Как устроено
eachOfвui/action.ts), и каждая запись пишется в журнал как раньше. Пачкой ходит только добавление в группу пациентов, у него свой маршрут.Проверки
e2e-linux.compose.yml. Ответы API не менялись.