create: Merge from master to dev#4744
Conversation
Translations update from Hosted Weblate
Merge Dev to Main
Fix: chim-1158 sidebar routing
Fix: Campaign listing mobile browser zoom and assignment filtering
Root cause: partners is in text[], but the query applied scalar ilike, causing error
fix: Program Tab Filter Is Not Working
Root cause: the percentage fields are calculated values, not program_metrics database columns.
fix: Program Listing Filters Are Not Working
Fix: Restrict program creation access for program managers
Return uniqueSubjects as part of the campaign assignments response and consume it in the UI. CampaignAssignmentTab no longer calls getCampaignSubjectsByCampaignId; it uses response.uniqueSubjects and clears subjects on error. ServiceApi types updated to include uniqueSubjects. SupabaseApi parses unique_subjects from the RPC result, maps them, and adjusts total calculation. Database types updated for unique_subjects. Tests updated to reflect the new payload shape and behavior.
fix: disable sorting for school performance column for ops users in school list
fix: Message Editing Remains Enabled After Campaign Cancellation.
Include uniqueSubjects in campaign assignments
Fetch available grades from classes linked to the selected schools instead of showing all program grades. This ensures users only see grade options that are actually available for their school selection. Changes include: - Added getCampaignGradesForSchools API method to fetch grades by school IDs - Implemented lazy loading of grades when schools are selected - Filter selected grades to only include available options after schools change - Add close-on-click-away behavior to grade dropdown for better UX - Update tests to verify grade filtering and selection behavior
fix: Load grades dynamically based on selected schools
Update CampaignSetupActions disableNext logic to handle activeStep === 4. Previously other steps defaulted to false, which allowed advancing during submission on step 4. Now disableNext uses isSubmitting for step 4 to prevent duplicate submits. (File: src/ops-console/components/campaignSetup/CampaignSetupActions.tsx)
…ession restoration logic in SupabaseAuth
fix: Add Subjects to School button is missing for OPS user in Teacher mode
Disable Next while submitting on step 4
Chim 1187 selectmode school flow
refactor: improve refresh token persistence and implement automatic session restoration logic in SupabaseAuth
…udents client-side
fix: fetching classes and class users separately, then aggregating students client-side
CHIM-1187 Preserve school profile selection flow
Added API to get the schools related to USer
…ate Parent WhatsApp Invitation feature.
fix: Optimize campaign reach using batched school metrics queries and campaign reach queries
…" and "Alternate Weeks (≥2)" Schedules
…" and "Alternate Weeks (≥2)" Schedules
fix: Campaign Message Tab Displays Incorrect Data for "Alternate Days" and "Alternate Weeks (≥2)" Schedules
fix: Subject Dropdown Is Not Displaying All Subjects in campaign
fix: Once loaded, only campaign Grades and Subjects are shown.
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces several enhancements and fixes to the campaign management system, including dynamic grade filtering based on selected schools, optimized audience summary retrieval via a database RPC, and support for campaign frequency. It also restricts program creation to authorized roles, adds web session refresh capabilities using stored tokens, and refines routing permissions for field coordinators. Feedback on these changes highlights a potential crash in SubjectSelection.tsx due to missing optional chaining on roles, a SQL injection risk in SqliteApi.ts from direct string interpolation, and a performance concern in SupabaseApi.ts where database-level pagination is bypassed for percentage filtering. Additionally, throwing a generic error in SupabaseAuth.ts discards crucial error properties, and the custom click-away logic in CampaignMultiSelect.tsx is redundant and can be simplified.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| if (response?.error) { | ||
| throw new Error( | ||
| 'Web session refresh failed: ' + response.error.message, | ||
| ); | ||
| } |
There was a problem hiding this comment.
Wrapping the original response.error in a generic Error object with a custom prefix message discards the original error type and its specific properties (such as status or code). Since this.isInvalidRefreshTokenError(error) likely inspects these properties to identify stale or invalid refresh tokens, throwing a plain Error will prevent it from correctly identifying the error. Throwing response.error directly preserves the full error context.
if (response?.error) {
throw response.error;
}| const hasSubjectModificationRole = SUBJECT_MODIFICATION_ROLES.some((role) => | ||
| roles.includes(role), | ||
| ); |
There was a problem hiding this comment.
If roles is undefined or null (e.g., before the authentication state is fully loaded), calling roles.includes(role) will throw a TypeError and potentially crash the page. Using optional chaining (roles?.includes) or a fallback array ensures robust and defensive execution.
| const hasSubjectModificationRole = SUBJECT_MODIFICATION_ROLES.some((role) => | |
| roles.includes(role), | |
| ); | |
| const hasSubjectModificationRole = SUBJECT_MODIFICATION_ROLES.some((role) => | |
| roles?.includes(role), | |
| ); |
| LEFT JOIN ${TABLES.Assignment_user} au | ||
| ON a.id = au.assignment_id | ||
| AND au.user_id = "${studentId}" | ||
| AND au.is_deleted = 0 |
There was a problem hiding this comment.
Directly interpolating studentId into the SQL query string poses a SQL injection risk. Even though this query runs on client-side SQLite, it is highly recommended to use parameterized queries with placeholders (e.g., ?) to ensure security, robust escaping, and query plan caching.
| const rootRef = useRef<HTMLDivElement | null>(null); | ||
| const handleClickAway = (event: MouseEvent | TouchEvent) => { | ||
| const target = event.target as Node | null; | ||
| if (target && rootRef.current?.contains(target)) { | ||
| return; | ||
| } | ||
|
|
||
| setOpen(false); | ||
| }; | ||
|
|
||
| return ( | ||
| <Autocomplete | ||
| multiple | ||
| disableCloseOnSelect | ||
| disablePortal | ||
| openOnFocus | ||
| options={options} | ||
| value={value} | ||
| loading={loading} | ||
| getOptionLabel={getOptionLabel} | ||
| isOptionEqualToValue={isOptionEqualToValue} | ||
| slotProps={{ | ||
| popper: { | ||
| placement: 'bottom-start', | ||
| modifiers: [ | ||
| { | ||
| name: 'flip', | ||
| enabled: false, | ||
| <ClickAwayListener onClickAway={handleClickAway}> | ||
| <div ref={rootRef} className="campaign-multi-select-root"> |
There was a problem hiding this comment.
The ClickAwayListener component from @mui/material already handles detecting clicks outside of its child element. Therefore, maintaining a custom rootRef and manually checking rootRef.current?.contains(target) inside handleClickAway is redundant. We can simplify this by removing the ref and the custom handler entirely, and passing an inline arrow function directly to onClickAway.
return (
<ClickAwayListener onClickAway={() => setOpen(false)}>
<div className="campaign-multi-select-root">
| const pagedQuery = requiresCalculatedPercentageFiltering | ||
| ? orderedQuery | ||
| : orderedQuery.range(from, to); |
There was a problem hiding this comment.
When requiresCalculatedPercentageFiltering is active, database-level pagination is bypassed, and the entire dataset is fetched into memory to perform filtering and slicing. As the number of programs grows, this can lead to severe performance degradation, high memory usage, and increased network bandwidth. Consider calculating these percentage fields on the database side (e.g., via a database view or generated columns) so that PostgREST can filter them directly, preserving database-level pagination.
![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: 6d4a1d0851
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 availableGrades = useMemo(() => { | ||
| const relevantGradeIds = new Set( | ||
| subjects.flatMap((subject) => subject.gradeIds), | ||
| ); | ||
|
|
||
| return grades.filter((grade) => relevantGradeIds.has(grade.id)); |
There was a problem hiding this comment.
Against the checked-in get_campaign_assignments contract in src/services/database.ts (lines 7964-7967), each unique_subjects entry contains only subject_id and subject_name; SupabaseApi consequently defaults every missing grade_ids field to []. This makes relevantGradeIds empty here and removes every option from the Grade filter for real RPC responses, even though assignments have grades. The RPC contract or the client derivation needs to supply the subject-to-grade mapping before relying on it.
Useful? React with 👍 / 👎.
| const pagedQuery = requiresCalculatedPercentageFiltering | ||
| ? orderedQuery | ||
| : orderedQuery.range(from, to); | ||
| const { data, error, count } = await pagedQuery; |
There was a problem hiding this comment.
When any calculated-percentage filter is active, this branch removes .range() and assumes the response contains the complete program_metrics relation before filtering and slicing it in memory. PostgREST still applies its configured maximum response-row limit to an unbounded select, so once the table exceeds that limit, matching programs beyond the returned prefix disappear and both total and later pages are incorrect. Fetch all batches explicitly or move the calculated filtering into an RPC/database expression.
Useful? React with 👍 / 👎.
| const shouldUseDatabasePagination = Boolean( | ||
| nativeSortColumn || shouldUseRelationSort, | ||
| ); | ||
| const shouldUseDatabasePagination = Boolean(nativeSortColumn); |
There was a problem hiding this comment.
For manager and programName sorts, nativeSortColumn is undefined, so this change disables database pagination and later sorts only the unbounded query response in memory. If the campaign count exceeds the PostgREST response-row limit, campaigns beyond that limit are omitted, totalCount reports only the truncated response length, and users cannot reach them. Relation sorting needs a database/RPC ordering strategy or explicit batched retrieval before client-side sorting.
Useful? React with 👍 / 👎.
10 participants
No description provided.