Repository navigation
Schema: validation-rule editor + client-side row validation #19
Description
Scope
Expose each field's validation object and pre-validate rows before they hit the API.
| Rule | Applies to | Server error message |
|---|---|---|
min |
number, date, datetime |
<label> must be at least <min> / ... must be after <min> |
max |
number, date, datetime |
<label> must be at most <max> / ... must be before <max> |
minLength |
text, textarea |
<label> must be at least <n> characters |
maxLength |
text, textarea |
<label> must be at most <n> characters |
pattern |
text, textarea, email |
<label> format is invalid |
step |
number |
increment hint, UI only — not enforced server-side |
Type-level checks always run regardless of validation: email must be a valid address, url a valid URL, boolean a real boolean, select/multiselect values must be within options (<label> must be one of: ...), and a required empty value yields <label> is required.
Rules are applied by the server on POST/PUT /api/lists/{id}/data — not when the schema is created.
Acceptance criteria
- Editor exposes every applicable rule per field type (and hides the inapplicable ones).
- Row entry validates client-side first, showing the message beside the offending field.
- A server
400still maps back to the right field rather than a generic toast. -
stepis applied to the number input only, and is not treated as a constraint.
Activity
Adron commentedon Sep 16, 2026
Live-verified findings from the schema service (#17 / PR #143) — three change the design here
propertiesform accepts only SIX types, not twelvepropertyTypeis also mandatory on every item (omitting it gives the same error for'undefined'/'null').Consequence: a list using
textarea,datetime,tel,select,multiselectorprioritycannot be edited non-destructively at all — every change to it has to go through the destructive DSL rebuild. That is a real, permanent constraint on #22's guard, not an edge case: the moment a user picks one of the six richer types in #18's builder, they lose safe editing on that list forever.Exposed as
ListFieldType.PropertiesEditableandListSchema.SupportsPropertiesEditso the UI can tell the user before they choose.2. "Destructive" means columns, not rows — so #22's confirmation copy must not say "rows will be deleted"
Confirmed live. The DSL rebuild:
validation/options/visibilitythe new DSL omits;rowData— still stored, not shown, not validated.And it renames the list:
schema.nameoverwrote the title ('ZZ Throwaway Schema Probe'→'ZZ Renamed By DSL Rebuild'), andschema.descriptionoverwrites the list description too — clearing it when omitted. #22's confirmation needs to say all three things.3. Deleting a data-bearing column requires
?force=true?force=truethen strips that key from every row. The sharedEnsureSuccessAsyncdiscards thepropertiesWithDataarray, so a dedicatedListSchemaExceptioncarriesIssues/PropertiesWithData/RequiresForce. #18 should name the affected columns in its confirmation rather than showing a bare 409.4. The server never validates conditional visibility — #20 must
An unknown operator, a missing watched key, and a forward reference all return
200and then silently never match. So there is no server-side safety net:ListSchema.Validate()does these checks client-side, and #20's editor should surface them rather than letting a rule quietly do nothing.5.
propertiesPUT definitively preserves row data — #22's guard is soundA rename + column-add on a list holding two rows returned
200and both rows came back with identical ids, identicalversion(still 1), and identicalrowData. Column ids were preserved too, and thevalidationRules/helpText/visibilityConditionthat thepropertiesshape cannot even express survived untouched.Note the envelope differs between the two directions:
GETreturns{"data":{…}}, thepropertiesPUTreturns{"properties":[…]}ordered bydisplayOrder— nodatawrapper.6. Schema-less lists are unaffected
Created without a schema they answer
{"data":{"name":"<list title>","description":…,"fields":[]}}and accept rows with arbitrary, even nested, keys. Unknown keys are accepted on a schema'd list too — nothing is stripped.fields: []cannot be PUT back (DSL must have at least one field), soListSchema.Validate()catches that locally.Also relevant to #21
Row-write validation returns
422withdetails:[{field,message}], which the current plumbing flattens — so #21 can't currently attach a message to the offending field without unpacking that itself.Separately, #144/PR #146 fixed a crash on the row-read path (
ListDataRow.ListIdwasrequiredbut absent) and modelledversion, which is the hook for optimistic concurrency on row edits.