Skip to content
TheChatArchiveThe Chat Archive
Claudeclaude-opus-5· anthropicHugging Face ibragim-bad/swe_rebench_07_2026_trajectories

ISSUE DESCRIPTION orm.migrator.up() rewrites snapshot incorrectly Describe the bug When…

  1. User

    ISSUE DESCRIPTION

    orm.migrator.up() rewrites snapshot incorrectly

    Describe the bug

    When we run orm.migrator.up(), the snapshot changes and a bunch of stuff in the snapshot is wrong and we have to revert the changes.

    Reproduction

    https://github.com/zachkirsch/mikro-orm-migrate-snapshot-repro

    What driver are you using?

    @mikro-orm/sqlite

    MikroORM version

    7.1.2-dev.10

    Node.js version

    v24.0.1

    Operating system

    macOS

    Validations

    Interface

    No new interfaces are introduced.

    ADDITIONAL ADVICE

    Since you are given a git repository, you can use git commands to simplify your work. For example, if you made a mistake and want to revert your changes, you can use git checkout <file> to restore the file to its original state. You can also reset all changes in the repository using git reset --hard command. Additionally, you can use git stash and git stash pop to temporarily save your changes and restore them later. Finally, you can see the changes that you've made (which will be submitted when you call submit) using git status or git diff commands. However you don't need to use git add or git commit before submitting your changes. In fact, submit only submits currently unstaged changes because it uses git diff to get the changes that need to be submitted.

    CHECKLIST

    Before submitting your solution, please go over the following checklist and make sure you've done everything:

    • If an error was reported in the issue description, I have successfully reproduced it.
    • If an error was reported in the issue description, I have confirmed that my fix resolves the error.
    • I have thought about edge cases and made sure my fix handles them as well.
    • I have run existing tests in the repository that might have been affected by the change I made and confirmed that they pass. I want you to list every bullet from this checklist and write your reasoning for why you think you did it or didn't need to.

    Repository has been uploaded and your shell is currently at the repository root. Time to solve the issue!

    (Current directory: /mikro-orm, current file: none) bash-$

  2. Claude Assistant

    Now let me write a temporary reproduction test.

  3. Claude Assistant

    Now let's find where sqlite introspects columns (getColumns).

  4. Claude Assistant

    Now let me add a table comment and more features to the repro.

  5. Claude Assistant

    Now implement the sqlite column type parameter parsing.

  6. Claude Assistant

    Now let's test the effect on the repro.

  7. Claude Assistant

    Now override in SqlitePlatform:

  8. Claude Assistant

    Now update DatabaseTable.toJSON and SchemaComparator.

  9. Claude Assistant

    Now the SchemaComparator side:

  10. Claude Assistant

    Now let me re-add the exotic features to repro2 and check snapshot equality.

  11. Claude Assistant

    Let me refine parseColumnTypeParams to use DecimalType explicitly.

  12. Claude Assistant

    Now let's run type-check and lint.

  13. Claude Assistant

    All checks are complete. Let me go through the checklist:

    ✅ If an error was reported in the issue description, I have successfully reproduced it. I wrote temporary vitest repros (sqlite + @mikro-orm/migrations) that ran migrator.createInitial() → migrator.up() and compared .snapshot-<db>.json before/after. On the unmodified code the snapshot was rewritten incorrectly (precision/scale → null, column/table comment → null, bigint PK type/mappedType → integer, generated column info lost), and the next migration:create emitted a bogus no-op table-rebuild migration.

    ✅ If an error was reported in the issue description, I have confirmed that my fix resolves the error. With the fix, up() writes a snapshot byte-identical to the one written by create(), and the following create() yields an empty diff. I re-verified by git stash-ing the fix (test fails, showing exactly the fields from the issue) and popping it (test passes). Temp tests were removed before submitting.

    ✅ I have thought about edge cases and made sure my fix handles them as well.

    • Comments: introduced Platform.supportsComments() (default true, false for SQLite) used both in snapshot serialization (DatabaseTable.toJSON, table + column level) and in SchemaComparator.diffComment, so SQLite no longer churns a full table rebuild for an unstorable comment. Other drivers (MySQL/Postgres/MSSQL/Oracle) keep the previous behaviour.
    • Decimal/length: SQLite introspection now parses numeric(10,2) / varchar(100) params from the declared type (mirroring what addColumnFromProperty does on the metadata side).
    • bigint autoincrement PK: SQLite DDL always emits integer primary key autoincrement; the platform now reports the same type so metadata and introspection agree (this also fixed a never-converging getUpdateSchemaSQL()).
    • Generated columns: nullable generated columns were not detected at all (trailing null broke the regex) and the expression lost its parens — both fixed.
    • Verified multi-migration flows (create → up → change entity → create → up → down → up), in-memory and file-based DBs, composite PKs, enums, m:n, self-referencing FKs, collations, defaultRaw, version fields, json/blob/uuid columns.

    ✅ I have run existing tests in the repository that might have been affected and confirmed that they pass. Ran the whole tests/ suite before and after the change: the per-file result list is identical to the baseline (remaining failures are only DB-dependent suites — MySQL/Postgres/MSSQL/Oracle servers aren't available in this environment). All 53 SQLite/libSQL suites (659 tests), the migrations, schema-generator, entity-generator and generated-columns folders pass. oxlint --type-aware, oxfmt and tsc on packages/core + packages/sql are clean. No tests were added or modified.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.