test: 'fix' memory leak in graceful-fs #211
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
graceful-fs and jest does not like each other, so this removes the graceful-fs implementation before running tests and adds it back afterwards. jestjs/jest#6814
|
wow, ugly. |
|
Yeah, ugly. Hopefully this can be removed soon, but currently CI builds started to fail randomly because I had too much test cases. And it reached 1g heap size, without this hack... Its sad that but I will rather do this than manually rerun failed builds and hope it will fit the limit. Also adding more test files would make this more and more often... |
|
Ha! I guess it is possible to do too much unit testing... =) Yea... shrug... just a test hack, no biggie. |
seyfer
reviewed
Nov 13, 2019
| @@ -36,8 +36,10 @@ | |||
| "homepage": "https://mikro-orm.io", | |||
| "scripts": { | |||
| "build": "rimraf dist && tsc", | |||
| "test": "jest --runInBand", | |||
| "coverage": "rimraf temp tests/generated-entities && jest --runInBand --coverage", | |||
| "pretest": "node tests/pre-test", | |||
There was a problem hiding this comment.
is any performance test before and after available? I have added this to my project - see no difference in heap size usage
There was a problem hiding this comment.
travis builds were randomly reaching 1g memory limit, with this change the max heap size is quite stable around 300m. you can take a look at the travis builds before and after:
before: https://travis-ci.com/mikro-orm/mikro-orm/jobs/248182459
after: https://travis-ci.com/mikro-orm/mikro-orm/jobs/248297817
unfortunately i do not have the logging of heap size there before this actual change.
There was a problem hiding this comment.
maybe you have actual leak in your app, or it is caused by another dependency than graceful-fs...
B4nan
added a commit
that referenced
this pull request
Nov 20, 2019
* fix(core): disable auto flushing by default [BC] (#79) * feat(core): use knex to generate sql + enable connection pooling [BC] (#76) * fix(mapping): remove deprecated `fk` option from 1:m and m:1 decorators (#87) * feat(mapping): allow overriding getClassName() in NamingStrategy (#88) * chore: remove @Class jsdoc from tests as it does not work in WS anymore * chore: use ObjectId instead of deprecated ObjectID in mongo * feat(metadata): create instance of metadata instead of static one [BC] (#91) * feat(hooks): add onInit hook fired after entity is created (#92) * feat(core): add support for virtual property getters (#93) * feat(sql): support multiple conditions in JOINs (#94) * feat(schema): use knex in schema generator (#81) * feat(schema): add basic entity generator (#98) * chore: update roadmap * fix(mapping): do not override user defined nullable value in m:1 and 1:1 * feat(schema): add basic schema update (#97) * feat(cli): add basic CLI tool (#102) * feat(core): improve logging - add namespaces, colors and highlighting (#109) * refactor: improve handling of unknown types in entity/schema generators * feat(mongo): improve query logging, use `inspect` instead of `stringify` * feat(metadata): improve validation during metadata discovery * fix(schema): prefer user-defined collection names in naming strategy * feat(core): add support for read connections (#116) * feat(core): add `Reference<T>` wrapper to allow improved type safety (#117) * feat(drivers): add support for MariaDB (#120) * docs: update docs for v3 * fix(mapping): remove obsolete parameter in UnderscoreNamingStrategy (#134) * docs: fix typo in reference wrapper section (#137) * docs: comment out comments (#141) * refactor(core): propagate reference PK from entity instead of its clone * feat(core): add findOneOrFail method to entity manager and repository (#142) * refactor(mapping): allow passing Reference to Reference.create() * feat(mapping): add NoopNamingStrategy * refactor(cli): make MIKRO_ORM_CLI env variable highest priority * refactor(validation): validate that entitiesDirs are not file globs * feat(cli): add debug command to help with setting up the CLI * docs: merge entity reference and reference wrapper sections * chore: fix build * docs: fix link to mysql integration tests (#147) * refactor: rename NoopNamingStrategy to EntityCaseNamingStrategy * refactor: allow replication in postgres * refactor: add dependency versions to debug cli command output * refactor: change sqlite dependency to sqlite3 as that is now used (knex) * refactor: change sqlite dependency to sqlite3 in docs * refactor: use globby to check for paths in cli debug command * refactor: allow default export in cli-config * docs: fix schema generator docs (get query methods vs run query methods) * fix(core): always query inverse side of 1:1 association * feat(validation): warn when failing to get metadata of an entity * feat(logging): allow logging full query including params (#155) * refactor: return entity from IEntity.assign() * refactor: improve auto-detection of ts-node * chore: change schema in default postgres connection string (#154) * chore: fix links to repository and homepage after org transfer * feat(mapping): add type-safe way to define relationships * feat(postgres): use timestamps with time zone by default * refactor: use relative paths in metadata cache * feat(core): allow populating all relations via `populate: true` * feat(validation): validate one to one relationship metadata * feat(mapping): auto-wire missing references from owner to inverse side * refactor(core): make em.find() where parameter required * feat(core): add `em.findAndCount()` method * docs: add `findAndCount` method and describe query object in EM section * feat(core): allow empty where condition in `em.count()` * feat(core): allow filtering and sorting by nested query * docs: declare correct type of codeblock (#170) * feat(core): add support for eager loading * docs: declare correct type of codeblock (#173) * feat(cli): rename default config name to mikro-orm.config.js * feat(core): allow whitelisting entity fields in em.find() * fix(core): allow object constructor parameters in entities * feat(core): allow assigning PK to undefined/null * fix(core): do not use request context in transactional/user forks * fix(core): auto-wire 1:1 owner to inverse side * feat(core): add support for deep nested conditions with operators (#185) * fix(cli): rename `--no-fk` schema command parameter to `--fk-checks` * fix(core): ignore inverse side of 1:1 when computing change set * feat(drivers): add native UUID postgres type (#188) * chore(types): improve types in EntityGenerator (#192) * feat(metadata): auto-detect optional properties * docs: fix bash highlighting * fix(generator): fixed default values and types for nullable properties (#191) * feat(core): add @repository decorator * feat(core): simplify entity definition and rework typings of FilterQuery (#193) * docs: add basic upgrading guide (WIP) * fix(query-builder): fix malformed query when populate and join are used * chore: export @SerializedPrimaryKey decorator, omit some internal types * feat(core): propagate nested where and orderBy when populating * refactor: use Set for QB flags * refactor: em `transactional` method return value & improve types (#197) * refactor: em and connection generic types (#198) * chore: remove unused import * refactor: add support for self referencing when auto-joining * fix(core): always init collections when creating entity via em.create() * tests: fix running tests on windows (#199) * refactor(typings): support nested collections in FilterQuery * docs: improve entity definition section, add more examples * docs: add deployment sections * refactor: add Reference.get(prop) method * feat(query-builder): allow mapping to entities directly via getResult() * feat(core): add support for filtering and ordering of Collection items * feat(core): add support for bundling with Webpack (#200) * chore: fix code style and increase timeout for CI tests * docs: add section about bundling with webpack (#202) * refactor: enable strictPropertyInitialization * docs: improve code style of webpack section examples * chore: hide knex's warning about sqlite FK failures * refactor: try inlining deep conditions before renaming * refactor: simplify webpack support test * feat(drivers): allow passing additional driver options * feat(core): use composite PK in many to many relations (#204) * fix(serializing): add check for circular references in toObject() * feat(core): do not require entity attribute in collection decorators (#207) * feat(core): add support for migrations via umzug (#209) * test: 'fix' memory leak in graceful-fs (#211) * refactor: implement allOrNothing in Migrator * docs: add note about possible TS inference problems with FilterQuery * docs: add section about unit of work to readme * refactor: run on init hooks after hydration * fix(core): make sure constructor params are sniffer from the constructor * fix(core): allow persisting 1:1 from inverse side * test: run tests in ci on node 8, 10, 12 and 13 * test: speed up MikroORM.init validation tests * chore: remove lgtm alert in MigrationGenerator * docs: update deployment section (#223) * refactor: do not use some umzug types that are badly defined * feat(schema): allow dropping migrations table via `schema:drop` cli cmd * fix(schema): do not make FK fields nullable if not needed * docs: add section with limitations of migrations * chore: fix build, remove stub of new docs * fix(serializing): do not ignore already visited collection items * feat(cli): add cache:generate command to warm up production cache * feat(core): add support for enums via `@Enum()` decorator (#232) * docs: add not about enums to deployment section * docs: use `!` and `?` in entity props definition * refactor: remove `AUTO_GROUP_BY` flag from query builder * fix(core): fix querying by m:n primary keys * refactor(core): do not propagate flat order by * chore(dependencies): downgrade ts-morph to v4 as v5 requires TS 3.7 * perf(sql): use multi-insert when populating m:n collections * fix(sql): support self-referencing m:n in pivot tables * docs: add link to realworld example app * docs: add link to migrations section into sidebar * chore: fix typo and remove unreachable getReference() signature * feat(schema): add support for create/drop database (#237) * feat(cli): add `database:import` command to run external sql dumps * feat(sqlite): ensure the directory with database exists * chore: update mongodb to 3.3 and use new unified topology (#241) * feat(metadata): add `ReflectMetadataProvider`, rename the `ts-morph` one (#240) * refactor: fix create/drop db commands in mariadb driver * chore: move typescript to dev dependencies * chore: improve upgrading docs * chore: add note about v2 docs * chore: add script to generate the changelog * chore: update changelog
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
graceful-fsandjestdoes not like each other, so this removes thegraceful-fsimplementation before running tests and adds it back afterwards.jestjs/jest#6399
jestjs/jest#6814
isaacs/node-graceful-fs#102