Comparing changes

…s directly from the library (#1159) * chore: call non-hermetic functions since we're installing node_modules directly from the library Source-Link: googleapis/synthtool@4c4063f Post-Processor: gcr.io/cloud-devrel-public-resources/owlbot-nodejs:latest@sha256:8b6a07a38d1583d96b6e251ba208bd4ef0bc2a0cc37471ffc518841651d15bd6 Co-authored-by: Owl Bot <gcf-owl-bot[bot]@users.noreply.github.com>

* A simple test with multidb support * A simple test with multidb support Adds databaseId to the request options, client options and code for carrying over the client options to the request options as well as a test to make sure it actually works. * Add additional checks with get The test should include additional checks to ensure get is running the way it is supposed to. * Add a test for saving with another database * other datastore should delete the entity Since we are saving things to another datastore we need to delete the data from that datastore. * Add a test for the getDatastore method The user needs a way to get the database Id from the client so that they can keep track of their clients more easily. * Add a unit test for database id Ensure that the database id gets passed all the way down to the generated layer and gets included in the request. * Do an additional check for the sake of tests We should check for the existance of datastore.options so that the code doesn’t fail if they are not provided at all. * initialize generated client so it can be mocked The test on kokoro is failing because in that environment a project id isn’t defined. The solution is to eliminate the get request so that a call never reaches the backend and initialize the gapic client so it can be mocked out and then catch any call that reaches the mock. * comments and setDatabaseId Add docs comments for newly added functions and add a test that ensures setDatabaseId works correctly. * Refactor of database name Don’t use the same string multiple times. When the database name we are using for testing changes then we should only have to make that change in one place. * correction to test against old database This was a typo. We want to test against the database we were using before * getRequestWithDatabaseId function Add the getRequestWithDatabaseId function and change it so that it can easily be used in multiple places and the tests still pass. * Change function name to addDatabaseIdToRequest The request options should contain databaseId so rename the function as appropriate. * Comments in all the places to add reqOpts This commit contains a comment in all the places we want to add the reqOpts change. * Add a warning to a certain test for easier debug Add a warning because the way the assert checking was done before, the test suite was not receiving any indication of which test was failing which makes debugging really hard. * This fixed the bug for the save operation * datastore request id Add request ids back in all the places they are required * addDatabaseIdToRequest Add addDatabaseIdToRequest to the transaction test framework so that it can be used in the tests. * Solve all the test errors for createReadStream Mock out add database id for the tests in createReadStream that correspond to createReadStream. * addDatabaseIdToRequest mock for runQueryStream Add the addDatabaseIdToRequest mock to the tests for every runQueryStream test so that the tests pass. * By default, mock out addDatabaseIdToRequest addDatabaseIdToRequest should be a default function in datastore object of request so that we don’t have to mock it out in many different places. * Revert "addDatabaseIdToRequest mock for runQueryStream" This reverts commit a3bdb92. * Revert "Solve all the test errors for createReadStream" This reverts commit 7b5f902. * linting lint an import * Revert "datastore request id" This reverts commit f167ce3. * Revert "This fixed the bug for the save operation" This reverts commit dd7b06c. * Revert "Comments in all the places to add reqOpts" This reverts commit 8ba512a. * Introduce addDatabaseIdRequest in request Introduce adding the databaseId in request layer again. * Declare a constant so that tests don’t run Importing test into the system-test folder causes the system-test command to also run the tests. Eliminate this import. * Remove setDatabaseId Remove setDatabaseId as it provides a way for the user to get confused about the current state of the client. * Add comment for addDatabaseIdToRequest function The function needs a comment to be consistent with docs. * Address header issue The new util file needs a header * Add databaseId for encoding and decoding keys The databaseId should be encoded on keys so that the user can write on a record by record basis * Revert "Add databaseId for encoding and decoding keys" This reverts commit 0ebabd5. * Try a test with the namespace Try out a new test that looks at data details * Use the other datastore in the test Use another datastore with key * Create a test that looks at specific details The test should look at specific details in the key and should return the data with the right author. * Add additional checks to existing test Add additional checks to existing test to show default database is unaffected. * Revert "Revert "Add databaseId for encoding and decoding keys"" This reverts commit 6449d10. * Revert "Revert "Revert "Add databaseId for encoding and decoding keys""" This reverts commit e694341. * Rename the second database to secondDatabase * Use the database called multidb-test multidb-test is the name of the database set up on the project used in Github automation so we need to change this name in code to use that new database. * Eliminate function on index that adds the id Currently a function is used as a passthrough to the function on the util file. This unnecessarily exposes a function to the user that they shouldn’t be concerned with. * Fix the test so that the assertion error bubbles The assertion error should bubble up to the user if we send it through the callback. * Add types so that purpose of function is clear Types currently set to any should be more specific. This will help clarify the purpose of the function that is used. * Add types and names so that the function works The types should be more specific and the names should be more explicit and clear. * Eliminate repeated code fragments The test case has the same lines for each secondary database key and these lines should not be repeated so we regroup them into better data structures for the job * Better names and code organization The names of some variables were changed to align with others and code was moved so that code pertaining to the same function is grouped together * Change the variable name of the default key There is only one default key so we change the variable name so that it doesn’t mention 1 * Refactor the post key hierarchy out Refactor the hierarchy out for a cleaner test * Add only to the test we are interested in * Revert "Add only to the test we are interested in" This reverts commit 0378dea. * Add second database id In the transaction tests we add the databaseId parameter. The databaseId parameter will run tests against the secondary database. * Remove unnecessary mocks The mocks for addDatabaseIdToRequest are not necessary for these two files since they were only useful for the way the code was written before so should be removed. * Remove unused imports Imports that were used in old mocks no longer exist in the file so should be removed. * Add parameterized testing to the test/index file Parameterized testing can be added to the test/index file to test to see how the tests behave with databaseId set to the second database and without this setting too. * Add parameterized testing to test/transaction.ts Parameterized testing should be used on transaction.ts. This commit adds the testing with async. * This ensures that the tests still run on the deft Tests need to run on the default database as well as the secondary database. This commit ensures that they do. * Add parameterized testing for the system tests Parameterized testing for the system tests will ensure that tests run with the default database as well as with the secondary database. * Use default datastore variable in new tests The intent of the datastore variable in the new tests is to be the default database so that we are able to see if reads/writes affect the default or secondary database in these tests. They should not use the parameterized version of datastore. * inline add database id Now that the tests don’t need to mock out adding database id function, we can inline the code that adds the database id to the request options. * Modify the namespace in parameterized tests Parameterized tests should have a different namespace for the second set of tests for sure so we add this prefix. * Move databaseId to SharedQueryOptions for alignme The databaseId should be moved to shared query options as discussed. * Run the linter Linting was broken so that the PR could be read. Ran the linter again.

…unning on docker (#1164) * Check if DATASTORE_EMULATOR_HOST is set Instead of checking to see if the endpoint is local, we should check to see if the datastore emulator variable is set. This is because there is an edge case where the user is using an emulator, but the emulator is not listening from local host and in this case we want to use insecure credentials. * Add tests for combinations of emulator / custom Add unit tests for evaluating value of ssl credentials. Tests evaluate what happens when the user uses remote custom endpoints or local custom endpoints in combination with different values for the DATASTORE_EMULATOR_HOST variable. * Revert "Check if DATASTORE_EMULATOR_HOST is set" This reverts commit dea1706. * Revert "Revert "Check if DATASTORE_EMULATOR_HOST is set"" This reverts commit a3b50df. * Clean up the tests The tests should use a concise apiEndpoint variable and they should set the host for the remote case to be the same as the apiEndpoint variable. * Add more test cases describing all behaviour More test cases are needed to describe what happens when the DATASTORE_EMULATOR_HOST variable is not set both in a remote and in a localhost environment because those behaviors should be different. Setting the apiEndpoint to localhost means the user wants to use the emulator without specifying the environment variable so insecure credentials should be provided. Setting the apiEndpoint to remote means the user likely wishes to use regional endpoints so should not be using insecure credentials to skip authentication. * Allows test cases to pass from previous commit If the user uses localhost as the endpoint or they provide DATASTORE_EMULATOR_HOST then it is assumed that the user is using the emulator and authentication is skipped. Otherwise, authentication is not skipped. * Change old test title back to the way it was The old test title should be what it was before. setHost is now only needed in one place. * Add comments to each test The comments need to be added to explain why the test is written the way that it is written so that we know the motivation behind each test.

…1114) * Add a test for testing without mocks Testing without mocks allows us to look at the code how it is functioning right now so that we can evaluate the intended behavior. * Do a refactor on the test Objects shared by both tests should be moved into the outer block so that they can be shared. * improve the test so that the error bubbles up Now the test fails properly and passes properly by introducing callbacks * isolate the latter test right now The failure should be visible to the user and in the current state it is so now by isolating the test it becomes easier to explore solutions that work. * test fixes indent a describe block, modify tests to make sure they pass and refactor a function out that easily allows for testing modified parameters. * Add two more tests for excludeFromIndexes on array Add two more tests that address the array case for excludeIndexesFromPath. Also ensure that the conditions which check for these test cases properly ignore exclude from indexes when an array is used * Move the tests over to the index file We should not create a nomocks file because we can just group the tests under a unique describe block of test/index.ts. * PR update Remove typeof and remove check for string ‘undefined’ * Create a function to refactor the tests Refactor the tests so that they are more readable. There is a lot of repeated code in each test which makes comparisons hard. * All tests for this feature now use same script Replace all tests exercising the feature so that they all use the same test script. This will make the tests way easier to read. * Add parameterized testing Add parameterized testing to eliminate repeated code. These tests need to be easier to read for the sake of the reviewer. * Eliminate function and inline code This function is only used once now. We should inline this code in the test. * Add JSON stringify to test title In the test title include the properties passed in so that it is easy to track what passed/failed. * Organize code so that similar variables are used Similar variables should be used together. Organize the code so that there is less confusion about which variables are related. * inline getExpectedConfig function The getExpectedConfig does not need to be used anymore because its code can just be inlined in the one place it is used. * Separate tests into blocks and inline code Separate the test into blocks and inline some parameters for better test readability. * Add types to parameters passed into async Stronger typing makes the code easier to read. * Eliminate unnecessary variable assignment Don’t assign to the Datastore variable. This is not necessary. Just use the OriginalDatastore variable directly. * Ran linter and simplified source changes The diff will be easier to read if we do not add a nested if statement and instead apply a condition on the if and change the else to else if. * Add blank line back for easier diff reading * Try again, eliminate the blank line * Additional adjustment to entity first path part Apply the first suggestion in the PR to all examples * Define isFirstPathPartUndefined There is a repeated code fragment for checking if the first path part is undefined that we should refactor. * Rename the variable to align with what it does * run linter * Revert "run linter" This reverts commit a101f27. * Revert "Rename the variable to align with what it does" This reverts commit 4f54d94. * Revert "Define isFirstPathPartUndefined" This reverts commit df3f376. * Refactor check out for seeing if defined We do four checks to see if the first path part is defined. We should refactor these checks out. * Move comment to more appropriate place The comment should reflect the next line of code. * Replace comments with description Replace comments with description property so that the tests will report the description when running the tests instead of the properties in the description. * Eliminate prefix. Only use description * fix typo * lowercase convention

* Extend the sleep time to allow save to work The sleep time should be increased to allow the save time to vary significantly from the variable savedTime so that any time different between the client and the server’s clock is negligible compared to the sleep time. * test: Extend the sleep time to allow save to work * run linter * Reapply sleep commit that was there first

Co-authored-by: release-please[bot] <55107282+release-please[bot]@users.noreply.github.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Comparing changes

Open a pull request

Commits on Sep 21, 2023

Commits on Sep 26, 2023

Commits on Sep 27, 2023

Commits on Sep 29, 2023

Commits on Oct 2, 2023

This comparison is taking too long to generate.

Uh oh!