Fixes:
The streaming protocol is now tightened to ensure that only the last message for a result set contain isCompleted=true. Now isCompleted=true is never sent in isAvailable message.
Tightened logic in sending messages to make sure that no duplicate messages get sent out due to concurrent processing.
Made a fix to a null reference exception when processing special action which was a pre-existing benign bug.
Testing: Added 1 more new test that runs existing tests concurrently 1000 times to make sure no random timing issues are observed and tightened verifications to existing tests to ensure no duplicate messages and only one isComplete=true message is sent across.
* tightening the protocal to ensure only one message (the last one) with completed=true is sent back for a single result set within a query batch
This changes adds the following two notifications from the results processing within a batch. These new notifications allows a consumer to stream results from a resultset instead of getting them all at once after the entire resultset has been fetched.
ResultsAvailable
This is issued after at least 1 row has been fetched for this resultset.
ResultsUpdated
This is issued periodically as more rows are available on this resultset. The final send of this notification when all rows have been fetched has the property 'Complete' set to true in the ResultSummary object.
Detailed Change Log:
* Initial completed implementation of QueryResults stream feature. 3 unittests still need fixing
* Fix for the 3 failing test. I will look into making MockBehavior strict again for the three tests later
* Making GetReader/GetWriter use filestream objects in FileShare.ReadWrite mode so the file can be concurrently read and written
* Changing resultsAvailable also to fire off on a timer instead of after 1st row
* adding a project for clr TableValuedFunction to produce result set with delays after each row. This is helpful in end to end testing.
* Fixing up some tests and simplifying implementation of result update timer
* Address review comments
* Some test fixes
* Disabled flaky test verification
* Reopen connections prior to creating query execution data readers
* Reopen connection if an exception was rasied executing the query
* Fix unit tests
* Support `GO N` syntax to execute multiple times
- Plumbed through the batch execution count from the parser and used in the batch execution code path
- Functionality matches SSMS:
- Outputs loop start/end messages that match SSMS if you're doing multi-batch execution
- Outputs an "ignoring failure" error if an error happens during a batch
- Added tests for this
- Manually verified end to end also
* Fixing test error
* Fix to make sql exceptions surface properly to user (with important notes!)
* Adding support for detecting column size issues when updating a cell
* Adding unit tests for the exception on read scenario
* change simple execute to open a new connection and close it every query
* updated tests for simple execute
* removed an unnecessary connect
* refactored code to be more readable
* global try catch on simple execute
* added multiple execution test
* update execution to be asynchrous; update tests to account for asynchrounous nature
* inital request
* refactored query execution failure callback to take exception
* added failure callback to execute and return
* added test for query execute and return
* updated params
* removed dead code
* addressed feedback; added multiple active result set support; updated tests
* addessed feedback and added testing and errors and verification
* change <= to ==
* changed name of trashQ to removedQuery
This change ensures that when calling `requestContext.SendError` you are only able to supply parameters that match the language service beta protocol expected Error object. In other words, you have to provide an error message and optionally and error code.
# **BREAKING API CHANGES**
This will break displaying errors in Microsoft/vscode-mssql. I will be making changes to properly handle the error object shortly.
* Adding contract for returning Error objects as per LanguageService "protocol"
* Fixes throughout codebase to send only error message in error cases
Cleanup of CredentialServiceTest unit test class
Adding standard error handling for event flow validator
* Adding optional data field as per protocol spec
https://github.com/Microsoft/language-server-protocol/blob/master/protocol.md
* Adding optional validation for error objects
The main goal of this feature is to enable a command that will
1) Generate a parameterized command for each edit that is in the session
2) Execute that command against the server
3) Update the cached results of the table/view that's being edited with the committed changes (including computed/identity columns)
There's some secret sauce in here where I cheated around worrying about gaps in the updated results. This was accomplished by implementing an IComparable for row edit objects that ensures deletes are the *last* actions to occur and that they occur from the bottom of the list up (highest row ID to lowest). Thus, all other actions that are dependent on the row ID are performed first, then the largest row ID is deleted, then next largest, etc. Nevertheless, by the end of a commit the associated ResultSet is still the source of truth. It is expected that the results grid will need updating once changes are committed.
Also worth noting, although this pull request supports a "many edits, one commit" approach, it will work just fine for a "one edit, one commit" approach.
* WIP
* Adding basic commit support. Deletions work!
* Nailing down the commit logic, insert commits work!
* Updates work!
* Fixing bug in DbColumnWrapper IsReadOnly setting
* Comments
* ResultSet unit tests, fixing issue with seeking in mock writers
* Unit tests for RowCreate commands
* Unit tests for RowDelete
* RowUpdate unit tests
* Session and edit base tests
* Fixing broken unit tests
* Moving constants to constants file
* Addressing code review feedback
* Fixes from merge issues, string consts
* Removing ad-hoc code
* fixing as per @abist requests
* Fixing a couple more issues
The goal of this make sure that test code is correctly organized to ensure that test suites aren't dependent on each other.
* UnitTests get their own project now (renaming Microsoft.SqlTools.ServiceLayer.Test to Microsoft.SqlTools.ServiceLayer.UnitTests) which is about 90% of the changes to the files.
* IntegrationTests no longer depends on UnitTests, only Test.Common
* Any shared components from TestObjects that spins up a "live" connection has been moved to IntegrationTests Utility/LiveConnectionHelper.cs
* The dictionary-based mock file stream factory has been moved to Test.Common since it is used by UnitTests and IntegrationTests
* Added a overload that doesn't take a dictionary for when we don't care about monitoring the storage (about 90% of the time)
* The RunIf* wrapper methods have been moved to Test.Common
* OwnerUri and StandardQuery constants have been moved to Test.Common Constants file
* Updating to latest SDK version available at https://www.microsoft.com/net/core#windowscmd
* Moving unit tests to unit test folder
* Changing namespaces to UnitTests
* Moving some constants and shared functionality into common project, making the UnitTests reference it
* Unit tests are working!
* Integration tests are working
* Updating automated test runs
* Fixing one last broken unit test
* Exposing internals for other projects
* Moving edit data tests to UnitTest project
* Applying refactor fixes to unit tests
* Fixing flaky test that wasn't awaiting completion
