Snippet completions from JDTLS not inserting text properly

[valentinegb]

Check for existing issues

• Completed

Describe the bug / provide steps to reproduce it

When attempting to autocomplete a snippet completion with the Zed Java extension (which uses Eclipse JDT Language Server, aka. JDTLS), Zed inserts the name of the snippet instead of the snippet's insertText.

For example, if I autocomplete class in a file called DriveSubsystem.java, Zed just inserts the text class when I would've expected:

/**
 * DriveSubsystem
 */
public class DriveSubsystem {

  ${0}
}

Here are the LSP RPC logs around the time of autocompleting:

// Receive:
{"jsonrpc":"2.0","method":"$/progress","params":{"token":"5c0bc3c5-48af-4bb5-aa79-3d587a7dd3cc","value":{"kind":"report","message":"Publish Diagnostics - 0% ","percentage":0}}}
// Send:
{"jsonrpc":"2.0","id":"22","result":null}
// Receive:
{"jsonrpc":"2.0","method":"textDocument/publishDiagnostics","params":{"uri":"file:///Users/valentinebriese/Developer/SwerveDrive2025/src/main/java/frc/robot/subsystems/DriveSubsystem.java","diagnostics":[{"range":{"start":{"line":1,"character":0},"end":{"line":1,"character":3}},"severity":1,"code":"1610612968","source":"Java","message":"Syntax error on token \"cla\", delete this token"}]}}
// Receive:
{"jsonrpc":"2.0","method":"$/progress","params":{"token":"5c0bc3c5-48af-4bb5-aa79-3d587a7dd3cc","value":{"kind":"end","message":"Publish Diagnostics"}}}
// Send:
{"jsonrpc":"2.0","method":"textDocument/didChange","params":{"textDocument":{"uri":"file:///Users/valentinebriese/Developer/SwerveDrive2025/src/main/java/frc/robot/subsystems/DriveSubsystem.java","version":4},"contentChanges":[{"range":{"start":{"line":1,"character":3},"end":{"line":1,"character":3}},"text":"ss"}]}}
// Send:
{"jsonrpc":"2.0","id":16,"method":"completionItem/resolve","params":{"label":"class","kind":15,"documentation":"/**\n * DriveSubsystem\n */\npublic class DriveSubsystem {\n\n\t\n}","sortText":"999999212","filterText":"class","insertText":"/**\n * DriveSubsystem\n */\npublic class DriveSubsystem {\n\n\t${0}\n}","insertTextFormat":2,"insertTextMode":2,"textEdit":{"newText":"class","insert":{"start":{"line":1,"character":0},"end":{"line":1,"character":3}},"replace":{"start":{"line":1,"character":0},"end":{"line":1,"character":3}}}}}
// Receive:
{"jsonrpc":"2.0","id":16,"result":{"label":"class","kind":15,"documentation":"/**\n * DriveSubsystem\n */\npublic class DriveSubsystem {\n\n\t\n}","sortText":"999999212","filterText":"class","insertText":"/**\n * DriveSubsystem\n */\npublic class DriveSubsystem {\n\n\t${0}\n}","insertTextFormat":2,"insertTextMode":2,"textEdit":{"newText":"class","insert":{"start":{"line":1,"character":0},"end":{"line":1,"character":3}},"replace":{"start":{"line":1,"character":0},"end":{"line":1,"character":3}}}}}
// Receive:
{"jsonrpc":"2.0","id":"23","method":"window/workDoneProgress/create","params":{"token":"a597e761-6ad4-4d01-acb7-6d83b82b2691"}}
// Receive:
{"jsonrpc":"2.0","method":"$/progress","params":{"token":"a597e761-6ad4-4d01-acb7-6d83b82b2691","value":{"kind":"begin","title":"Validate documents","message":"Validate documents"}}}
// Receive:
{"jsonrpc":"2.0","method":"$/progress","params":{"token":"a597e761-6ad4-4d01-acb7-6d83b82b2691","value":{"kind":"report","message":"Validate documents - 0% ","percentage":0}}}
// Send:
{"jsonrpc":"2.0","id":"23","result":null}
// Receive:
{"jsonrpc":"2.0","method":"$/progress","params":{"token":"a597e761-6ad4-4d01-acb7-6d83b82b2691","value":{"kind":"end","message":"Validate documents"}}}

Environment

Zed: v0.165.2 (Zed Preview)
OS: macOS 15.2.0
Memory: 8 GiB
Architecture: aarch64

If applicable, add mockups / screenshots to help explain present your vision of the feature

No response

If applicable, attach your Zed.log file to this issue.

Zed.log

[osiewicz]

I think we're actually doing the correct thing here; the completion you've listed has both insertText and textEdit fields, and in the LSP spec it is clearly stated that when textEdit is provided, insertText should be ignored.

/**
	 * An edit which is applied to a document when selecting this completion.
	 * When an edit is provided the value of `insertText` is ignored.
	 *
	 * *Note:* The range of the edit must be a single line range and it must
	 * contain the position at which completion has been requested.
	 *
	 * Most editors support two different operations when accepting a completion
	 * item. One is to insert a completion text and the other is to replace an
	 * existing text with a completion text. Since this can usually not be
	 * predetermined by a server it can report both ranges. Clients need to
	 * signal support for `InsertReplaceEdit`s via the
	 * `textDocument.completion.completionItem.insertReplaceSupport` client
	 * capability property.
	 *
	 * *Note 1:* The text edit's range as well as both ranges from an insert
	 * replace edit must be a [single line] and they must contain the position
	 * at which completion has been requested.
	 * *Note 2:* If an `InsertReplaceEdit` is returned the edit's insert range
	 * must be a prefix of the edit's replace range, that means it must be
	 * contained and starting at the same position.
	 *
	 * @since 3.16.0 additional type `InsertReplaceEdit`
	 */
	textEdit?: [TextEdit](https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#textEdit) | [InsertReplaceEdit](https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#insertReplaceEdit);

I'd suggest to file an issue with the language server. Feel free to reopen this issue if anything.

[valentinegb]

Hello again @osiewicz, I've spoken with a JDTLS dev here and we think this is in fact the fault of Zed. We noticed a discrepancy between the completion that Zed receives and the completion that Zed asks to autocomplete:

// Send:
{"jsonrpc":"2.0","id":38,"method":"textDocument/completion","params":{"textDocument":{"uri":"file:///Users/valentinebriese/Developer/SwerveDrive2025/src/main/java/frc/robot/subsystems/Test.java"},"position":{"line":0,"character":3},"context":{"triggerKind":1}}}
// Receive:
{"jsonrpc":"2.0","id":38,"result":{"isIncomplete":false,"items":[{"label":"class","kind":15,"documentation":"package frc.robot.subsystems;\n\npublic class Test {\n\n\t\n}","sortText":"999999212","filterText":"class","insertText":"package frc.robot.subsystems;\n\npublic class Test {\n\n\t${0}\n}","insertTextMode":2,"textEditText":"package frc.robot.subsystems;\n\npublic class Test {\n\n\t${0}\n}"},{"label":"class","kind":14,"sortText":"999999213","filterText":"class","textEditText":"class","command":{"title":"","command":"java.completion.onDidSelect","arguments":["10","0"]},"data":{"pid":"0","rid":"10"}}],"itemDefaults":{"editRange":{"insert":{"start":{"line":0,"character":0},"end":{"line":0,"character":3}},"replace":{"start":{"line":0,"character":0},"end":{"line":0,"character":3}}},"insertTextFormat":2,"data":{"completionKinds":[3]}}}}
// Send:
{"jsonrpc":"2.0","id":39,"method":"completionItem/resolve","params":{"label":"class","kind":15,"documentation":"package frc.robot.subsystems;\n\npublic class Test {\n\n\t\n}","sortText":"999999212","filterText":"class","insertText":"package frc.robot.subsystems;\n\npublic class Test {\n\n\t${0}\n}","insertTextFormat":2,"insertTextMode":2,"textEdit":{"newText":"class","insert":{"start":{"line":0,"character":0},"end":{"line":0,"character":3}},"replace":{"start":{"line":0,"character":0},"end":{"line":0,"character":3}}},"data":{"completionKinds":[3]}}}
// Receive:
{"jsonrpc":"2.0","id":39,"result":{"label":"class","kind":15,"documentation":"package frc.robot.subsystems;\n\npublic class Test {\n\n\t\n}","sortText":"999999212","filterText":"class","insertText":"package frc.robot.subsystems;\n\npublic class Test {\n\n\t${0}\n}","insertTextFormat":2,"insertTextMode":2,"textEdit":{"newText":"class","insert":{"start":{"line":0,"character":0},"end":{"line":0,"character":3}},"replace":{"start":{"line":0,"character":0},"end":{"line":0,"character":3}}}}}

This comment may be particularly informative.

I believe this means this issue should be reopened, which I am unable to do myself ^^'

[SomeoneToIgnore]

Oh, then it's #21277

[valentinegb]

Looks like it, sorry for making a duplicate issue then