* Fix lspFrame.source NPE on stackTrace request

Follow up to:

- microsoft#614
- microsoft#609

With the change to set the line number to 0 the jdiLineNumber !=
lspFrame.line comparison can evaluate to true:

    dap> lspFrame
    Types$StackFrame@78
      column: 1
      id: 6
      line: 0
      name: "0x000000002f0bc000.invokeVirtual(Object,Object)"
      presentationHint: "subtle"
      source: null

    dap> jdiLineNumber
    -1

`source` being null caused an NPE

* Fix line number comparison in StackTraceRequestHandler

---------

Co-authored-by: Changyong Gong <shawn.gong@hotmail.com>

* fix: lspFrame.source.path is null

* fix: update bundle location

* feat: support suspend all setting

* fix: handle stoppedEvent for suspend policy change

* fix: not allow change policy during a live debug session

* fix: format issue

* fix: let step command execute for current thread

* fix: remove unused import

* fix: apply suspend policy to exception breakpoint

* fix: improve naming

…ft#621)

Map.Entry objects were incorrectly treated as lazy-loading candidates
because they are included in COLLECTION_TYPES. This caused the debugger
to display entries as 'HashMap$Node@id' instead of showing the actual
key:value details inline.

Map.Entry's details computation (getKey + getValue) is lightweight,
so eager evaluation is safe and significantly improves UX when debugging
large Maps.

Fixes microsoft/vscode-java-debug#1605

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Replace manual jdt:// URI construction in getFileURI(IClassFile) with
JDTUtils.toUri(classFile) to align with eclipse.jdt.ls PR #3666.

This ensures the debugger generates the same URI format as the language
server, fixing the issue where classes with source code are opened twice
(once as .java, once as .class) during debugging.

Fixes microsoft#623
Related: eclipse-jdtls/eclipse.jdt.ls#3729

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…T Core (microsoft#622)

Catch NoSuchMethodError when calling SourceMethod.isMainMethodCandidate()
in ResolveMainClassHandler and ResolveMainMethodHandler. This method was
added in JDT Core 3.36 and is unavailable on older versions, causing
the debugger to fail entirely when resolving main classes.

Fixes microsoft/vscode-java-debug#1598

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Changyong Gong <shawn.gong@hotmail.com>

… definition (microsoft#625)

* added attributes to VariablePresentationHint as specified in protocol definition

* Fixed issue with lazy VariablePresentationHint.
Restricted constant to Primitive & StringRefs

…rosoft#626)

* Ignore VMDisconnected on detach; reorder terminate

Wrap vm.dispose() in a try/catch to ignore VMDisconnectedException when the VM has already disconnected (e.g. process terminated) to avoid spurious errors during detach. Reorder calls in DisconnectRequestHandler to invoke terminate() before detach() when terminateDebuggee is true and the session is not attached, ensuring the debuggee is terminated prior to detaching.

* Always detach debug session on terminate

Wrap debugSession.terminate() in a try/finally so debugSession.detach() is always invoked when terminateDebuggee is true and the session is not attached. This ensures the session is detached even if terminate() throws, avoiding potential dangling sessions. Other detach logic is unchanged.

* Prefer sourcePaths before JDT source URIs

When JDT source lookup returns a non-file URI, check debug sourcePaths before returning that URI to the client.

Add a regression test covering cached JDT URI results being overridden by sourcePaths.

* Address source path review feedback

Use a null-safe SourceType comparison and convert sourcePaths matches to the client's expected path format.

Extend regression coverage for URI clients and null source types.

Co-authored-by: titanniya542-spec <titanniya542-spec@users.noreply.github.com>