SWTBot Test Support

Release 0.1.0 ยท All versions

Screenshots

Create one instance per test/session using your application's existing Display:

Screenshots screenshots = new Screenshots(
    CaptureOptions.defaults(outputDirectory), testIdentity, () -> applicationDisplay);
screenshots.requireCheckpoint("01_precondition", () -> applicationShell);
screenshots.guard("failure_action", () -> applicationShell, () -> {
    performUserAction();
    assertTargetState();
});
screenshots.requireCheckpoint("02_target_state", () -> applicationShell);

Import Screenshots and CaptureOptions from io.klib.swtbot.support. Application variables/actions above belong to your test. The library does not create a Display, find windows or assert application state. Shell suppliers run on the supplied Display's UI thread. Actions passed to guard run on the calling thread.

Capture Options

Defaults capture client area only, with a five-second completion budget per mode. Explicitly opt into other regions:

new CaptureOptions(outputDirectory, Duration.ofSeconds(10),
    List.of(CaptureMode.SHELL_CLIENT_AREA, CaptureMode.SHELL_DISPLAY_BOUNDS));
Mode Artifact label Region
SHELL_CLIENT_AREA shell Shell client area via SWT graphics
SHELL_DISPLAY_BOUNDS swt Shell rectangle including window decorations
DESKTOP screen Display bounds, potentially multiple monitors

Screen-based capture can include occlusion, other applications and private information. Keep the target visible. Scaling and window manager affect pixel dimensions. Do not publish test screenshots without reviewing their contents.

Outcomes and Artifacts

capture returns one CaptureResult; checkpoint returns outcomes for configured modes. Status is CAPTURED, SKIPPED, or FAILED. requireCaptured and requireCheckpoint reject unavailable evidence explicitly.

Each instance creates a separate run-<UUID>/test-<sanitized-id>/ directory. Unique PNG paths include checkpoint and mode; names are bounded. directory() gives the actual artifact directory. ArtifactPaths.allocate creates an empty reserved file, not screenshot evidence.

failure logs outcomes and suppresses diagnostic errors onto the original throwable. guard captures before caller teardown and rethrows that same throwable. Each instance diagnoses a throwable once and retains it for duplicate suppression.

Threading and Timeouts

UI-thread capture runs immediately. Other callers queue work and wait for the configured budget. Timeout/interruption cancels queued work; executing native operations cannot be forcibly stopped and may finish later. Interrupted callers retain interruption status. Native graphics and images are disposed after capture. Capture does not modify global SWTBot preferences.