Bug 169315
| Summary: | Web Inspector: addition test coverage for DOM breakpoints | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Matt Baker <mattbaker> |
| Component: | Web Inspector | Assignee: | Matt Baker <mattbaker> |
| Status: | NEW | ||
| Severity: | Normal | CC: | inspector-bugzilla-changes |
| Priority: | P2 | ||
| Version: | WebKit Nightly Build | ||
| Hardware: | All | ||
| OS: | All | ||
| Bug Depends on: | 168101 | ||
| Bug Blocks: | |||
Matt Baker
We should expand DOM breakpoint test coverage to include the following:
Bad protocol command input:
- DOMDebugger.setDOMBreakpoint with invalid nodeId
- DOMDebugger.setDOMBreakpoint with invalid type string
- DOMDebugger.removeDOMBreakpoint with invalid nodeId
- DOMDebugger.removeDOMBreakpoint with invalid type string
Manipulate the DOM while paused:
- Pause on a DOM "node-removed", then move the Node being removed while paused, then continue
- Pause on a DOM "subtree-modified", then modify the subtree again while paused, then continue
- Pause on a DOM "attribute-modified" for "style" attribute (this appears to have a different code path that is untested (InspectorDOMDebuggerAgent::didInvalidateStyleAttr))
Subframe breakpoints:
- Breakpoint should become unresolved when subframe removed
- Breakpoint should become unresolved then resolved when subframe resource reloaded
- Breakpoint should become unresolved then resolved when subframe removed from the DOM and reinserted
Misc:
- Should pause on a DOM "subtree-modified" when removing a Node from the subtree
- Should pause once if a parent has a "subtree-modified" breakpoint, a child has a "node-removed" breakpoint, and the child is removed (the node-removed breakpoint should be the expected pause reason)
- A Node with all possible breakpoint types ("node-removed", "subtree-modified", "attribute-modified")
- Pause on a regular breakpoint, then continue through code that would trigger a DOM breakpoint
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |