Bugzilla – Full Text Bug Listing
| Summary: | Incorrect use of getElementsByTagName in DBGp property decoding | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Technology] DLTK | Reporter: | Julien Desgats <jdesgats> | ||||
| Component: | Common-Debug | Assignee: | dltk.common-inbox <dltk.common-inbox> | ||||
| Status: | RESOLVED FIXED | QA Contact: | |||||
| Severity: | major | ||||||
| Priority: | P3 | CC: | alex.panchenko, contact | ||||
| Version: | 3.0 | ||||||
| Target Milestone: | --- | ||||||
| Hardware: | All | ||||||
| OS: | All | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
Committed the patch. Thanks! |
Created attachment 200663 [details] Use only on direct child nodes for property parsing When property elements are parsed, the DOM method getElementsByTagName is used. This does not search in children but in all descendant nodes. This lead to problems with nested properties because the nodes are returned in the document order. For example, take the following Python dictionary { "a": "foo", "b" : { "c": 12, "d": [1, True, "vf"]}, "e": False } It is not displayed like expected because: * The first 3 children returned by getElementsByTagName are "a", "b" and "c" instead of "a", "b" and "e" * The first "name" element is used and it could be found in a descendant node (at least it is with my debugger) The proposed patch search only in direct children on property parsing. I've made a quick review of other getElementsByTagName uses on DBGp part, and except for property elements, they look fine (property is the only recursive structure in DBGp protocol AFAIK).