Bugzilla – Full Text Bug Listing
| Summary: | An OCRHelper class to read widget text | ||
|---|---|---|---|
| Product: | [Eclipse Project] Platform | Reporter: | Mickael Istria <mistria> |
| Component: | UI | Assignee: | Platform-UI-Inbox <Platform-UI-Inbox> |
| Status: | NEW --- | QA Contact: | |
| Severity: | enhancement | ||
| Priority: | P3 | CC: | akurtakov, azerr, daniel_megert, wim.jongman |
| Version: | 4.8 | Keywords: | helpwanted |
| Target Milestone: | --- | ||
| Hardware: | All | ||
| OS: | All | ||
| Whiteboard: | |||
Remove target. |
The Test harness should provide an OCRHelper method that would delegate to software like `tesseract` to read the screen content. That would make it very handy to test some cases of rendering like code minings. Command I tried successfully is `tesseract /path/to/screenshot.png stdout -psm 6` for a text editor; we can replace the path/to/screenshot by stdin and send the image of the widget to stdin stream. Usage could look like ``` styledText.setText("class A {}"); assertEquals("0 references\nclass A {}", OCRHelper.read(styledText)) ```