How to record unique covered regions

I am wondering how the covered regions in fuzzbench are being identified. I check code and found `coverage_utils.extract_covered_regions_from_summary_json` extracts information from Clang's sourced base coverage but not pretty sure how this approach guarantee the unique code regions.

```
for function_data in functions_data:
for region in function_data['regions']:
if region[hit_index] != 0 and region[type_index] == 0:
covered_regions.append(region[:hit_index] +
region[file_index:])
```
The above loop checks all covered regions in each function and construct the code locations (starting and ending line/column, file id, etc.) that have been executed. I'd like to understand what actually happen when we have **the exact starting and ending lines/columns but from different functions**.

Let's say we have two executed regions a) `3:6 to 3:12 in function foo()` and another region b) `3:6 to 3:12 in function bar()`, would these be considered different region?

Also, I'd like to get some ideas why fuzzbench uses region coverage as the primary coverage measurement. I notice clang source-based also offers branch coverage which will perhaps give more granular view, are there any advantages of region coverage over branch coverage? Does fuzzbench also support branch coverage now?

Thank you very much.