Is it a bug here when setting breakpoints at return address of multiple functions

[hitzhangjie]

trafficstars

func setBreakpoint(t *Session, ctx callContext, tracepoint bool, argstr string) ([]*api.Breakpoint, error) {
        ...
	switch t := loc.(type) {
	case *locspec.NormalLocationSpec:
		shouldSetReturnBreakpoints = t.LineOffset == -1 && t.FuncBase != nil
	case *locspec.RegexLocationSpec:
		shouldSetReturnBreakpoints = true
	}


	if tracepoint && shouldSetReturnBreakpoints && locs[0].Function != nil {
		for i := range locs {
			if locs[i].Function == nil {
				continue
			}

                        // here, if we use RegexLocationSpec for multiple functions, 
                        // is `locs[0].Function.Name()` a bug here?
			addrs, err := t.client.(*rpc2.RPCClient).FunctionReturnLocations(locs[0].Function.Name())
			if err != nil {
				return nil, err
			}
			for j := range addrs {
				_, err = t.client.CreateBreakpoint(&api.Breakpoint{
					Addr:        addrs[j],
					TraceReturn: true,
					Line:        -1,
					LoadArgs:    &ShortLoadConfig,
				})
				if err != nil {
					return nil, err
				}
			}
		}
	}
	return created, nil
}
```go

[aarzilli]

Yes, it does look like a bug, I think it should be locs[i] and the nilness check should be inside the loop and also be on locs[i].