Unless I'm missing something, the addYield method leaks memory.
func (r *renderer) addYield(name string, binding interface{}) {
funcs := template.FuncMap{
"yield": func() (template.HTML, error) {
buf, err := r.execute(name, binding)
// return safe html here since we are rendering our own template
return template.HTML(buf.String()), err
},
"current": func() (string, error) {
return name, nil
},
}
r.t.Funcs(funcs)
}
r.execute method Gets a buffer from the buffer pool, but the buffer is never Put back to the pool. I suggest getting rid of the r.execute method (it's two lines of code) and essentially in-lining its functionality. Doing so will allow deferring the buffer Put immediately after the the buffer Get. Seeing these paired together makes it much easier to ensure the buffer is returned.
Unless I'm missing something, the addYield method leaks memory.
r.execute method Gets a buffer from the buffer pool, but the buffer is never Put back to the pool. I suggest getting rid of the r.execute method (it's two lines of code) and essentially in-lining its functionality. Doing so will allow deferring the buffer Put immediately after the the buffer Get. Seeing these paired together makes it much easier to ensure the buffer is returned.