pdfcpu / pdfcpu

The filters usually return a *bytes.Buffer, so copying it into a temporary bytes.Buffer to get the underlying []byte can be avoided.

Example:

pdfcpu/pkg/pdfcpu/types/streamdict.go

Lines 292 to 297 in fc87a22

    
           var buf bytes.Buffer 
        
           if _, err := io.Copy(&buf, c); err != nil { 
        
           	return err 
        
           } 
        
           sd.Content = buf.Bytes()

Same applies for the filters itself which can be chained and in such casees also don't need to copy from the passed io.Reader if it is actually a *bytes.Buffer.

	var buf bytes.Buffer
	if _, err := io.Copy(&buf, c); err != nil {
	return err
	}

	sd.Content = buf.Bytes()

Avoid unnecessary copying with filters.