This study investigated the potential of comprehensive corrective feedback forms as editing and learning tools and examined their effect on learners' cognitive and attitudinal engagement. Low-intermediate second language writers (N = 139) were randomly assigned to four experimental conditions (direct corrections of grammatical errors, metalinguistic codes for grammatical errors, direct corrections of grammatical and nongrammatical errors, or metalinguistic codes for grammatical and nongrammatical errors) and a control group (self-correction). Results from mixed-effects linear models showed that although direct corrections and codes were effective for enhancing learners' immediate grammatical and nongrammatical accuracy (i.e., during text revision), a long-term advantage (i.e., 4 weeks after feedback provision) was only evident for direct corrections. A mental effort measure of cognitive load revealed that participants' cognitive load was significantly lower when processing direct corrections targeting grammar issues. Questionnaire answers also yielded a significant attitudinal difference between the direct feedback groups and their metalinguistic counterparts.