Download raw body.
double lock in tog(1)
If match_commit() fails there's no point in trying to re-grab the mutex that the thread already owns. Simply stop the iteration after having release the lock. Ok? diff --git tog/tog.c tog/tog.c index db59543..fc01887 100644 --- tog/tog.c +++ tog/tog.c @@ -1414,20 +1414,16 @@ queue_commits(struct got_commit_graph *graph, struct commit_queue *commits, commits->ncommits++; if (*searching == TOG_SEARCH_FORWARD && !*search_next_done) { err = match_commit(&have_match, id, commit, regex); - if (err) { - pthread_mutex_lock(&tog_mutex); - break; - } } errcode = pthread_mutex_unlock(&tog_mutex); if (errcode && err == NULL) err = got_error_set_errno(errcode, "pthread_mutex_unlock"); - if (have_match) + if (err || have_match) break; } return err;
double lock in tog(1)