doveadm deduplicate: Optimize by deduplicating in a single transaction

diff --git a/src/doveadm/doveadm-mail-deduplicate.c b/src/doveadm/doveadm-mail-deduplicate.c
index 2854a4a746a3efb12e14790061c8f604d9f1e3bc..d0ce3d2999ab6ef345421383e714043262541d73 100644 (file)
--- a/src/doveadm/doveadm-mail-deduplicate.c
+++ b/src/doveadm/doveadm-mail-deduplicate.c
@@ -8,63 +8,11 @@
 #include "doveadm-mail-iter.h"
 #include "doveadm-mail.h"
 
-struct uidlist {
-       struct uidlist *next;
-       uint32_t uid;
-};
-
 struct deduplicate_cmd_context {
        struct doveadm_mail_cmd_context ctx;
        bool by_msgid;
 };
 
-static int cmd_deduplicate_uidlist(struct doveadm_mail_cmd_context *_ctx,
-                                  struct mailbox *box, struct uidlist *uidlist)
-{
-       struct mailbox_transaction_context *trans;
-       struct mail_search_context *search_ctx;
-       struct mail_search_args *search_args;
-       struct mail_search_arg *arg;
-       struct mail *mail;
-       ARRAY_TYPE(seq_range) uids;
-       int ret = 0;
-
-       /* the uidlist is reversed with oldest mails at the end.
-          we'll delete everything but the oldest mail. */
-       if (uidlist->next == NULL)
-               return 0;
-
-       t_array_init(&uids, 8);
-       for (; uidlist->next != NULL; uidlist = uidlist->next)
-               seq_range_array_add(&uids, uidlist->uid);
-
-       search_args = mail_search_build_init();
-       arg = mail_search_build_add(search_args, SEARCH_UIDSET);
-       arg->value.seqset = uids;
-
-       trans = mailbox_transaction_begin(box, _ctx->transaction_flags, __func__);
-       search_ctx = mailbox_search_init(trans, search_args, NULL, 0, NULL);
-       mail_search_args_unref(&search_args);
-
-       while (mailbox_search_next(search_ctx, &mail))
-               mail_expunge(mail);
-       if (mailbox_search_deinit(&search_ctx) < 0) {
-               i_error("Searching mailbox '%s' failed: %s",
-                       mailbox_get_vname(box),
-                       mailbox_get_last_internal_error(box, NULL));
-               doveadm_mail_failed_mailbox(_ctx, box);
-               ret = -1;
-       }
-       if (mailbox_transaction_commit(&trans) < 0) {
-               i_error("Committing mailbox '%s' transaction failed: %s",
-                       mailbox_get_vname(box),
-                       mailbox_get_last_internal_error(box, NULL));
-               doveadm_mail_failed_mailbox(_ctx, box);
-               ret = -1;
-       }
-       return ret;
-}
-
 static int
 cmd_deduplicate_box(struct doveadm_mail_cmd_context *_ctx,
                    const struct mailbox_info *info,
@@ -77,9 +25,8 @@ cmd_deduplicate_box(struct doveadm_mail_cmd_context *_ctx,
        struct mail *mail;
        enum mail_error error;
        pool_t pool;
-       HASH_TABLE(const char *, struct uidlist *) hash;
+       HASH_TABLE(const char *, void *) hash;
        const char *key, *errstr;
-       struct uidlist *value;
        int ret = 0;
 
        if (doveadm_mail_iter_init(_ctx, info, search_args, 0, NULL, 0,
@@ -113,35 +60,16 @@ cmd_deduplicate_box(struct doveadm_mail_cmd_context *_ctx,
                        }
                }
                if (key != NULL && *key != '\0') {
-                       value = p_new(pool, struct uidlist, 1);
-                       value->uid = mail->uid;
-                       value->next = hash_table_lookup(hash, key);
-
-                       if (value->next == NULL) {
-                               key = p_strdup(pool, key);
-                               hash_table_insert(hash, key, value);
-                       } else {
-                               hash_table_update(hash, key, value);
-                       }
+                       if (hash_table_lookup(hash, key) != NULL)
+                               mail_expunge(mail);
+                       else
+                               hash_table_insert(hash, key, POINTER_CAST(1));
                }
        }
 
        if (doveadm_mail_iter_deinit_keep_box(&iter, &box) < 0)
                ret = -1;
 
-       if (ret == 0) {
-               struct hash_iterate_context *iter;
-
-               iter = hash_table_iterate_init(hash);
-               while (hash_table_iterate(iter, hash, &key, &value)) {
-                       T_BEGIN {
-                               if (cmd_deduplicate_uidlist(_ctx, box, value) < 0)
-                                       ret = -1;
-                       } T_END;
-               }
-               hash_table_iterate_deinit(&iter);
-       }
-
        hash_table_destroy(&hash);
        pool_unref(&pool);