5 files changed, 86 insertions, 0 deletions

diff --git a/activitypub.c b/activitypub.c
index d699fee..5fb60ba 100644
--- a/activitypub.c
+++ b/activitypub.c
@@ -3377,6 +3377,12 @@ void process_queue_item(xs_dict *q_item)
         rss_poll_hashtags();
     }
     else
+    if (strcmp(type, "fsck") == 0) {
+        srv_log(xs_fmt("started deferred data integrity check"));
+        data_fsck();
+        srv_log(xs_fmt("finished deferred data integrity check"));
+    }
+    else
         srv_log(xs_fmt("unexpected q_item type '%s'", type));
 }
 
diff --git a/data.c b/data.c
index b36e31e..1533305 100644
--- a/data.c
+++ b/data.c
@@ -3605,6 +3605,17 @@ void enqueue_collect_outbox(snac *user, const char *actor_id)
 }
 
 
+void enqueue_fsck(void)
+/* enqueues an fsck */
+{
+    xs *qmsg   = _new_qmsg("fsck", "", 0);
+    const char *ntid = xs_dict_get(qmsg, "ntid");
+    xs *fn     = xs_fmt("%s/queue/%s.json", srv_basedir, ntid);
+
+    qmsg = _enqueue_put(fn, qmsg);
+}
+
+
 int was_question_voted(snac *user, const char *id)
 /* returns true if the user voted in this poll */
 {
@@ -4362,3 +4373,61 @@ const char *lang_str(const char *str, const snac *user)
 
     return n_str;
 }
+
+
+/** integrity checks **/
+
+void data_fsck(void)
+{
+    xs *list = user_list();
+    const char *uid;
+
+    xs_list_foreach(list, uid) {
+        snac user;
+
+        if (!user_open(&user, uid))
+            continue;
+
+        {
+            /* iterate all private posts and check that non-public posts
+               from this user are also linked into the public directory,
+               to avoid the don't-fucking-delete-my-own-private-posts purge bug */
+
+            xs *priv_spec = xs_fmt("%s/private/""*.json", user.basedir);
+            xs *posts = xs_glob(priv_spec, 0, 0);
+            const char *priv_fn;
+
+            xs_list_foreach(posts, priv_fn) {
+                xs *pub_fn = xs_replace(priv_fn, "/private/", "/public/");
+
+                /* already there? look no more */
+                if (mtime(pub_fn))
+                    continue;
+
+                /* read the post */
+                FILE *f;
+                if ((f = fopen(priv_fn, "r")) == NULL)
+                    continue;
+
+                xs *post = xs_json_load(f);
+                fclose(f);
+
+                if (!xs_is_dict(post))
+                    continue;
+
+                const char *attr_to = get_atto(post);
+
+                if (!xs_is_string(attr_to) || strcmp(attr_to, user.actor) != 0) {
+                    /* not from this user */
+                    continue;
+                }
+
+                /* link */
+                snac_debug(&user, 1, xs_fmt("fsck: fixed missing link %s", xs_dict_get(post, "id")));
+                link(priv_fn, pub_fn);
+            }
+        }
+
+        user_free(&user);
+    }
+}
diff --git a/httpd.c b/httpd.c
index 9707f9c..50f56f2 100644
--- a/httpd.c
+++ b/httpd.c
@@ -805,6 +805,8 @@ static void *background_thread(void *arg)
 
     srv_log(xs_fmt("background thread started"));
 
+    enqueue_fsck();
+
     while (p_state->srv_running) {
         int cnt = 0;
 
diff --git a/main.c b/main.c
index f767355..19cfe1a 100644
--- a/main.c
+++ b/main.c
@@ -34,6 +34,7 @@ int usage(const char *cmd)
         "httpd {basedir}                      Starts the HTTPD daemon\n"
         "purge {basedir}                      Purges old data\n"
         "state {basedir}                      Prints server state\n"
+        "fsck {basedir}                       Performs a non-destructive data integrity check\n"
         "webfinger {basedir} {account}        Queries about an account (@user@host or actor url)\n"
         "queue {basedir} {uid}                Processes a user queue\n"
         "follow {basedir} {uid} {actor}       Follows an actor\n"
@@ -200,6 +201,11 @@ int main(int argc, char *argv[])
         return 0;
     }
 
+    if (strcmp(cmd, "fsck") == 0) { /** **/
+        data_fsck();
+        return 0;
+    }
+
     if ((user = GET_ARGV()) == NULL)
         return usage(cmd);
 
diff --git a/snac.h b/snac.h
index e4f5262..e344f46 100644
--- a/snac.h
+++ b/snac.h
@@ -299,6 +299,7 @@ void enqueue_webmention(const xs_dict *msg);
 void enqueue_notify_webhook(snac *user, const xs_dict *noti, int retries);
 void enqueue_collect_replies(snac *user, const char *post);
 void enqueue_collect_outbox(snac *user, const char *actor_id);
+void enqueue_fsck(void);
 
 int was_question_voted(snac *user, const char *id);
 
@@ -477,3 +478,5 @@ xs_str *rss_from_timeline(snac *user, const xs_list *timeline,
                         const char *title, const char *link, const char *desc);
 void rss_to_timeline(snac *user, const char *url);
 void rss_poll_hashtags(void);
+
+void data_fsck(void);