[Solved] Composer does not clear after submitting a reply.


  • Yeah, I'll try to update as I can. I'm running it through a proxy (IIS/ARR) and I found a couple 502 errors in my Chrome console, so I'm wondering if that might be related. I was able to load the URLs it was complaining about 502s on, so it could be IIS misbehaving.

    I have access to the direct forum since it's self-hosted, so I will try running it that way and report back if I find it is still having the issue without the proxy in the middle.


  • Okay, so I thought this might have been a problem with my IIS proxy, because Chrome's console was reporting 502 (bad gateway) errors, but I was connecting directly to the local IP and the same issue came up. The screenshot shows the console while connecting through the local network.

    ChromeConsoleErrors.png

    Configuration details:
    Ubuntu Server 14.04.1 running on a VM (VMware Workstation 9)
    nodejs v0.10.25
    NodeBB 0.4.3 master
    The rest of the stack is whatever was installed by following the directions here: https://docs.nodebb.org/en/latest/installing/os/ubuntu.html

    I have added the imgur plugin (nodebb-plugin-imgur) and the dbsearch plugin (nodebb-plugin-dbsearch) and the local emailer plugin (nodebb-plugin-emailer-local). Everything else is a pretty much stock install of NodeBB. I run this through a proxy usually, but in this case I am circumventing the proxy and directly accessing the Ubuntu server by it's local IP. NodeBB was started with sudo ./nodebb -start (too lazy to worry about the proper permissions for port 80 right now).

    I still can't cause it to happen. It seems to recur more often once it's happened once (and closing the browser and opening a new session seems to clear it up for a while) but I haven't found any rhyme or reason to it. And clicking the links listed in the console results in correctly loading items, usually something like:

    OOSHD_6bdZFPLw2_Uok7:60:60:websocket,xhr-polling,jsonp-polling,flashsocket

    Please let me know if there's any additional information I can provide.

  • GNU/Linux Admin

    Could it be that your server is crashing? When you try accessing those links, the server may be back up.

    Try running your server in development mode (./nodebb dev) and see if you can trigger it again.


  • @julian can confirm this bug, it happened to me today a few times, (I've been downloading about 50GB worth of stuff onto my new PC and not sure if it's that that caused it, will see with this reply.

    Edit: it didn't happen, but I will try when I'm maxing my net out again, as it could be linked.


  • @julian said:

    Could it be that your server is crashing? When you try accessing those links, the server may be back up.

    Try running your server in development mode (./nodebb dev) and see if you can trigger it again.

    It could be, but the log is clean. I'll run it in development mode for a while and see.


  • So it does seem that something is causing the NodeBB service to restart. I created a test thread with lots of markdown and submitted a bunch of replies to myself and then started doing some quoted replies and I noticed the NodeBB service repeating the info and warnings it gives when it first starts up.

    Whatever happened before then was quickly scrolled off the screen, and I don't know where I might find any logs to see the problem.


  • @Shaun Run the forum in ./nodebb dev and leave the Terminal window open, when it crashes, it will show you the full output, not just from when you went to nodebb log. 👍


  • @a_5mith I was, but I'm not the most proficient linux user so I hadn't figured out how to scroll back in the log. Now that I have, this seems to be the error:

    error: TypeError: Cannot read property 'uniqueId' of null
    at shouldPush (/home/shaun/nodebb/src/notifications.js:163:30)
    at /home/shaun/nodebb/src/notifications.js:132:5
    at /home/shaun/nodebb/node_modules/async/lib/async.js:125:13
    at Array.forEach (native)
    at _each (/home/shaun/nodebb/node_modules/async/lib/async.js:46:24)
    at Object.asnyc.each (/home/shaun/nodebb/node_modules/async/lib/async.js:124:9)
    at /home/shaun/nodebb/src/notifications.js:127:10
    at /home/shaun/nodebb/src/notifications.js:34:12
    at /home/shaun/nodebb/src/database/redis/main.js:68:4
    at try_callback (/home/shaun/nodebb/node_modules/redis/index.js:573:9)


  • Further detail:

    The composer window is locked open, but pressing discard will remove it. I can toggle the display with the thread button (bottom right corner) but I cannot submit the reply (probably because it was already submitted).

    If I click the reply buttons in the thread while the composer is open, it does not clear the existing composer window and the submit button is still unavailable.

    If I quote an existing reply, it updates the text of the composer with the quoted post, but submit button is still unavailable.

    Pressing discard results in the [OK/Cancel] modal popup, and confirming on that popup does clear the composer.


  • Seems to be happening every time I submit a quoted reply right now. Same error, same line numbers.


  • @Shaun said:

    Seems to be happening every time I submit a quoted reply right now. Same error, same line numbers.
    Quoting seems to work for me here. :s

  • NodeBB Admin


  • @baris I can probably figure it out, but would you care to supply the specific commands to pull that down from git so that I can't screw it up?

  • NodeBB Admin

    If you are on the master branch just a git pull should be enough to get the latest.


  • Haven't seen it happen since the pull.


Suggested Topics

| | |

© 2014 – 2022 NodeBB, Inc. — Made in Canada.