dropdown becomes inconsistent with change effects

[pseudotensor]

Describe the bug

When over internet connection (not local) there is a 0.5s-1s delay in the change operation, and if user changes the dropdown on a shorter timescale, then the effects are somehow not queued in sequence properly and the UI state becomes inconsistent with the cumulative changes.

Have you searched existing issues? 🔎

• I have searched and found no existing issues

Reproduction

• Need to run over slower internet, like server in India and UI in USA, or go over slow connection to mimic.
• Run script
• Do what video shows, by quickly deselecting some, see how not consistent in end.
• Worse when there is queue of events, so lag is even longer

import random
import time

import gradio as gr


def sentence_builder(quantity, animal, countries, place, activity_list, morning):
    return f"""The {quantity} {animal}s from {" and ".join(countries)} went to the {place} where they {" and ".join(activity_list)} until the {"morning" if morning else "night"}"""


demo = gr.Blocks()

with demo:
    num = 10
    choices = [str(x) for x in list(range(0, num))]
    choices0 = [str(x) for x in list(range(0, num - 1))]
    vis_list = gr.Dropdown(
        choices,
        value=choices0,
        multiselect=True,
        label="Model",
    )

    msg = gr.Textbox()

    def respond(*args):
        message = args[0]
        chat_historys = args[1:]
        bot_message = random.choice(["How are you?", "I love you", "I'm very hungry"])
        for chat_history in chat_historys:
            chat_history.append((message, bot_message))
        time.sleep(2)
        return tuple(chat_historys)

    chatbots = [gr.Chatbot(visible=str(x) in choices0, label=str(x)) for x in range(0, num)]
    msg.submit(respond, [msg] + chatbots, chatbots)

    def vis_change(vis_list1):
        vis_list1 = [int(x) for x in vis_list1]
        ret_list = [gr.update(visible=xi in vis_list1) for xi in list(range(0, 10))]
        return tuple(ret_list)
    vis_list.change(vis_change, inputs=vis_list, outputs=chatbots)
demo.queue()
demo.launch(server_name='0.0.0.0')

Screenshot

vokoscreenNG-2023-09-17_22-04-18.mp4

Logs

Not quite blocking all usage, but leads to poor UX.

System Info

gradio==3.41.0
gradio_client==0.3.0

Severity

Blocking usage of gradio

[abidlabs]

This is actually the same issue as #1954 -- we'll fix this in 4.0 @pseudotensor

[abidlabs]

Fixed in v4 branch, thanks @dawoodkhan82!