Derpitron/Discord-OTP-Forcer

[Bug]: ElementNotInteractableException raised and program crash, when code submission button randomly becomes uninteractible.

Closed this issue · 30 comments

SMG54321 commented

Description

Program crashes with NoSuchElementException after random uptime of running.
Occurs independently of program mode.
Cause:
Code submission button on code entry page may randomly have an uninteractible state (appears as 3 pulsing "loading" dots) (probably due to a pending network request in the page). When the program attempts to click the button after the specified sleep timeout elapses, it tries to interact with the uninteractible button, raising ElementNotInteractableException. This then causes the rest of the program to collapse and crash.

When crashing, the program raises a bunch of other exceptions and prints random hex-code stack traces to the terminal.

Minimal Reproduction

Conditions for crash are random.

  1. Run the program normally.
  2. Either on a slow internet connection or after considerable time elapsed, wait for the code submission button to appear "loading".
  3. Program crashes if the button does not re-fix itself before the internal program timeout elapses.

Error

Redacted Images

Operating System Version

Widows 10

Python Version

3.11, 3.12

Program Version

d1a5b03

Any other information or context?

Luminaex commented

I'm unable to recreate this issue so I will need more information.
What happens with the Chrome window during all of this?

SMG54321 commented

I'm unable to recreate this issue so I will need more information. What happens with the Chrome window during all of this?

everything that you would expect it to do. go to the right screen and start punching in codes at randomly spaced out pulses. then suddenly, it doesnt do that and breaks. if you want i can try to figure out how to record one of my screens. maybe me clicking off the screen(s) was enough of a problem?

SMG54321 commented

I'm unable to recreate this issue so I will need more information. What happens with the Chrome window during all of this?

everything that you would expect it to do. go to the right screen and start punching in codes at randomly spaced out pulses. then suddenly, it doesnt do that and breaks. if you want i can try to figure out how to record one of my screens. maybe me clicking off the screen(s) was enough of a problem?

everything looks like its working fine on chrome's end i think.

SMG54321 commented

is this something on my end?

SMG54321 commented

i think its the correct code thats breaking it, can you try to extract it from the exceptions and give it to me? in like email or something?

SMG54321 commented

GetHandleVerifier [0x009572A3+45731]
(No symbol) [0x008E2D51]
(No symbol) [0x007D880D]
(No symbol) [0x0080B940]
(No symbol) [0x0080BE0B]
(No symbol) [0x0083D1F2]
(No symbol) [0x00828024]
(No symbol) [0x0083B7A2]
(No symbol) [0x00827DD6]
(No symbol) [0x008031F6]
(No symbol) [0x0080439D]
GetHandleVerifier [0x00C60716+3229462]
GetHandleVerifier [0x00CA84C8+3523784]
GetHandleVerifier [0x00CA214C+3498316]
GetHandleVerifier [0x009E1680+611968]
(No symbol) [0x008ECCCC]
(No symbol) [0x008E8DF8]
(No symbol) [0x008E8F1D]
(No symbol) [0x008DB2C7]
BaseThreadInitThunk [0x75F1FCC9+25]
RtlGetAppContainerNamedObjectPath [0x770F7C6E+286]
RtlGetAppContainerNamedObjectPath [0x770F7C3E+238]

    GetHandleVerifier [0x009572A3+45731]
    (No symbol) [0x008E2D51]
    (No symbol) [0x007D86D0]
    (No symbol) [0x00808118]
    (No symbol) [0x0080774F]
    (No symbol) [0x00827FDC]
    (No symbol) [0x00804A4E]
    (No symbol) [0x00828254]
    (No symbol) [0x0083B7A2]
    (No symbol) [0x00827DD6]
    (No symbol) [0x008031F6]
    (No symbol) [0x0080439D]
    GetHandleVerifier [0x00C60716+3229462]
    GetHandleVerifier [0x00CA84C8+3523784]
    GetHandleVerifier [0x00CA214C+3498316]
    GetHandleVerifier [0x009E1680+611968]
    (No symbol) [0x008ECCCC]
    (No symbol) [0x008E8DF8]
    (No symbol) [0x008E8F1D]
    (No symbol) [0x008DB2C7]
    BaseThreadInitThunk [0x75F1FCC9+25]
    RtlGetAppContainerNamedObjectPath [0x770F7C6E+286]
    RtlGetAppContainerNamedObjectPath [0x770F7C3E+238]

its gotta be be one of these two! can you tell me the codes these alligned up to?

Derpitron commented

Man why are you creating so many issues, especially with weird ambiguous titles?

You've already made one issue before. Re-open that one, and put your other screenshots in it. Spamming this repo with issues will only make it harder for us to track each one.

SMG54321 commented

Man why are you creating so many issues, especially with weird ambiguous titles?

You've already made one issue before. Re-open that one, and put your other screenshots in it. Spamming this repo with issues will only make it harder for us to track each one.

i can do that?

SMG54321 commented

Man why are you creating so many issues, especially with weird ambiguous titles?
You've already made one issue before. Re-open that one, and put your other screenshots in it. Spamming this repo with issues will only make it harder for us to track each one.

i can do that?

if you want an idea on how dumb i am, here it is!

Derpitron commented

@SMG54321 this might be too much of a time waste to ask.

But I think the only way we could find the cause of this bug is to literally monitor the program the entire time.

You could:

  1. Stare at your screen with the program and terminal open, and keep an eye on when and what conditions the program crashes (assuming it even does)
  2. Record your screen right when you start the program, keeping the Selenium window and the Terminal window open at the same time. Then stopping the recording if/when the program crashes. Then sending the recording to us/me.

I think this is the only way any of us can find the cause.

SMG54321 commented

by program/selenium window, you mean the chrome one right? i can do that! i have OBS!

Derpitron commented

I mean recording both the chrome window AND your command prompt at the same time. Do it if you're able to. Would be very helpful

SMG54321 commented

this link will let you download the vid that i recorded

SMG54321 commented
SMG54321 commented
  • ❤️1
Derpitron commented

Thanks for the video. Now I know what's going on!

It was not a problem with the CSS Class name.

This is the entire course of the glitch:

  1. The Discord website takes ~4 seconds to accept the code after the submit button is pressed, locking the submission button.
  2. But the program expected the submission button to be interactible after 4 seconds. I.e the program didn't wait for the page to finish loading.
  3. The program has no idea what to do, because the submit button isn't interactible.
  4. The program raises the error: selenium.common.exceptions.ElementNotInteractableException: Message: element not interactable.
  5. Since the program has an error, the rest of it falls down and crashes painfully.

I think the solution to this problem is to somehow make the program "wait indefinitely" for the button to become interactible, without raising an exception.

Luminaex commented

@SMG54321 I've implemented a check to see if the confirm button is clickable so hopefully this'll fix this issue.

SMG54321 commented

@SMG54321 I've implemented a check to see if the confirm button is clickable so hopefully this'll fix this issue.

ok, either me replacing the files of the last version broke the fix, or something else is going on!

i really dunno why.

Derpitron commented

Please manually pull the repo using git. Do not try to manually replace any files except for your own cfg.yml file.

In the future maybe I could implement some easy updater.

SMG54321 commented

THATS WAS THE PROBLEM!!!

Derpitron commented

Hey, please don't give vague descriptions like "this is weird" when you upload another screenshot/recording of the bug. Instead describe the issue in text (whatever you see/think is happening). Im not at my computer all day so am not able to check and watch the recordings all the time. Hope you can understand 🙂

Derpitron commented

@LuXeZs The fix you made should be applied in both normal and backup mode; You've mistakenly added it only for the backup mode.

Discord-OTP-Forcer/src/backend.py

Lines 106 to 116 in 3f2ce8e

match configuration['codeMode']:
case 'normal':
login_fields['TOTP'] = driver.find_element(by=By.XPATH, value="//input[@placeholder='6-digit authentication code']")
case 'backup':
driver.find_element(By.XPATH, "//*[contains(text(), 'Verify with something else')]").click()
WebDriverWait(driver, 10).until(EC.element_to_be_clickable((By.XPATH, "//*[contains(text(), 'Use a backup code')]"))) # Waits until element is clickable
driver.find_element(By.XPATH, "//*[contains(text(), 'Use a backup code')]").click()
driver.implicitly_wait(1)
login_fields['TOTP'] = driver.find_element(by=By.XPATH, value="//input[@placeholder='8-digit backup code']")
driver.implicitly_wait(1)

I think I know the issue more clearly now, I may have misdiagnosed it before. It's not just the code submission button, but also the code entry field that becomes uninteractible.

image

See how the entry field (blue) is greyed out? That means it's uninteractible. I.e you can't click on it or type anything, which is what the program tries to do.

Meanwhile, the code submit button (red) merely shows the loading icon, is also uninteractible. I can't click on it when it's in the loading state.

@LuXeZs You implemented the fix only for the button, not for the field.

Discord-OTP-Forcer/src/backend.py

Line 186 in 3f2ce8e

WebDriverWait(driver, 10).until(EC.element_to_be_clickable((By.XPATH, "//*[contains(text(), 'Confirm')]")))

Luminaex commented

Alright added checks to see if the submit button is active before submitting the code, as well I added a check to see if the text field is also interactable.

SMG54321 commented

Hey, please don't give vague descriptions like "this is weird" when you upload another screenshot/recording of the bug. Instead describe the issue in text (whatever you see/think is happening). Im not at my computer all day so am not able to check and watch the recordings all the time. Hope you can understand 🙂

FUCC, i keep doing that and wonder why i dont get any decent help. i am so sorry!

Derpitron commented

@SMG54321 program has been updated. Please redownload (with your cfg.yml as usual) and test away, no need to record this time.

Derpitron commented

image
Seems to not work for me.

Derpitron commented

image
This is the code working as intended believe it or not.

Discord-OTP-Forcer/src/backend.py

Line 198 in 63891d7

WebDriverWait(driver, 10).until(EC.element_to_be_clickable((By.XPATH, "//*[contains(text(), 'Confirm')]")))

Discord-OTP-Forcer/src/backend.py

Line 243 in 63891d7

WebDriverWait(driver, 10).until(EC.element_to_be_clickable((login_fields['TOTP']))) # Waits until element is clickable

The program waits 10 seconds for the button to load. Since it didn't, it crashed. That's just how it works.

Derpitron commented

Should be guarded against in 2beb3ba. This commit should hopefully fix the whole issue.

Derpitron commented

image
Works for me! @SMG54321 try it out.

Derpitron commented

No crashing. Im considering this solved, and @SMG54321 I am deleting the videos you posted since they contain sensitive info.