1827190 - The reply button doesn't work on solely one of my accounts.

Status: RESOLVED WORKSFORME

(bugzilla.mozilla.org :: User Interface, defect)

Description

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

Environment

User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/111.0

Steps to reproduce

I try to use the reply button.

Actual results

It doesn't do anything.

Expected results

It should add something, like what id=1701123#c17 includes to the start of my comment.

Comment 1

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

Attachment: Reproduction screencast.

This has now persisted across multiple installations of multiple browsers:

1. ⪅ google-chrome-canary-138.0.7167.0-1.x86_64

2. The desktop edition:

   1. ⪅ firefox-138.0.1-1.fc42.x86_64:

      Installed packages
      Name            : firefox
      Epoch           : 0
      Version         : 138.0.1
      Release         : 1.fc42
      Architecture    : x86_64
      Installed size  : 240.2 MiB
      Source          : firefox-138.0.1-1.fc42.src.rpm
      From repository : updates
      Vendor          : Fedora Project
      

   2. ⪅ github.com/microsoft/winget-pkgs/pull/178197:

      PackageIdentifier: Mozilla.Firefox.Nightly
      PackageVersion: 116.0a1
      

...on multiple OSes:

1. ⪅ cpe:/o:microsoft:windows_11_24h2

2. cpe:/o:opensuse:tumbleweed

3. cpe:/o:fedoraproject:fedora :40 to :42:

   KDE Plasma Version: 6.3.4
   Kernel Version: 6.14.5-300.fc42.x86_64 (64-bit)
   Graphics Platform: Wayland
   

Comment 2

[David Lawrence [:dkl]]

This is working for me Firefox 127. I am able to reply and the entire comment displays in the new comment text area. No errors are displayed in the javascript console. When you click the Reply do you see any errors on the console yourself?

Comment 3

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

Attachment: A Screencast That Demonstrates That The JS Console Does Not Produce Logs When The Button Is Pressed

#c2

I don't see anything caused by the button.

Comment 4

[08xjcec48]

Can you reproduce this issue with a fresh profile?

Comment 5

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

Attachment: A Screencast That Demonstrates That This Reproduces In A New Profile

Yes. I've attached a screencast of me creating the profile, authenticating at index.cgi?GoAheadAndLogIn=1 for the first time, then being unable to utilise the button.

Comment 6

[David Lawrence [:dkl]]

(In reply to Mr. Beedell, Roke Julian Lockhart from comment #5)

Created attachment 9486549 [details]
A Screencast That Demonstrates That This Reproduces In A New Profile

Yes. I've attached a screencast of me creating the profile, authenticating at index.cgi?GoAheadAndLogIn=1 for the first time, then being unable to utilise the button.

Is there any javascript or other type of error in the developer console that shows up right after hitting the reply button?

Comment 7

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

Attachment: A Screencast Of The JS And Network Consoles

#c6

No; no difference to #c3. Does your question differ to what you requested in #c2?

Comment 8

[David Lawrence [:dkl]]

(In reply to Mr. Beedell, Roke Julian Lockhart from comment #7)

Created attachment 9486705 [details]
A Screencast Of The JS And Network Consoles

#c6

No; no difference to #c3. Does your question differ to what you requested in #c2?

Sorry. Missed that in your first screencast. I am not sure then why this is not working for you and works for others.

Comment 9

[:glob ✱]

As this is working for everyone else, this has to be something local to your setup, and without any error messages or even activity from your browser it's impossible for us to diagnose your issue further, sorry.

In case it helps - clicking reply should issue an XHR request to fetch the comment text for the relevant bug; eg: https://bugzilla.mozilla.org/rest/bug/comment/17482851?include_fields=text&Bugzilla_api_token=[redacted].

Comment 10

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

Attachment: A Screencast Of Chrome Canary And Fenix Nightly On AOSP

As this is working for everyone else, this has to be something local to your setup.

I would agree with you, except that this doesn't merely occur on one new profile on my desktop OS's Firefox installation. It occurs on:

1. ...a new profile on the Chrome Canary RPM on Fedora 42's KDE Workstation Edition;

2. ...my current profile on the Fenix Nightly APK on my device OEM's stock AOSP distribution:

   I've attached a screencast to demonstrate this.

3. ...a new installation of the Chrome Canary APK on my device OEM's stock AOSP distribution:

   I've attached a screencast to demonstrate this.

If you'd like, I can give you the gold standard of it reproducing on iOS too, if that'll convince you.

I don't even have a modified gateway (nor APs) – all run the stock/OEM firmware, with their providers’ DNS IPs.

Comment 11

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

In case it helps - clicking reply should issue an XHR request to fetch the comment text for the relevant bug; eg: https://bugzilla.mozilla.org/rest/bug/comment/17482851?include_fields=text&Bugzilla_api_token=[redacted].

Thanks. I'm really out of my depth with this, but I managed to find bug_modal.js in the Debugger and enable the "Log" option adjacent to "Event Listener Breakpoints", yet still don't see anything. I tried to capture every exception possible, but none fire. There's no XHR whatsoever.

Were I to, I presume I would remove the:

1. CSP nonce inside content-security-policy-report-only,

2. COOKIE key and its values, and

3. URIs' API keys?

Anyway, is there no possibility that something could be misconfigured for my account server-side...?

Comment 12

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

Attachment: A Screenshot That Accompanies Comment 11

Comment 13

[:glob ✱]

If you'd like, I can give you the gold standard of it reproducing on iOS too, if that'll convince you.

I'm convinced that you are hitting this bug.

I don't understand it either, and I'm at a loss as what the next steps could be. However, as per comment 9 impossible for us to diagnose your issue further.

Anyway, is there no possibility that something could be misconfigured for my account server-side...?

No, this isn't possible.

Comment 14

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

No, this isn't possible.

...Would creating another account to see whether it's usable when authenticated as that confirm it's not on my end, though? Were the button usable, when authenticated with another account, that would appear to indicate that the sole possible cause would be that it's somehow account-specific.

Comment 15

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

Attachment: A Screencast Demonstrating That The Problem Is Account-Specific

Were the button usable, when authenticated with another account

It works on a new account, so it's account-specific!

Comment 16

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

Attachment: A Screenshot Of The Button Operating Correctly Via The Previously-Affected Account

(In reply to Mr. Beedell, Roke Julian Lockhart from comment #15)

Created attachment 9487045 [details]
A Screencast Demonstrating That The Problem Is Account-Specific

Were the button usable, when authenticated with another account

It works on a new account, so it's account-specific!

...it works for me, now, even on my default account in my normal profile. What the heck? I suppose WORKSFORME is correct, ironically. This is the weirdest bug I've ever come across, but at least it's finally resolved.

Comment 17

[Mr. Beedell, Roke Julian Lockhart (RJLB)]

Most incredibly, I can confirm that it now works on my mobile devices (etcetera) too, despite me not having re-authenticated on them.

Consequently, the conclusion must be that this was server-side, affected solely one account, and was remediated when I authenticated with another. In retrospect, I should have evaluated whether it was LAN-specific, or occurred on a cellular WAN, too.