Firefox Community Forum

On our web site we have a page that includes mp3 files, which play okay on Firefox; but on another page on the site with mp3 files, Firefox (including different versions on different computers) produces the error message, No video with supported format and MIME type found." FWIW, other browsers like Safari (Mac), Chrome (Mac and Win), and Edge (Win) all play the problematic files without a hitch.

Is there a fix for this problem that can be attempted by the non-tech types?

I noticed today that there's a small blue cube appears at the tip of the mousecursor inside a firefox window , I guess it's some keyboard shortcut I accidentally activated but can't find any info on where to disable

I have a website that I go to frequently that does not require a username, only a password. However, on Firefox it insists on a username, and won't let me enter the site without one. The website owner simply does not require a username. I just tried to check into another site and it insisted my password was wrong. I had just changed it and knew it to be correct in format. This relay system is screwing up my browsing and I want it off my machine. Help me!

This is a bug report & fix I found.

Firefox x64 DE 130.0b1 (however, it is not specific to this version of FF.) AMD 3900X CPU (all hardware clocks default unless otherwise noted) AMD RX 6900XT Corsair HX-850 PSU Latest device and chipset drivers from AMD and the respective manufacturers, from their sites Gigabyte Aorus Master motherboard (rev 1) 32GB DDR4 system ram (4x8) stock clocked at 3733Mhz, default timings Windows 11 Build 22635.4005 (Insider's beta-channel build--problem and solution are not build specific, I found)

Problem description:

• Two-three months ago, my previously rock stable system began regular crashing events, every 2-4 days average, sometimes twice a day, sometimes once every 4 days.

• Kernel Error 41, and I spent inordinate amounts of time trying to trace the problem as you might imagine from PSU to device drivers, you name it.

• Finally dawned on me that the hard crashes did not occur when playing a game or when punishing the CPU in some computational tasks--nope, I realized that every time the system crashed I was online running Firefox.

• I got it in my head (through some reasoning I don't even recall) that possibly the CPU stop error that triggered the critical Kernel error and crashed the system was due to a multi-threading bug in my trusty browser!

• So, I checked out Mozilla bug reports and stumbled upon this Windows environment variable, MOZ_FORCE_DISABLE_E10S 1, which I decided to try. It has now been two solid weeks without another episode--not even 1--and while I am not whooping and cheering wildly, quite yet, I do believe the problem boils down to some sort of obscure Ff multi-threading bug! Something was locking up process threads between cores! So, I thought I'd report it! If this gets ID'ed and fixed I'll be happy to try it on at some point...;)

Contents: Conditions, Problem, Observations, Motivation

Conditions

System: Linux Fedora fc20.x86_64, CPU i520M, 3.66 GiB RAM on DELL E5410

Libraries have been updated from rpm packages of Fedora distribution fc21.x86_64 to satisfy the requirements for Firefox 128 and now following versions of relevant libraries and packages are installed:

X.Org 1.14.4, libstdc++ 4.8.3, glibc 2.18, libglib 2.42.1, GTK+ 3.14.5

NetworkManager-0.9.9.0, dbus-1.6.30, GNOME 3.10.0, PulseAudio 5.0

Updating to newer version of Fedora is not possible option for me now.

Firefox 128.0.3 tarball has been unpacked to /usr/local/lib64/firefox128

Executable bash script, /usr/local/bin/firefox128 has been used to launch Firefox 128, it contains two lines

export LD_LIBRARY_PATH=/usr/local/lib64/firefox128:$LD_LIBRARY_PATH /usr/local/lib64/firefox128/firefox $1

Existing file $HOME/.mozilla has been temporarily renamed to avoid hardly possible conflicts while testing the new version of Firefox. (Later I have found that this is not necessary because new installs

of Firefox do not conflict with the previous).

Problem

Execution of the bash script, firefox128 from gnome-terminal command line started Firefox 128 but without GUI. Only the small white interval at left upper corner of the desktop appeared and it could be extended by using mouse to empty (white) window created by Gnome and titled, Firefox.

Observations

After Firefox 128.0.3 started the output of the command, ps -flu username revealed,

      /usr/local/lib64/firefox128/firefox
      /usr/local/lib64/firefox128/firefox-bin \
           -contentproc -parentBuildID 20240725162350 \
           -prefsLen 21615 -prefMapSize 245327 \
           -appDir /usr/local/lib64/firefox128/browser \
                   {d54171b1-8a79-44b3-a271-db4e09db2365} 3715 true socket

Two parameters preceding `true socket` vary from launch to launch.

Executions of, firefox128 --help

              firefox128 --full-version

display the expected output

In the way of successive approximations to highest Firefox version that starts with GUI in my system I have found this is Firefox 95.0.2

The binaries firefox in Firefox versions from 96.0 to 99.0.1 do not launch firefox-bin and the application hangs requiring explicit kill

Firefox versions 100.0.2, 101.0, 106.0, 106.0.5, 117.0, 117.0.1, 123.0.1, 128.0 I have tested like Firefox 128.0.3 behave in the same way as Firefox 128.0.3 with the same effects.

Output of the command, ps -flu username obtained when Firefox 95.0.2 is running with GUI in my system differs specifically from outputs of this command for Firefox versions above 100 Last argument for firefox-bin in those Firefox versions i.e. 100.0.2 and newer is `socket` whereas in Firefox 95.0.2 the last argument is `tab` as in the case,

       /usr/local/lib64/firefox95/firefox-bin \
                    -contentproc -childID 2 -isForBrowser \
                    -prefsLen 4814 -prefMapSize 241644 -jsInitLen 278884 \
                    -parentBuildID 20211218203254 \
                    -appDir /usr/local/lib64/firefox95/browser 9563 true tab

       /usr/local/lib64/firefox95/firefox-bin \
                    -contentproc -childID 6 -isForBrowser \
                    -prefsLen 5627 -prefMapSize 241644 -jsInitLen 278884 \
                    -parentBuildID 20211218203254 \
                    -appDir /usr/local/lib64/firefox95/browser 9563 true tab

       /usr/local/lib64/firefox95/firefox-bin \
                    -contentproc -childID 7 -isForBrowser -prefsLen 5627 \
                    -prefMapSize 241644 -jsInitLen 278884 \
                    -parentBuildID 20211218203254 \
                    -appDir /usr/local/lib64/firefox95/browser 9563 true tab

During executing both the bash scripts firefox128 and firefox95, launching the respective Firefox versions, Firefox writes messages to stdout (gnome-display) but the outputs do not differ qualitatively.

For firefox128 the only intriguing mesasage is the last one that appears while closing Firefox 128, [GFX1-]: RenderCompositorSWGL failed mapping default framebuffer, no dt

These messages not differing qualitatively for Firefox 95.0.2 and Firefox 128 can be summarized as follws (firefox:5370): Gtk-CRITICAL **: gtk_widget_get_style_context: assertion 'GTK_IS_WIDGET (widget)' failed gtk_style_context_get_state: assertion 'GTK_IS_STYLE_CONTEXT (context)' failed gtk_style_context_get_property: assertion 'GTK_IS_STYLE_CONTEXT (context)' failed gtk_style_context_get_background_color: assertion 'GTK_IS_STYLE_CONTEXT (context)' failed gtk_style_context_get_color: assertion 'GTK_IS_STYLE_CONTEXT (context)' failed

(firefox:5370): GLib-GObject-CRITICAL **: \ g_value_unset: assertion 'G_IS_VALUE (value)' failed

(firefox:2901): GLib-GObject-CRITICAL **: \ g_object_ref: assertion 'object->ref_count > 0' failed and g_object_unref: assertion 'object->ref_count > 0' failed

[GFX1-]: glxtest: libEGL missing methods

For Firefox 128 they are preceded by warnings like, [Parent 4185, Main Thread] WARNING: \ g_value_unset: assertion 'G_IS_VALUE (value)' failed: \ 'glib warning', file /builds/worker/checkouts/gecko/toolkit/xre/nsSigHandlers.cpp:187

Moreover during starting Firefox in versions 96.0 - 117.0.1 there is the message,

                    os_mmap failed.: Cannot allocate memory 

This message is absent either in Firefox 95.0.2 and Firefox 128.0, 128.0.3

Motivation

The effect just described looks very strange but may be common to older versions of Linux distributions still being used by number of people with various motivations.

Various institutions, particularly banks, require "newest" versions of a browser and I am afraid that Firefox 95.0.2 working in my system will be very soon considered outdated that will result in lack of electronic access to bank account.

Therefore solving the problem just presented appears to be important.

Thank you in advance

On that piece of crap Windows 11 I keep getting the "Can't get your tabs back". I get it EVERY TIME I open firefox. I have uninstalled (several times), deleted directories and tried the "Refresh Firefox" and nothing works. I do not have this issue on any of my Linux computers. I only have a few programs left that I need Windows for so I guess I need to look at dual booting this computer.

Is that ANY way to do a clean install of firefox? I keep getting the you have installed firefox before message when I try to reinstall. I really really really don't want to have to do a clean install of windows to try to fix this.

I have been a devoted Firefox users for many years. Just in the past week or so I have started having multiple tabs crash daily and now am getting 1 or 2 blue screens a day. I run on Windows 11 Pro. What can I do to help this issue?

On Firefox for Windows the touch and pen logic is a mess. When I hover over something with the pen or scroll using touch, Firefox seems to make both elements hover and triggers even JavaScript related code at very high rate. These 2 elements are where the real mouse/cursor is on the screen and where the currently touching or hovering over element is. For example rest your mouse on the YouTube HTML5 player, then use a pen to hover over the browsers close button or even just something else on the web page like the like button. This will ramp up the CPU and make both the player and the like button your are hovering flicker between ::hover and ::normal. This will also call lot's of instances of these related events in the website's JavaScript. Another semi-related issue is when using the F12 developer tools in "Device Simulation" mode, It's impossible to for some reason because Firefox fakes real touch injections into the OS (and they show up as if I touched the screen by the OS) which cancels the precision touchpad input movement until I lift my finger from the touchpad. This is specifically a "Windows Precision Touchpad" issue, and There is no known fix to this apart from not using a such a touchpad... On old 10+ year old laptops with PS/2 or proprietary Touchpads that report themselves as a normal Mouse this issue won't occour but Firefox will still input system-wide touch input. As a developer myself this seems very silly because you own the source code and could solve this way more elegantly than asking the OS to input Touch events... Let me know if you need anything else, and point me out where the source for the input (specifically for touch and pen) live so I can take a look at it and make some improvements. Thanks

My Windows 10 system is very slow on startup. Applications are taking a long time to even respond to clicking icons, but eventually start. In Task Manager, I see that several instances of Firefox are running at startup, although it doesn't appear in the startup list. When I end these Firefox processes, the system immediately resumes normal operating speed, and things that are waiting to open appear (Speccy, CC, AV's, mail client,etc.). How do I stop Firefox from running in the background without permission at startup?

Dont know how it happen but "https://ww25.waterdropswiter.xyz" has taken over comes up everywhere when tying in web adress and has give it self free access to everything

For about 2 weeks now my Firefox Desktop v. 128.0.3 (Mozilla-Flatpak - 1.0) freezes within 30 mins of use.

No good waiting, nothing responds except the exit button.

Clearly, there's a flaw in a recent update.

I'm on Ubuntu 22.04.4 LTS and my Firefox is from Flatpak.

While writing this I got a Flatpak update for Firefox and Firefox.Locale - whatever that is.

So I'll park my question till I close and reopen Firefox - hopefully with a release from the big freeze-up.

Since the auto-update to version 128.0.3 (64-bit) two weeks ago, my Firefox crashes as soon as I switch over to a different programme. I am on Windows 10. Whenever I try to re-open the programme, I get the error message that Firefox is already running and if I want to restart or quit. Can you please deploy a fix for this? Thank you.

After your an update some weeks ago, I have not been able to access my bank account using Firefox. Got some advice from support from the bank, but since I have an US version, the labels to be accessed differed, and were not compatible with the advice. Presently using M Edge, which I of several reasons don`t want to. Any suggestions?

When I start typing 'subscr..' into the URL bar, I'd like it to autocomplete / select 'Subscriptions' from my bookmarks.

However, instead the default item is a Google search for 'subscr' as shown in the image.

Is there a way to skip this and make it select the top suggestion instead, much like Chrome does? (I've jumped ship from Chrome)