Firefox regression in RHEL but not in RL (9.2)

I wonder why the new Firefox package is working in RL while a RHEL installation has some problems:

Just out of curiosity. I would expect some kind of 1:1 bug compat. Any clue?

RHEL has the same problem in Thunderbird

Curious indeed. Can you doubly check that you are running the version that was reported to have the issue?

Also, have you tried https:// rather than http:// ?

Sure. Its a RL workstation as a daily driver with 102.14.0esr as browser (https/http). Everything works …

OK, I confirm that I cannot reproduce the problem reported for RHEL-9 firefox.

This is both good and bad. Good because Rocky users do not suffer from the issue, bad because Rocky is not bug-for-bug with RHEL…

I am using RHEL 9.2. When Firefox ESR updated itself recently, I was suddenly unable to load any website. So I installed the Flatpak version of Firefox so that I would have at least something. Yesterday, I reinstalled the Red Hat Firefox. They still have not patched the problem. I have since installed the Rocky version of Firefox ESR on my Red Hat.

I like Red Hat, but I think it’s pathetic that they haven’t yet patched this bug.

Actually, I think Rocky did everything right. I assume that RH compiled it (firefox and thunderbird) against the wrong (newer) NSS in their internal brew/koji instances. Currently EL8 and EL9 has nss-3.79.0 and RH are preparing a rebase for both to 3.90.0. I made some test here that confirm this carelessness of RH. RH’s new update restores the 1:1 bug compat again :slight_smile:

They just released a patched version today. From the changelog:

* Wed Aug 09 2023 Jan Horak <> - 102.14.0-2
- Rebuild due to rhbz#2228948

Precisely. Because “Rocky did everything right”, they could not reproduce the serious bug that RH introduced. :wink:

1 Like

Are you sure you are using the equivalent sources in your repos?

Maybe you have added repos in Rocky which you don’t use in RHEL, or on one OS you are using the ESR version, & on the other you are using the normal version.

Delicious about the definition of “right” and “bug”.
It is “right” to build a functional program, but it is also “right” to build practically identical to RHEL version.
Rocky had a “bug” that it did not have the same “bug” as the RHEL version, but if had not have that bug, then the program would have had a bug nevertheless.
While it was “bad” to make a “mistake”, the result was a lesser of two evils.

You might try using Firefox developer edition - I use it on 8.8 but imagine it should work on 9.x and it shouldn’t affect the system version if you follow the instructions at FF Dev as the user in the thread mentions.

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.