Future Of Firefox Quantum On Windows 7

[Dixel]

On 7/9/2023 at 12:12 PM, Sergiaws said:

Chrome introduced a special javascript api to recognice the operating system used. Will it be added to FF?

This API is there for ages, this is how both Chrome and FF generate their real UA.

[yoltboy01]

12 hours ago, Dixel said:

Basically, yes. 116 beta requires Windows 10

116 works on Windows 7

117 alpha builds works up to 2023-07-18

118 does not work

 

I'm testing new nightly builds on 8.1 every day and it still seems to work. Let's see for how long. Would be nice to get another ESR working on 8.1. This would indicate that support would extend for another year.

Edited August 17, 2023 by yoltboy01

[D.Draker]

13 hours ago, yoltboy01 said:

116 works on Windows 7

Officially it doesn't, I think the question was about the official support. For 116 to start, it needs to be done like this. I also saw some reports of video decoding/playback troubles.

The installer of 116 is blocked for anything below Win 10.

About the video and text problems, I can't confirm or deny, I don't use Firefox myself.

https://www.reddit.com/r/windows7/comments/15i9dfd/firefox_116_on_windows_7/

[veso266]

OK, can someone tell me why do they have to introduce the newer win10 only apis

 

What do theese apis solve?

Or is it just because the said they will drop support (because this was quite quickly, only 2 versions and it breaks

If a new compiler cannot compile for windows 7, are there particular reasons they choose it or its just because its a shiny new thing

 

What is wrong with a warning that says: If you are still using windows 7 we sadly cannot provide you support anymore, so if something breaks in the future you will have to ask a fellow windows 7 user for help

 

I know that the best way forward is to use kernel extender, but right now, it seams its easier to upgrade to Vista, because the only kernel extender: https://github.com/vxiiduu/VxKex that we have died, I hope the situation does improve

Also can't backwards compatibility issue be somehow solved with more modularization in windows itself (idk, each api would be a plugin that could be loaded by a kernel so when new api is created, it could be bundled in new version of windows, while older windows versions could just install it manualy (or in a API pack or something like that)

I understand drivers can have problems with backwards compatibility because they are tied to the kernel so much but why do user programs need to suffer from the same problem

Let the normal users that want shiny new shell and new features that look different but do the same thing as the old ones upgrade to newer versions of windows and leave us that want to use a particular windows version to do so, without forcing us to upgrade to do the same thing that worked perfecly fine on older os

 

Because lets be real here, each of us that use particular version of windows have a good reason to do so

For instance I use windows 7 because I like Aero and Microsoft Virtual PC 2007 does not work on anything newer and 3 drivers for 3 pieces of hardware (that the company went under) do not work on anything newer

And I don't want to have a separate computer (because lets be honest, at one point you will need a separate computer to run a particular OS, VM can only go so far) because the modern os does not support a driver anymore and I cannot use some other software on older os

 

 

 

Edited September 5, 2023 by veso266

[vinifera]

i think they do it mostly to track current versions of OS
if you look at it, win7 was released well over 12 years ago

what is good about the win7 is that it can be still supported with apps
but people choose not to (probably some compiler BS)

i mean there is a reason why
7 was 6.1
8 was 6.2
10 was 6.3
dunno about 11 (is it 6.4 ???)

>>> BACKWARD COMPATIBILITY !!!

[D.Draker]

3 hours ago, vinifera said:

i think they do it mostly to track current versions of OS
if you look at it, win7 was released well over 12 years ago

i mean there is a reason why
7 was 6.1
8 was 6.2
10 was 6.3
dunno about 11 (is it 6.4 ???)

>>> BACKWARD COMPATIBILITY !!!

win7 released 14 years ago

8.1 was 6.3

10 was 10.0  (Windows NT 10.0;) (still is)

[Saxon]

6 hours ago, vinifera said:

dunno about 11 (is it 6.4 ???)

6 hours ago, vinifera said:

10 was 6.3

 

Both, Windows 10 and 11, are NT 10.0. Weird, right?

[ED_Sln]

4 hours ago, D.Draker said:

10 was 10.0  (Windows NT 10.0;) (still is)

It's a marketing name to make users think the kernel is new, in fact it's still 6.x. The kernel is certainly updated, but in fact the difference between Win7 and Win10 is the same as between Vista and Win8.1.

[D.Draker]

3 hours ago, ED_Sln said:

It's a marketing name to make users think the kernel is new, in fact it's still 6.x. The kernel is certainly updated, but in fact the difference between Win7 and Win10 is the same as between Vista and Win8.1.

Sounds very much like a theory, I'm eager to know, where and how could you (or anyone else) make a comparison of these completely closed source products.

[broken120x120]

5 hours ago, D.Draker said:

Sounds very much like a theory, I'm eager to know, where and how could you (or anyone else) make a comparison of these completely closed source products.

Beta versions of Windows 10 were originally known as Windows NT 6.4. It isn't exactly a stretch to assume they bumped the kernel version up to 10.0 to match its name.

[UCyborg]

On 9/5/2023 at 12:36 PM, veso266 said:

OK, can someone tell me why do they have to introduce the newer win10 only apis

 

What do theese apis solve?

Can't be more technically specific here since I'm not a software engineer, but generally they say to accomplish one thing better than was possible in previous versions lacking the new API. Using new API directly breaks compatibility with old OS right away, which implies you need a fallback, so you have to be explicit about implementing fallback, which in turn will be practically used only on an old OS and if they don't test on old OS, they won't leave untested code inside (maintenance burden) and call it unsupported. It gets more convoluted quickly, as if it wasn't enough as it is.

I also got impression from reading certain discussions in the past that multi-process architecture is more prone to security bugs, so I think they'll want to use every last trick OS provides to tighten it. I'm sure I don't need to emphasize how security is being shouted about from every rooftop.

I imagine API packs for old OS would complicate matters in similar way. Then you add the into consideration how many users still use the old OS and there's even less inclination to bother supporting old ways of doing things. I'm sure some developers will also relate to the thoughts of the police officer in the following video.

[veso266]

On 9/6/2023 at 7:20 PM, UCyborg said:

Can't be more technically specific here since I'm not a software engineer, but generally they say to accomplish one thing better than was possible in previous versions lacking the new API. Using new API directly breaks compatibility with old OS right away, which implies you need a fallback, so you have to be explicit about implementing fallback, which in turn will be practically used only on an old OS and if they don't test on old OS, they won't leave untested code inside (maintenance burden) and call it unsupported. It gets more convoluted quickly, as if it wasn't enough as it is.

I also got impression from reading certain discussions in the past that multi-process architecture is more prone to security bugs, so I think they'll want to use every last trick OS provides to tighten it. I'm sure I don't need to emphasize how security is being shouted about from every rooftop.

I imagine API packs for old OS would complicate matters in similar way. Then you add the into consideration how many users still use the old OS and there's even less inclination to bother supporting old ways of doing things. I'm sure some developers will also relate to the thoughts of the police officer in the following video.

I am not a buisness man, but couldn't backwards compatibility (in this case called forward compatibility) be marketed to companies (where upgrading everytime a new things comes is expensive (because you have to test if a new thing will not break an existing thing)

I always thought that backwards compatibility should be top priority when building something (like an os) people build upon

Imagine you have a gun and cannot buy ammo anymore, because the company that made it, said they will stop making it (1000 of guns rendered useless because you cannot use them anymore)

 

Imagine you could upgrade only parts of the OS you actualy need (not everything), lets say you need some new APIs because your programs don't work anymore, and you upgrade just that, (without changing the UI and the rest as well)

Windows backwards compatibility is great for the most part, now it only needs forward compatibiliy

 

I am sure companies and other tech people would upgrade sooner if this would be the case

[AstragonQC]

Hi, am I the only one to have my FxNightly install got automatically updated to a newer build despite it shouldn't be possible?

I mean, I kept my FxN on the v117 July 18 build (as I still use it for some things), and when I wanted to open the browser this morning I noticed it self-updated to a v119 build (2023-09-06) (which obviously doesn't run due to missing DLL entries). So I manually downgraded my install to the July 13 build, and as soon I restarted, the browser immediately started to download and install a newer build (2023-09-12), definitely proving the update ability is not blocked anymore...

I put the policy file to completely block updates but now I'm wondering why (and especially how) it happened, as it never behaved like this during the two past months (It just showed the orange triangle like on my FxDE install)
(Maybe it happened because it was a v117 build and not a v115 one. Or there's something else I didn't understand yet)

[ED_Sln]

21 hours ago, AstragonQC said:

and when I wanted to open the browser this morning I noticed it self-updated to a v119 build

Yes, I have it the same way, updates to 119 which doesn't run.

[Mark-XP]

On 9/13/2023 at 11:39 AM, ED_Sln said:

Yes, I have it the same way, updates to 119 which doesn't run.

I observed this kind of behavior impudence by mozilla already many years ago, leaving add-ons unsupported after an update.

But updating to a version that isn't supported by the used os would lift mozillas improfessionalism and recklessness to an new level imo - is that really true?

And beware: even if you copy a profile (with auto-update disabled) to a profile-folder of another (newer) version of firefox, it will imediately change that pref to enabled. Simply awful.