Supermium

[NotHereToPlayGames]

I'm not finding accurate info regarding the "19.0.0".  That IS what OFFICIAL Chrome also shows on my Win10.

But "per spec", anything over 13.0 is supposed to indicate Win11.

[Dave-H]

22 minutes ago, NotHereToPlayGames said:

For what it's worth, Supermium showing a BITNESS of 32 is a DEAD GIVEAWAY.

Chrome has not released a 32bit version since 2018 with v69.

At least by default, the user had to jump through some major hoops to obtain a 32bit version.

Well I guess the '32' gives away that it isn't official Google Chrome in that case but a fork which still provides 32-bit versions.
I think the old-school UA string has always said 'x64' in recent times, 360Chrome was/is the same. 
The discrepancy doesn't seem to be causing any problems with Cloudflare-protected sites as yet, at least none of the ones which I use.
It's not pretending to be a higher version of Chromium now at least, which it used to with earlier versions.
Does Cloudflare even look at the old legacy UA string now?

[Dave-H]

26 minutes ago, NotHereToPlayGames said:

I'm not finding accurate info regarding the "19.0.0".  That IS what OFFICIAL Chrome also shows on my Win10.

But "per spec", anything over 13.0 is supposed to indicate Win11.

Now that would be a bit of a giveaway, as there has never been a 32-bit version of Windows 11 of course!

I suppose you could legitimately be running a 32-bit browser on Windows 11, but why would anyone do that?

[NotHereToPlayGames]

27 minutes ago, Dave-H said:

but why would anyone do that?

When I was on XP x64, I only ran 32bit browsers for purposes of lower RAM consumption.

But nowadays, with the improved methods of memory management, that doesn't work as well as it used to.

[Dave-H]

My main bugbear with Supermium on XP is that tabs often fall over with out-of-memory errors on sites like Instagram and Facebook.
I gather this is because 32-bit processes can each only access a maximum of 4GB of RAM (sometimes even less depending on the system configuration).
My understanding is that nothing can be done to work around this, even patching XP to access all of my 16GB of RAM makes no difference because the RAM limit for each process still exists.
It means I can potentially open a lot more tabs, but each tab will still crash if it goes over the RAM usage limit.
That's the main reason I would never use a 32-bit browser from choice.

[user57]

it seems these are somewhat all called somewhat "user agent (UA)"
for the normal UA a function is offered by the browser in this case supermium

then somewhat it either talks about java offering these values or HTTP
both are controlled by the browser
either way it is very likely they access the same data just with a different request form  (that has to be in the browser)

why would there be a problem to find where these strings maybe values are stored ?

they dont look very different to me:
https://browserleaks.com/client-hints#client-hints-description

then you already find a description (and unlikely a plugin you have the source-code - something far beyond a plugin)

https://developer.mozilla.org/en-US/docs/Web/API/NavigatorUAData/getHighEntropyValues

navigator.userAgentData .getHighEntropyValues

([
"architecture",
"model",
"platformVersion",
fullVersionList",
])
then somewhat "Low entropy hints"
https://developer.mozilla.org/en-US/docs/Web/HTTP/Guides/Client_hints

they are part of HTTP thats a old protocol for webbrowsers

microsoft offers this code 

in short:
userAgentData.getHighEntropyValues(["platformVersion"]
then:
majorPlatformVersion = parseInt(ua.platformVersion.split('.')[0]);   (ua probaly mean user agent - while this is going into majorplattformversion) (this makes it move to majorplattformversion)
then:
if (majorPlatformVersion >= 13)

https://learn.microsoft.com/en-us/microsoft-edge/web-platform/how-to-detect-win11

"User-Agent Client Hints (recommended) - Sec-CH-UA HTTPS header navigator.userAgentData JavaScript method"

"User-Agent string (legacy) - User-Agent HTTPS header navigator.userAgent JavaScript method"

so the 2 java methods are: navigator.userAgent and navigator.userAgentData

Edited Sunday at 10:17 PM by user57
typo error

[j7n]

I only use 32-bit Supermium and haven't experienced any significant problems on normal sites, even the heavy app ones like Discord or YouTube. The multi-process nature allows the browser to use all available memory. On XP you can "only" access about 1.5 gigs per process. The site I ran into problems tried to load ffmpeg webassembly for converting media files in the browser. The native exe is just over 100 MB. But the site used over 3 GB of RAM after switching to the 64-bit version. That is insane.

[Dave-H]

I've never had any issues with YouTube either. I don't use Discord.

Facebook and Instagram are the main problem, the tab always crashes after doing any prolonged scrolling on either of them. If I look at the Windows Task Manager, I can see the available RAM disappearing before my eyes!
😯

[NotHereToPlayGames]

I don't have a Facebook account (*zero* "social media" for me, can you imagine my OT-ish-ness on any of those types of sites, lol).

 

BUT...  Sounds to me like there is an *EASY* solution  --  convert Facebook's "infinite scroll" to a PAGED VIEW instead.

This is sctrictly AI-Generated, but using AI has always been a good STARTING POINT for me, but will require a Facebook account to test and adjust accordingly:

	// ==UserScript==
// @name         Facebook Scroll to Pages
// @namespace    https://example.com/
// @version      1.0
// @description  Replace Facebook infinite scroll with manual page navigation
// @match        https://www.facebook.com/*
// @grant        none
// ==/UserScript==
	(function () {
    'use strict';
	    // Configuration
    const POSTS_PER_PAGE = 10; // Number of posts per page
	    let currentPage = 1;
    let posts = [];
	    // Utility: Create navigation buttons
    function createNavButtons() {
        const nav = document.createElement('div');
        nav.style.position = 'fixed';
        nav.style.bottom = '10px';
        nav.style.right = '10px';
        nav.style.zIndex = '9999';
        nav.style.background = 'white';
        nav.style.padding = '8px';
        nav.style.border = '1px solid #ccc';
        nav.style.borderRadius = '5px';
        nav.style.fontSize = '14px';
	        const prevBtn = document.createElement('button');
        prevBtn.textContent = 'Previous';
        prevBtn.disabled = true;
        prevBtn.onclick = () => changePage(currentPage - 1);
	        const nextBtn = document.createElement('button');
        nextBtn.textContent = 'Next';
        nextBtn.style.marginLeft = '5px';
        nextBtn.onclick = () => changePage(currentPage + 1);
	        nav.appendChild(prevBtn);
        nav.appendChild(nextBtn);
        document.body.appendChild(nav);
	        return { prevBtn, nextBtn };
    }
	    // Change page
    function changePage(page) {
        if (page < 1) return;
        const totalPages = Math.ceil(posts.length / POSTS_PER_PAGE);
        if (page > totalPages) return;
	        currentPage = page;
        renderPage();
	        // Update button states
        navButtons.prevBtn.disabled = currentPage === 1;
        navButtons.nextBtn.disabled = currentPage === totalPages;
    }
	    // Render posts for current page
    function renderPage() {
        posts.forEach((post, index) => {
            const start = (currentPage - 1) * POSTS_PER_PAGE;
            const end = start + POSTS_PER_PAGE;
            post.style.display = (index >= start && index < end) ? '' : 'none';
        });
    }
	    // Stop infinite scroll
    function disableInfiniteScroll() {
        window.onscroll = null;
        document.addEventListener('scroll', e => e.stopImmediatePropagation(), true);
    }
	    // Initialize script
    function init() {
        disableInfiniteScroll();
	        // Collect posts
        posts = Array.from(document.querySelectorAll('[role="article"]'));
        if (posts.length === 0) return;
	        renderPage();
    }
	    const navButtons = createNavButtons();
	    // Wait for posts to load
    const observer = new MutationObserver(() => {
        const newPosts = Array.from(document.querySelectorAll('[role="article"]'));
        if (newPosts.length !== posts.length) {
            posts = newPosts;
            renderPage();
        }
    });
	    observer.observe(document.body, { childList: true, subtree: true });
	    // Run after initial load
    window.addEventListener('load', init);
})();

[Dave-H]

Thanks!
I can't try that out at the moment as I'm away from home for a few days, but I will certainly try it out on my desktop machine when I get back home!

[j7n]

I am currently limited to using Server 2008 R2, and Supermium Version 132.0.6834.224 (Official Build) (32-bit). I haven't updated because other than the ffmpeg issue, it works great. Don't fix what isn't broken. There might be a glitch where it uses more memory on WinXP.

After scrolling Facebook's main feed for a long time, I got the commit charge to about 1 GB and it would fluctuate around that number. I have 8 GB of memory. This was more continuous scrolling that I could ever read. I wouldn't keep the tab open continuously after finishing with the task.

https://i.imgur.com/peXNaAq.png