Jump to content

The UOC and QUOC Patch - Optimize Firefox (and derivatives) for old hardware!


Recommended Posts

On 11/4/2019 at 4:09 AM, looking4awayout said:

EDIT: I have fixed the issue on both the 52 ESR and the NM28 version of the patch. It was caused by this value:

pref("network.http.referer.XOriginPolicy", 2);

Removing this value has restored the user avatar and the feedback count on Ebay. Please re-download and update! :)

EDIT 2: I have implemented multithreaded image decoding but only on the 45 and 52 ESR versions of the UOC Patch. The 38 ESR and NM 28 versions haven't been touched, as they do not support E10S. Ironically, enabling multithreaded image decoding has somewhat sped up the process of loading images on my Tualatin. Never thought multithreading would work on single core systems...

However, all the versions have been modified to reduce the maximum amount of rendered text per page, from 8191 to 4095.

Please update!

Thanks for checking it out. I'll test the newest version this evening. On a somewhat unrelated note, Firefox Quantum appears to be capable of leveraging WebGL on almost any GPU, regardless of  DirectX version support (it uses OpenGL when a DX11 GPU is not present). Under Windows 7, I can achieve silky smooth scrolling, and even 720p video playback with a Pentium 4 670, and a 32MB Radeon 7500. I've tested this on a GeForce 4 Ti 4200 as well, and observed the similar results.

Edited by sparty411
Apparently Quantum leverages OpenGL when a DX11 GPU isn't present
Link to comment
Share on other sites


On 11/6/2019 at 1:55 AM, sparty411 said:

On a somewhat unrelated note, Firefox Quantum

... I think they themselves stopped calling it that way :P ; recently, I had the chance to view the "about" box of latest stable Mozilla Firefox v70.0.1 (in sister's Win7 SP1 laptop ;) ) and was surprised to see it was now referred to as "Firefox Browser":buehehe:; any bet what it'll be called come v85.0 (I guess it'll make it there, as they'll be increasing major version numbers each month soon...) ? :whistle:

Edited by VistaLover
Link to comment
Share on other sites

1 hour ago, VistaLover said:

... I think they themselves stopped calling it that way :P ; recently, I had the chance to view the "about" box of latest stable Mozilla Firefox v70.0.1 (in sister's Win7 SP1 laptop ;) ) and was surprised to see it was now referred to as "Firefox Browser:buehehe:; any bet what it'll be called come v85.0 (I guess it'll make it there, as they'll be increasing major version numbers each month soon...) ? :whistle:

D'oh. I forgot all about the stupid name change :dubbio:

Link to comment
Share on other sites

UPDATE!

I have updated all the versions of the UOC Patch, as I have applied some tricks that further increase hardware acceleration, making scrolling smoother and faster. The tweaks work on every codebase, so please update to the latest version of the UOC Patch.

On the 45 ESR version, support for multithreaded image decoding has been implemented and it seems to bring benefits even on single core machines.

Thank you and happy RDDin'!

Link to comment
Share on other sites

Hi, I know you've said your RDD has a Radeon x1950xt, but do you have an nVidia  card to test on? The latest release of the patch causes horrendous stuttering while scrolling with my GeForce 6600. This can be observed in FF 45.9 ESR, as well as Pale Moon 27.

Link to comment
Share on other sites

4 hours ago, sparty411 said:

Hi, I know you've said your RDD has a Radeon x1950xt, but do you have an nVidia  card to test on? The latest release of the patch causes horrendous stuttering while scrolling with my GeForce 6600. This can be observed in FF 45.9 ESR, as well as Pale Moon 27.

Hello, unfortunately I don't own any of those cards. The ones I own are a GeForce 6800GT and an X1950 Pro. I cannot replicate the issue with those cards.

Do you still have the previous version of the Patch that worked on your system? If so please, send it to me as I do not have it.

Meanwhile, try to delete the entries related to APZ from the Patch. You will find them on top. See if deleting them makes a difference.

EDIT: try to set the values below:

pref("layers.use-deprecated-textures", true);
pref("apz.overscroll.enabled", true);
pref("apz.relative-update.enabled", true);
pref("apz.enlarge_displayport_when_clipped", true);
pref("apz.allow_zooming", true);
pref("apz.cross_slide.enabled", true);
pref("gl.multithreaded", true);
pref("privacy.donottrackheader.enabled", true);
pref("layout.css.prefixes.transforms", false);
pref("prompts.tab_modal.enabled", false);
pref("webgl.shader_validator", true);
pref("webgl.prefer-16bpp", true);

To False. Do it one by one, in order to find out the offending value. Let me know the outcomes. If you find the offending value, I will remove it from the patch.

Edited by looking4awayout
Link to comment
Share on other sites

URGENT UPDATE!

I have thoroughly inspected the latest update of the UOC Patch and the Enforcer, and I have found several double entries that might have been the culprits behind @sparty411's issue.

Please, redownload and update! All the versions of the UOC Patch and both versions of the Enforcer have been updated.

EDIT: the Patches have been updated again.

Edited by looking4awayout
Link to comment
Share on other sites

On 11/4/2019 at 1:08 AM, roytam1 said:

hope this rev can help (preventing them to be enabled by users): https://github.com/roytam1/UXP/commit/e578dccb6d69c327f5ad334064320ba25b24f5de

On 11/4/2019 at 6:18 PM, Mathwiz said:

When I get a chance, I'll test it. If it works, @looking4awayout won't need to maintain a separate NM28 version of his patch (NM28 will simply ignore the preferences that enable e10s).

Yes, I tested it with the latest PM 28, and it works! E10s remains disabled even with the "standard" FF 52 version of the patch. Thanks!

So, the separate NM28 version of the patch is no longer needed.

Link to comment
Share on other sites

2 hours ago, looking4awayout said:

@Mathwiz please don't forget to update to the latest version of the UOC Patch and the UOC Enforcer if you still haven't.

If I could figure out how to automate downloads from Mediafire I'd just put the code to do it in my batch file. Then every time I installed a new version it'd get the new UOC Patch!

Link to comment
Share on other sites

44 minutes ago, Mathwiz said:

If I could figure out how to automate downloads from Mediafire I'd just put the code to do it in my batch file. Then every time I installed a new version it'd get the new UOC Patch!

If/When @looking4awayout's build process slows down a bit, I can happily host it on my site. I can do it now but it wont be easy keeping up with his new builds every five minutes. :P

Link to comment
Share on other sites

NEWS!

After some testing on different platforms, I have noticed that SeaMonkey, at least version 2.48, has a broken implementation of Asynchronous Pan and Zoom (APZ), so if you use the 45 ESR version of the UOC Patch, previously recommended for that browser, the scrollbars will be unusable. It is not a fault of the Patch, but a bug in the APZ implementation that has never been fixed.

So, in order to speed up and optimize SeaMonkey, you will have to use the 52 ESR version of the UOC Patch, alongside the 45/52 ESR version of the UOC Enforcer. In that way, the browser will be optimized and run as fast as it allows to.

Link to comment
Share on other sites

Warning to users of Adobe's "Flash" plug-in: I have discovered that the 52 ESR version of the UOC Patch sets two preferences that interfere with Flash:

  1. privacy.resistFingerprinting
  2. dom.ipc.plugins.asyncdrawing.enabled

#1 is more serious. For some reason, it prevents Flash from working at all. Luckily, it also has the easiest workaround: set it back to false, or remove the line enabling it from the UOC_Patch_52.js file; instead, use an extension (I use "Canvas Defender" but there are several others) to resist canvas fingerprinting.

#2 may, or may not, affect your system. To test, first disable the above preference, then go to https://get.adobe.com/flashplayer/about/. If both the Flash animation and your currently installed Flash version appear, you're fine and can leave this preference set to true. If the page doesn't work properly, you'll have to set this preference to false also. Unfortunately I know of no workaround to retain its function if you have to turn it off.

Edit: It turns out that privacy.resistFingerprinting implicitly does lots of things "behind the scenes". Some of these changes, such as user agent overrides, could cause Web sites to malfunction or otherwise affect usability. Luckily, most of the changes didn't land until FF 55 (which won't run on XP) or later FF versions (but Serpent 55 may include some of them). Still, I think it would be best to remove this preference from the UOC Patch.

Edit 2: The troublesome privacy.resistFingerprinting also appears in the FF 45 version of the UOC Patch, but that preference doesn't do as much in that version, so it may not cause trouble with Flash there. I haven't tested FF 45 with the UOC Patch and Flash.

Edit 3: Found one more. This one's more obscure, but I'm posting it anyway, just in case anyone runs across it. The UOC Patch sets pref apz.allow_zooming, which conflicts with the Multi-Account Containers add-on. That add-on is the main reason I switched to Serpent 55, so my own customized version of the UOC Patch now omits privacy.resistFingerprinting and apz.allow_zooming, and sets dom.ipc.plugins.asyncdrawing.enabled to false to ensure Flash always works.

Hopefully I now have everything working again!

Edited by Mathwiz
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...