UCyborg

Yes, that would be better for consistency since expecting minified JS and escaped = is taken as literal without the added documented behavior. Now I've put that MIME-List into use and added the following rule: In = TRUE Out = FALSE Key = "Content-Type: Fix MIME types" Match = "(*|(^?))$URL([^/]+*.([a-z0-9]+{2,5}&&$LST(MIME-List))((^?)|\?))" Replace = "\0"

Ooops, that wasn't the case...it just uses update.googleapis.com instead of clients2.google.com. All documentation points to the latter... The way query is sent to that URL is different and response is in JSON rather than XML, which I haven't quite figured out yet. I prevented it from connecting to Google, but haven't got it to use local extension respository. Is it a secret? I certainly can't find anything about it.

uBlock Origin is enough for strong protection against Web tracking message. You're using tricks for randomized fingerprint, aren't you? About Edge and extension policies, with policies in general, some just aren't followed unless you're joined to a domain (or so it's written). I have "Configure extension management settings" policy set containing JSON specifying all extensions with "override_update_url" set to "true" and "update_url" set to locally hosted updates.xml file, along with this trick applied, but it doesn't care. There should be some better way than blocking access to clients2.google.com or side-loading the extensions.

Easiest way is downloading it from https://www.crx4chrome.com/ and you'll need KeePass from https://keepass.info/ along with KeePassRPC plugin. Portable version should work out-of-the-box on XP if you have at least .NET Framework 2.0 installed, best to have 4.0 for the best compatibility though, particularly with 3rd party extensions. This little app is totally unsupported on old Windows and presumably .NET Frameworks, but it's coded in a way that lack of features it can use simply makes them unavailable rather than entire application to not work at all. I don't remember when the installer was upgraded to Windows 7+ version, I can upload the one I made for KeePass version 2.49 a while back if you want the installer, retro-compatible (works on Win98) with translations (at least those 100% completed). The key that unlocks the database may be composed with the combination of a master password, a key file and a tie to Windows user account. I personally only use master password. Login form auto-filling is handled by the browser extension and should work for most sites as long as your password entries have URL field filled. Extension must also authenticate with KeePass to connect to it, you get a prompt to input generated password to authenticate. Then it's authenticated for a year before it expires and you get another prompt then. Of course, you can experiment with security settings to make it a pain for browser extension to connect to KeePass. Naturally, extension can only get data from the database when KeePass is running and the database is opened.

Ohhh, thanks for the hint! Putting application/x-javascript instead of application/javascript works. Both are considered legacy though (https://developer.mozilla.org/en-US/docs/Web/HTTP/Basics_of_HTTP/MIME_types - Legacy JavaScript MIME types): Note: Even though any given user agent may support any or all of these, you should only use text/javascript. It's the only MIME type guaranteed to work now and into the future. Yup, I'm clumsy and don't want to cut myself. Those rules from the starting point config break JS left and right on that site. Sure can, this was the odd one, all because of MIME type.

Haven't fired Edge in a while. Today I noticed it goes through configured proxy to check for extension updates installed from MS Store...but still connects directly for extension updates from Chrome store...I hate auto-updates.

KeePass (2.xx variant) Kee

Hello, have you tried following this guide? https://howtoubuntu.org/how-to-repair-restore-reinstall-grub-2-with-a-ubuntu-live-cd Where do you get stuck?

Of course, I don't expect Chinese actually added this to their own fork. Plus, the parameters probably have to be used from the very beginning, before the user profile is created for the first time. I haven't tested, but seems logical.

It's tied to user account's SID. Guess what happens if you make a new account or reinstall Windows without having noted your old account's SID. Ungoogled Chromium has command line switches to disable this feature.

Admittedly, I mostly compared the look with different ClearType parameters, so didn't try any other smoothing setting there or turned it off, because then everything generally looks much worse to my eyes. Now I notice subtle difference between different ClearType settings (have to compare screenshots...). With Firefox's cairo, the difference seems more apparent. Either way, can't say much about it, Firefox uses Skia or Cairo or D2D (latter only on later OS), Chrome uses Skia and there are probably differences due to how it's used, they could use different versions of the library etc. The site you mentioned, I get less edgy appearance with Skia on Firefox, regardless whether the checkbox is enabled with Standard selected on the dropdown menu while on 360Chrome, it's only less edgy with the checkbox enabled. I did restart browsers this time for the settings to apply.

Fonts always looks the same here in any Chromium browser on XP regardless of OS settings. OS settings only have influence on Firefox browsers when set to use cairo instead of skia for content rendering.

I got around playing with media.ffmpeg.hwacc_type setting mentioned in the release notes, it's supposed to enable HW video decoding through either CUDA and DXVA. Both crash. I wonder if it actually works for anyone.

I'm unable to get the following rules to work, built-in test shows correct output, but in practice, they just won't budge. [Patterns] Name = "Reddit - comments page regex 1" Active = TRUE URL = "www.redditstatic.com/desktop2x/CommentsPage.*.js" Limit = 127300 Match = "/\(\?:reddit\\.com\\/r\\/\)\(\?<subreddit>\[\\w\]\+\)\(\?:\\/comments\\/\)\?\(\?<postId>\[\\w\]\+\)\?/" Replace = "/\(\?:reddit\\.com\\/r\\/\)\(\[\\w\]\+\)\(\?:\\/comments\\/\)\?\(\[\\w\]\+\)\?/" Name = "Reddit - comments page regex 2" Active = TRUE URL = "www.redditstatic.com/desktop2x/CommentsPage.*.js" Limit = 127700 Match = "var s;const{subreddit:o,postId:r}=\(null===\(s=t.match\(i\)\)\|\|void 0===s\?void 0:s.groups\)\|\|{};" Replace = "const s=t.match\(i\),o=null===s\[1\]\?void 0:s\[1\],r=null===s\[2\]\?void 0:s\[2\];" They're supposed to replace: /(?:reddit\.com\/r\/)(?<subreddit>[\w]+)(?:\/comments\/)?(?<postId>[\w]+)?/ /(?:reddit\.com\/r\/)([\w]+)(?:\/comments\/)?([\w]+)?/ And: var s;const{subreddit:o,postId:r}=(null===(s=t.match(i))||void 0===s?void 0:s.groups)||{}; const s=t.match(i),o=null===s[1]?void 0:s[1],r=null===s[2]?void 0:s[2]; 1st line is original, 2nd is replacement. URL to the target file: https://www.redditstatic.com/desktop2x/CommentsPage.c501be48b07ce588906d.js

I meant... Pretty sure I saw it at least one more time in the past as well.

I mean, you keep writing MB where you mean GB. My next computer will probably have 32 GB of RAM and 8 CPU cores. I'll never run out of RAM again! I remember digital life being miserable with just 1 GB of RAM and single-core CPU in 2008.

You want my gas? Well, you can kiss my a**! This satire is 6 years old and sadly more relevant than ever. https://www.bbc.com/news/business-61237519

@NotHereToPlayGames Bruh, my first computer had 16 MB of RAM, in 1997! Anyway, I only trialed Mypal 68 for a short time in single-process mode, before figuring out IgnoreException trick usability for this browser, which I also use now successfully to get around access violation exception popup, but single-process mode was a definite path to (more) instability.

GUI limitation, can't input byte limit greater than 32767 for web page filter. Can be edited directly in CFG. BTW, this topic doesn't strike me as belonging to this sub-forum - it's not a web browser and works on most Windows OS I think, certainly it's not a derivative of other version of the program exclusive to modern OS. Maybe in Web Browsers since it's related to web browsing. @Dave-H, @Tripredacus, @dencorso ?

Hi . VP9 hardware decoding is only available since the seventh generation of PureVideo HD (C). I was actually referring to normal acceleration dealing with web page compositing, because the slowness doesn't necessarily come from inability to decode video on GPU. Obviously, on XP, no browser will even try to query GPU video decoding capabilities because expected APIs are missing. My CPU is doing fine in general decoding 1440p @ 60 FPS VP9 streams. But no browser I'm aware of on any Windows version I can run on my computer is actually able to handle that video resolution / frame rate smoothly without frame drops. Pale Moon on Linux handles it like a champ though (not that anyone here actually cares).

Things crashing left and right with HW accel is not the norm. There is a specific market of graphics cards for professional applications (including CAD), used to be NVIDIA Quadro and ATI FirePro/FireGL brands, now such cards are marketed under different brand. I doubt such market would exist if HW accel did not work, especially considering the high price tag.

I doubt Mypal users with BSOD problems are getting out of their way to enable OpenGL backend for hardware acceleration. Ignoring browser issues and their HW acceleration, which came later and is a bit specific, XP was all about HW accel, from graphics to sound - look at DirectX Diagnostic Tool - Win + R -> dxdiag. At the same time, the OS is old school and there is very little room for mistakes. A hiccup that might just cause a screen flash on Windows 10 may bring Windows XP down. Granted, hiccups are still possible in other areas on newer OS, but graphics drivers got a bit special treatment so not every hiccup crashes the entire OS. And properly working XP installations where everything adds up might be rarer than we think.

@Anbima roytam1's browsers might still be OK if GPU is up for it. And software fallbacks should still be faster, especially compared to EE V13. Just noticed 720p 60 FPS has a noticeable edge on my PC in MyPal 68 while they're only smooth in roytam1's browsers in a small window without GPU acceleration. I don't know what to say about Mypal and BSODs, but a browser shouldn't be able to crash the entire OS. Hard to guess, especially without crash dumps, what's the reason. Could be driver issues, could be unpatched bugs in the OS...I only have XP x64 install here and didn't manage to get the browser to crash the OS.

YouTube strikes me as rather conservative. Still works in Internet Explorer 11, last major update of that browser was in 2014. Of course, the interface is pretty sluggish with that browser. GPU doesn't matter on Chrome on XP as GPU acceleration is not implemented via APIs XP supports. From V11 to V13, it gets progressively slower. Same goes for the upstream Chromium they're based on. Haven't tried 13.5 though. Chrome's software presentation of frames is very slow, the bigger the surface the videos covers, the slower it gets. So even at the same resolution selected in YouTube's interface, there's big difference in running it in a small window vs full screen, particularly if screen is high resolution. MyPal 68 is much better at handling video, got smooth 720p @ 60 FPS playback without frame drops (VP9 codec) without GPU acceleration. So even if it works for someone well somehow on Chrome, more CPU time is spent presenting frames than would be needed if it was better optimized. Of course, if you don't mind the little inconvenience and lack of all YouTube features in one place (the browser), throwing a YouTube link to dedicated media player should be the most efficient as far as efficient hardware resource utilization is concerned.

I have Firefox 52.9 configured very much like Serpent and guess what, this issue does not occur there with same uBlock Origin and settings. Also spoofing the same user agent as on Serpent (Firefox 71) to get modern layout, though Firefox 52.9 does not support WebP images so GOOOOOGLE is missing on the bottom of the page above the links to navigate to next pages. Might be an issue upstream since it happens on official Pale Moon. Or maybe an odd extension issue, who knows. But what kind of issue since there's no error? You can right-click the result link, copy link, paste and go and that also doesn't redirect back to Google, just normally clicking it does. Edit: No problem in SeaMonkey neither (non-UXP one I mean, never tried UXP one).