Jump to content

Mathwiz

Member
 View Profile  See their activity

  • Posts

    1,858

  • Joined

  • Last visited

  • Days Won

    51

  • Donations

    0.00 USD 

  • Country

    United States

 Content Type 

Everything posted by Mathwiz

  1. Mathwiz
    Yep - same goes for me. Just as a reminder, the "web" doesn't use JPEG XL because Chrome pulled support for it, and no Web designer is going to use anything Chrome doesn't support. And Chrome pulled support for it because Google has their own alternative they want the Web to use instead. It's the "not-invented-here" syndrome.
  2. Mathwiz
    To be fair, MyPal 68 is still in beta, and its author lives in a country at war, so I'm not surprised that 68 still isn't very stable. It's stunning that it works on XP at all, given that Rust was supposedly fundamentally incompatible with XP. My comments were merely to express surprise that it's more stable in single-process mode, and to wonder whether that applies to Windows 7 or only to XP.
  3. Mathwiz
    @chermany4ever said he doesn't have enough RAM for multiprocess mode, so he should definitely make the settings changes to disable it, whichever FF-based browser he ultimately chooses. But my experience with using multiprocess mode in Serpent 55 has been generally positive. Besides higher RAM usage, it does have downsides I mentioned earlier, but instability hasn't been one of them. I haven't noticed an increase in crashes, and if a tab does crash, it rarely takes the whole browser down. Instead, you just refresh the crashed tab. I haven't tried MyPal 68, but I would've expected multiprocess mode to be even more stable by the time Mozilla got to version 68. Perhaps multiprocess mode has grown less stable on XP since Mozilla obviously doesn't test on XP, but I'd expect it to be OK on Win 7 if you have enough RAM. Make sure you get Waterfox Classic. AIUI, at some point "Waterfox" got sold to some corporation that wasn't committed to the original spirit of the project, so fans of the "original" Waterfox resurrected the original project as Waterfox Classic. Where it "comes from" is kind of complicated. AIUI, originally it came here from China - via Russia - but our own @NotHereToPlayGames made extensive improvements, removing telemetry, creating an "unGoogled" version, rebasing to reduce RAM use on XP, etc. It has a pretty good reputation on MSFN.
  4. Mathwiz
    Good suggestion. Lots of CSP errors helped me figure it out. Github now requires service workers merely to load its CSS! I use CSP to block service workers except when needed for a site to function. Had to add github.com and githubassets.com to the list of exceptions. In some ways I think M$ is worse than Google. Google just provides the rope; M$ actively uses it to try to hang you! Service workers just to load CSS. Sheesh. Could they possibly come up with a more complicated way to perform a simpler task? They must pay their Web developers by the number of lines of code they write!
  5. Mathwiz
    Well. Setting browser.tabs.remote.autostart to false is the "traditional" (FF vers. 54-67) means of disabling multiprocess mode. In FF 68+ this no longer works (at least not by itself), hence the article I linked to. I can think of a couple of reasons why it may still work in MyPal 68: @feodor2 intentionally re-enabled the setting in MyPal The other settings you listed in your post allow the "disabled" setting to work anyhow. If it's #2, the combination of four settings you listed may work in Firefox itself as well as MyPal 68. Should be worth a try, at least! Of course, if you're lucky enough to have the RAM (physical, or virtual via, say, a fast SSD) you should probably leave e10s enabled for better security and crash resistance. That advice doesn't necessarily apply to Serpent, though, where you have to forcibly enable e10s. (There's a pinned thread explaining how.) Many legacy extensions won't work correctly with e10s enabled, so you need to balance what you gain against what you lose.
  6. Mathwiz
    ... and/or (as others here have suggested) just switch off multiprocess mode: https://techdows.com/2019/08/multi-process-e10s-can-still-be-disabled-in-firefox-68-or-later-versions-here-is-how.html Mozilla made it harder in FF (and consequently MyPal) 68 and above; you now have to set a Windows environment variable vs. just changing a pref in about:config, but it's still doable. N16s is quite likely correct that forcing single-process mode increases vulnerability to Spectre/Meltdown attacks (and that probably applies to Serpent as much as FF or MyPal); hence the increased difficulty shutting it off. But if the browser hogs too much virtual RAM, it'll start thrashing and end up having to be shut down and restarted, so if you have little physical RAM (regardless of OS) it may be worth the security trade-off. But remember, a major reason for the OP preferring FF-based browsers was to keep as many of his legacy extensions as possible, which I why I suggested Waterfox Classic. If one uses a lot of extensions, switching browser platforms becomes a major hassle.
  7. Mathwiz
    For legacy extension support, you could try Waterfox Classic. (I've heard good things, but I haven't tried it myself.) For the memory consumption I'd suggest JustOff's "Lull the Tabs" extension, although I suspect it's a legacy extension that'll only run on Waterfox Classic.
  8. Mathwiz
    You're probably remembering JustOff's "Modify HTTP Response," which is how I initially solved the issue; but I think your user script is a better solution for me since, e.g., ViolentMonkey will run in multiprocess mode, while JustOff's extension won't. BTW has anyone yet noticed similar CSS breakage on GitHub? M$ seems determined to break any browser that won't run on Win 10+. Edit: Well, it works in a clean profile.... I wonder which extension or config setting only breaks GitHub? Edit 2: Well, it's uBO. (Grr.) Strangely, though, when enabled, GitHub breaks - yet uBO doesn't report blocking anything! How am I supposed to tell what to unblock?
  9. Mathwiz
    DPRK = "Democratic People's Republic of Korea" (known over here as North Korea), correct? AIUI all their Internet connections go through China. All this over a song. Hopefully things won't go that far. Still, it wouldn't hurt to make some contingency plans, just in case. Do we know anyone with the equipment to build Serpent/NM, the expertise to merge GitHub commits into Serpent/NM periodically, and the willingness to respond to issues that occur due to differences between NM and PM, or Serpent and Basilisk? I realize it's a tall order. (@feodor2 would seem an obvious choice, but he has his own geopolitical challenges.) It wouldn't have to be done weekly as @roytam1 has always done, if that makes things easier. It could also help if several folks picked up the pieces. There's no reason one person has to do everything.
  10. Mathwiz
    Serpent (52 or 55) fares a bit better with outlook.office.com if you enable multiprocess mode. (Can't do that with PM/NM.) Even Serpent will eventually start thrashing though, requiring either a restart, or at least closing and reopening the Outlook tab. Chromium-based browsers work better with Outlook. OTOH, we now have sites that work better on Serpent/NM (see previous page of this thread). So it looks like the ideal of one XP browser for all sites remains a dream.
  11. Mathwiz
    If so, an obvious thing to try would be Silverlight plus a custom user agent. You could try "Mozilla/5.0 (Windows NT 6.1; rv:45.9) Gecko/20100101 Firefox/45.9", which is what Serpent 52 uses.
  12. Mathwiz
    Yes, IIRC at the time (Dec. 2021) it was true, and there was some hostility between MCP and users of @roytam1's builds; hence my advice. I wanted to avoid giving MCP's Web site any "clues" (via the user agent string) that one was running an unofficial build. But that's not true any more, so I updated that old post - and thank you for releasing 32-bit builds! Even in the Win 7+ world, 32-bit PCs (and VMs) are occasionally found.
  13. Mathwiz
    IIRC Widevine 1.4.8 and 1.4.9 are too old and have been blacklisted by Google, so they won't work. Silverlight used to work for Netflix, but Netflix may have stopped supporting it.
  14. Mathwiz
    You could be right. In theory, I suppose it's possible that some newly-embraced JavaScript feature might confuse the JavaScript parsing code in an old browser like v11 so badly that it crashes rather than just signalling a syntax error. Perhaps one or all of the newly-common assignment operators (??=, &&=, or ||=) throws it for a loop.
  15. Mathwiz
    Martok is German, it seems; so there shouldn't be a language barrier. Also, just look for BigInt, not Promise.BigInt - BigInt is a data type, not specifically a Promise method. I must admit, I too am a bit puzzled as to why Promise.allSettled returns that message if you block the test script. The Promise class is supported - it's only BigInt that's missing. So I'm guessing the error message is wrong, or at least misleading, and that the true problem is related to needing the BigInt method and type after all. A big(int) problem with polyfilling BigInt, I suspect, would be performance. It might "work" but be too slow to be practical. Depends on just what FritzOS uses BigInts for.
  16. Mathwiz
    I'm not a JavaScript developer, but to me, both of those functions sound "polyfillable" at first blush. Of course there may be more functions needing to be pollyfilled; it looks to me like the software just checks for "Promise.BigInt" to see if a whole suite of "Promise" functions are implemented (probably because they were all implemented at the same time in Chromium and in Firefox, so checking for any effectively checks for them all). I'd give @martok a few weeks to take a crack at it. He's made great progress in the past few months in getting UXP caught up to Chromium 86 level, but the number of Googlisms to be implemented for full compatibility is evidently quite large.
  17. Mathwiz
    Oh, thank goodness! You have no idea how often I have the need to watch 3 videos at the same time!! </snark>
  18. Mathwiz
    It's all over my head. But as long as it works I'm happy. Been kind of quiet here lately....
  19. Mathwiz
    You might try again with the most recently posted version (2023.05.06) dynamic module imports were added to JS, so there's at least a chance That is incorrect. MITM will trigger a browser warning and DPI cannot decrypt packets encrypted with modern TLS ciphers. Please quit spreading misinformation.
  20. Mathwiz
    This release fixed Chase.com too! I guess define is defined now....
  21. Mathwiz
    Finally found the answer: if JS is disabled, HTML within <noscript> tags is executed, and it's properly formatted for our browsers: <noscript><link rel="stylesheet" href="https://support.microsoft.com/SocContent/css" /></noscript> <noscript><link rel="stylesheet" href="https://support.microsoft.com/SocContent/officeShared" /></noscript> <noscript><link rel="stylesheet" href="https://support.microsoft.com/SocContent/articleCss" /></noscript> <noscript><link rel="stylesheet" href="/css/TopNav/top-nav.css?v=y3fVhNR8laayLSfo-P3Q-CBl74RjRTQT6GeXgXCLJoc" /></noscript> <noscript><link rel="stylesheet" href="/css/MeControlCallout/teaching-callout.css?v=690pjf05o15fVEafEpUwgaF8vqVfOkp5wP1Jl9gE99U" /></noscript> <noscript><link rel="stylesheet" href="/css/SearchBox/search-box.css?v=bybwzGBajHicVXspVs540UfV0swW0vCbOmBjBryj9N4" /></noscript> <noscript><link rel="stylesheet" href="/css/sitewide/articleCss-overwrite.css?v=fnFBTMAbM2543ZbkNfpSyKgKIX54uJaVhbeyhZp8Uks" /></noscript> <noscript><link rel="stylesheet" href="/css/glyphs/glyphs.css?v=0Hf7KD3KuarPGDf55g1ICt-VY442qRabqObuIoFb6Bo" /></noscript> <noscript><link rel="stylesheet" href="/css/promotionbanner/promotion-banner.css?v=cAmflE3c6Gw7niTOiMPEie9MY87yDE2mSl3DO7_jZRI" /></noscript> <noscript><link rel="stylesheet" href="/css/ArticleSupportBridge/article-support-bridge.css?v=R_P0TJvD9HoRHQBEdvBR1WhNn7dSbvOYWmVA9taxbpM" /></noscript> <noscript><link rel="stylesheet" href="/css/StickyFeedback/sticky-feedback.css?v=weC9pd2Sy8mevUeLAfDK2H9-VuIOr3CQ8OeyytUpyO0" /></noscript> <noscript><link rel="stylesheet" href="/css/feedback/feedback.css?v=WbIIOpRmxm58LAO8kuENEUDlr_SNhBVl2chWF0yqRcY" /></noscript> Of course disabling JS also disables most of the functionality of the page, so a better solution would be preferred. @UCyborg proposed a solution, but it requires Proxomitron, which in turn requires ProxHTTPSProxy - seems like overkill to me.... So how about using the "Modify HTTP Response" add-on? After considerable frustration, I finally managed to create a filter that seems to work:
  22. Mathwiz
    Sure enough, Chase lowered the boom last weekend or so, killing Serpent once again. I don't even get a sign-on screen any more. At least MiniBrowser (Cr-87 based) still works (when using a Cr-95 user agent). Edit 360EE v13.0 (Cr-86 based) works too (again with the Cr-95 user agent). FWIW, here's a screen shot of (I think) the relevant portion of St 55's error console: I especially like the error messages that just say Error: Big help.
  23. Mathwiz
    But that's not what you said earlier: So first it was, UXP needs Rust to properly implement m10s mode; now adding Rust code to UXP would just be "useless cruft." Which is it? I also sense a straw-man argument. I didn't say MCP should incorporate Rust code into UXP, only that they could, if they felt it was necessary. (Actually, I'm surprised they didn't do so back when MAT was there, just to make our lives more difficult.) Come on; the "defining feature" of Chrome isn't multi-process, it's Googlisms: frequent additions to JS/CSS, proposed by Google, Inc., that require equally frequent browser updates to implement, thus ensuring obsolescence of any browser not backed by a development team large enough to implement the continual flow of new Googlisms. (Of course we have seen at least one "Mozilla-ism," so Google isn't the only one playing this game; but it's quite clear that Google is far and away the dominant player.) In theory that would probably work! The only problem is, companies at that scale are less committed to "philosophy," so it's easier for them either to just start with the dominant engine - Cr - and add their own flourishes (M$), or to work out a tech-sharing agreement with Google (Mozilla and - probably - Apple). That is true as well. E10s is certainly not for the smaller, slower systems many users are running XP on, and it makes less difference on 64-bit systems with essentially unlimited virtual RAM. It's probably best suited for larger, faster 32-bit systems, since you can use more virtual RAM without "maxing out" and crashing a single process. One nice thing about the implementation we have in Serpent is that it can keep e10s on a rather tight leash, limiting the number of processes to fit your PC's resources. Cr, OTOH, spawns processes like crazy - often several per tab. Even though they're generally rather small, they can quickly overwhelm a smaller PC if you open a lot of tabs.
  24. Mathwiz
    Since it provides important context but is only a few lines longer, here's the actual start of that thread: https://news.ycombinator.com/item?id=15278883 And that was written six years ago. Things have only gotten worse. For example, the Big Four are now the Big Three since Microsoft threw in the towel and jumped on the Cr bandwagon. I don't think Rust is a no-go for MCP, since they're targeting Win 7 and up. It would certainly complicate @roytam1's job and our lives, but @feodor2 has found ways to compile Rust code for XP; e.g., MyPal 68 - so Rust wouldn't necessarily be a show-stopper for us either. M10s doesn't fully solve the memory issues though. I have m10s active at work (32-bit WinXP) and while it lets me use Outlook.com for a while, eventually the process gets over 1 GB (!), the CPU usage goes to 99%, and I have to close the tab, wait for everything to calm down, then click "Undo Close Tab," after which I'm fine for a while longer until the cycle repeats. They really need to limit the cache size somehow, Rust or no Rust. It's worse than that: modern web pages are designed to work properly only on Chromium-based browsers. (Well, plus Safari, but I suspect if we could see Safari's code, it would look more like Cr than even modern FF. Whether we like it or not (in my case, definitely not, but it doesn't change the reality), Google's Chrome and Cr-based offshoots like Edge and (barely) Opera dominate today's browser landscape. While I criticize Mozilla for duplicating way too much of Chrome's look and feel, we're lucky that even modern FF exists as something of an alternative at all, let alone UXP.
  25. Mathwiz
    I must admit, I never tried GitHub's editor with multiprocess (m10s?) on; unlike you I rarely post anything there; besides, until recently, accessing GitHub required palefill, which is incompatible (i10e?) with m10s mode anyhow, so I had always opened GitHub links in a single-process profile instead. The issue you describe, though, sounded just like a bug that plagued several recent versions of Serpent. So I wondered if the same workaround would work? The workaround for the recent bug was setting the pref dom.keyboardevent.keypress.dispatch_non_printable_in_content (d58t?) to true. I had never reset that pref after the very recent fix, so I never realized there was still a problem while in m10s mode, but I can now confirm that resetting that pref to false (the default) does disable the normal functioning of BACKSPACE, and setting it to true makes BACKSPACE function normally again!
×
×
  • Create New...