Jump to content

My Browser Builds (Part 4)


Recommended Posts

On 1/13/2024 at 10:43 PM, roytam1 said:
On 1/13/2024 at 9:33 AM, DanR20 said:

Reposting this in case anyone else reproduces it, seems to be happening with noscript 5.1.9 enabled:

Problem signature:
  Problem Event Name:    APPCRASH
  Application Name:    firefox.exe
  Application Version:    4.0.5.8774
  Application Timestamp:    65a09bff
  Fault Module Name:    xul.dll
  Fault Module Version:    4.0.5.8774
  Fault Module Timestamp:    65a09c90
  Exception Code:    80000003
  Exception Offset:    003037a5
  OS Version:    6.1.7601.2.1.0.256.48
  Locale ID:    1033
  Additional Information 1:    0a9e
  Additional Information 2:    0a9e372d3b4ad19135b953a78882e789
  Additional Information 3:    0a9e
  Additional Information 4:    0a9e372d3b4ad19135b953a78882e789

Expand  

 

On 1/13/2024 at 6:44 PM, modnar said:

I too am encountering instability with the latest Serpent 52.9 - I do have noscript installed, but it's not configured to "forbid scripts globally".

Register.com crashes the browser, even with noscript disabled. Also majorgeeks.com makes it crash. For this week I have reverted to v52.9.0 (2024-01-05) (32-bit).

 

Clipboard_01-13-2024_01.png

Clipboard_01-13-2024_02.png

Expand  

you may redownload today's build and test again

and upstream created a new issue about regressions: https://repo.palemoon.org/MoonchildProductions/UXP/issues/2452

Edited by roytam1
Link to comment
Share on other sites


On 1/14/2024 at 4:49 AM, Mathwiz said:

if any extension crashes the browser, that has to be the browser's fault!

... Well, call me the odd one out, but (browser) extensions were created/exist to extend the browser they target, NOT the other way round :whistle:; especially in the case of a maintained browser, whose engine and features frequently change/are being modified to adapt to the (mostly Google governed) current web! 

As AstroSkipper has pointed out above (sorry ;), I've run out of "reactions" for today), NoScript-legacy is an abandonware now; why should the browser dev(s) commit to an obligation to support it (or other abandonware extension) indefinitely?

I understand the recent NoScript-related crashes may be related to some coding "mishap", but I have to emphasise once more that "upstream" do not code with NoScript in mind (or test the authored code with it installed); so, even if you

On 1/14/2024 at 3:25 AM, Mathwiz said:

wouldn't let "upstream" dictate my choice of extensions.

, you're indirectly impacted by their coding practices (and their determination NOT to support in code "your choice" of extensions ;) ) ...

Along my long journey using Mozilla Firefox (from v2.0.0.x to 52.9.1), I lost several extensions I considered "must have" at the time, because Mozilla devs made coding decisions (while still in the XUL era) that introduced incompatibilities with said extensions, that their authors were unable/not ar*ed to "fix"... Only very rarely and at older times, when Mozilla wasn't a Google subsidiary, did they care to "fix" popular "broken" extensions (DTA!, uBO, e.a.) by modifying themselves the core browser code, instead of transferring all burden to extension maintainers ...

Right now, I'm running a self-concocted version of Stylus (git-1.5.22+92) in St52 that isn't able to store its settings longer than a browser session, plus has other shortcomings even in that same session :(; I need that a recent version because the Fx52-EoL version (1.4.23) comes with an outdated CSS parser that prevents it from installing recent versions of "user.css" userstyles; no "legacy" userstyle manager exists that can support "user.css" styles; Stylem hasn't been updated for 1.5yrs, no plan to support Stylus-targeting styles, either...

It would be unrealistic to expect Roy Tam to come up with a "fixed" Stylus version for Serpent 52/55, on a par with latest Stylus 1.5.42; even more so if you direct your pleas to Stylus maintainers themselves, their mantra being

Quote

And anyway most people use modern browsers

UCyborg is (again) right:

On 12/23/2023 at 7:08 PM, UCyborg said:

Even with the theory that the fact that Moonchild crew explicitly opted out of supporting ChromeZilla extensions could encourage development of XUL versions of extensions, guess that's still a pipe dream in such a niche community.

MCP were delusional when they expected an XUL-extension-development resurgence, ready to cater to their no-WE-supporting browser platform... Frankly, I'm feeling very pessimistic over the challenges 2024 will put on "our" legacy browsers :dubbio:...

Edited by VistaLover
Link to comment
Share on other sites

On 1/13/2024 at 12:44 PM, modnar said:

I do have noscript installed, but it's not configured to "forbid scripts globally".

Register.com crashes the browser, even with noscript disabled. Also majorgeeks.com makes it crash.

I still am on the initial release of St52 (BuildID=20240111073248), late Friday night in my timezone (UTC+02:00), that doesn't contain the "NoScript mitigation":

https://github.com/roytam1/UXP/commit/c75d50addea4f134130f2f2d822aec7f56f4cbd3

NoScript, as posted already, ISN'T installed, both

https://www.majorgeeks.com/

and

https://www.register.com/

do NOT crash the browser and load OK; yet another proof that NS, when installed, does hook deeply into your browser, even when disabled :whistle:...

Though, in the case of "www.register.com", the browser mis-renders the site's CSS:

ZtJKg42.png

Can anybody else reproduce, or should I start digging into my dirty St52 profile (not an ideal way to pass one's Sunday evening :( ) ?

Edited by VistaLover
Link to comment
Share on other sites

20 minutes ago, VistaLover said:

Though, in the case of "www.register.com", the browser mis-renders the site's CSS:

ZtJKg42.png

Can anybody else reproduce, or should I start digging into my dirty St52 profile (not an ideal way to pass one's Sunday evening :( ) ?

Yes, can reproduce, it's also misrendered here. I don't have NoScript installed either, and never used it.

y-M50-ZEQy-Mx.png

Also, @roytam1, don't forget to create a Part 5 topic of your browser builds, this topic has reached 200 pages...

Edited by mina7601
Link to comment
Share on other sites

2 hours ago, VistaLover said:

I still am on the initial release of St52 (BuildID=20240111073248), late Friday night in my timezone (UTC+02:00), that doesn't contain the "NoScript mitigation":

https://github.com/roytam1/UXP/commit/c75d50addea4f134130f2f2d822aec7f56f4cbd3

NoScript, as posted already, ISN'T installed, both

https://www.majorgeeks.com/

and

https://www.register.com/

I'm sorry - I should have written https://www.theregister.com/

not just register.com - don't know about that one TBH. :-)

Thank you for your in-depth analysis. That noscript sure is a piece of work, but I guess I have it installed for historic purposes...

By the way - I do not see any need to break the topic into numerous parts...

Edited by modnar
Link to comment
Share on other sites

1 hour ago, modnar said:

I'm sorry - I should have written https://www.theregister.com/

not just register.com

Well then ;) , I can indeed tell you that:

https://www.theregister.com/

is able to crash Serpent v52.9.0 (2024-01-11) (32-bit) in xul.dll (without ever having installed NS in its profile), while the hotfix release Serpent v52.9.0 (2024-01-13) (32-bit) can load that site fine!

So, thank you @modnar :worship: for reporting a NS-unrelated case where the "microtasks" bug can be witnessed! :)

Kind greetings, good night :hello: and have a good week ahead, too!

Edited by VistaLover
Link to comment
Share on other sites

4 hours ago, modnar said:

By the way - I do not see any need to break the topic into numerous parts...

It's a decision made in the past by @Dave-H and @dencorso, and @roytam1 agreed, that whenever his topic reaches 200 pages or more, a new topic is started to continue discussions, and the old topic gets locked. See these:

Edited by mina7601
Link to comment
Share on other sites

Yes indeed.
:yes:
We usually do it when new builds are issued, so if you could create a new thread for your next builds @roytam1 that would be good, and I will then lock this thread.
Thanks, Dave.
:)

Link to comment
Share on other sites

On 12/17/2023 at 4:03 AM, Mathwiz said:

It's possible I got this wrong, but if not, a "straight" FF UA is no good for Intel.com either:

Mozilla/5.0 (Windows NT 10.0; rv:115.0) Gecko/20100101 Firefox/115.0

(Slightly OT: ) I recently became aware that the "geckorevision" value for FxESR 115.x.x has been frozen at "rv:109.0"; so, actual FxESR 115 on a Win10 x86 OS will report:

Mozilla/5.0 (Windows NT 10.0; rv:109.0) Gecko/20100101 Firefox/115.0

You can read more about that here (bugzilla#1805967) ;) ...

Additionally, some years ago, the decision was made by Mozilla to deprecate the "WOW64" slice, so, even if you're using the 32-bit release of FxESR on Win7 x64 OS, your browser will advertise its bitness as x64:

Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:109) Gecko/20100101 Firefox/115.0

Read more at bugzilla#1559747 :whistle:...

Edited by VistaLover
Link to comment
Share on other sites

7 hours ago, VistaLover said:

..Frankly, I'm feeling very pessimistic over the challenges 2024 will put on "our" legacy browsers :dubbio:...

Yeah, I sort of feel the same way; like, hate to put that out there. I think even with the new Chrome we are waiting for XP will become even more useless but I intend to keep it around so long as the (failing) hardware is still going, and honestly, it could be 'any day' for me with this system and on to my ancient Dell 3000 which is a terrible computer.

Sorry to go off topic on you guys, good thing you love mister blobfish.

Edited by XPerceniol
Link to comment
Share on other sites

On 1/13/2024 at 12:02 PM, Ben Markson said:

Good news, Fritz!Box now works with St52. I see several changes with respect to Promise

... Yes, and you'll have to thank @martok for that (the German creator of PaleFill :P ); martok is himself a Frtiz!Box owner, so he had immediate interest in getting this fixed (as it turned out, all by himself ;) ) ; look here ...

The "fix" materialised as Implement PromiseRejectionEvent but, while working on that, Align microtasks and promises scheduling with current spec also had to be implemented ;) ; however, that last one is the cause for the much-discussed here NoScript-related crashes, so the moto "fix something, break something else" :whistle: comes to mind (often inherent with coding, especially on such complex sourcebases) ...

The new issue opened about it looks like is heading to a "wontfix" (at least where Moonchild is concerned), but someone should alert martok (somehow) that at least one URI induces a browser crash without the presence of the "evil" NoScript extension :dubbio:...

Edited by VistaLover
Link to comment
Share on other sites

1 hour ago, VistaLover said:

but someone should alert martok (somehow) that at least one URI induces a browser crash without the presence of the "evil" NoScript extension :dubbio:...

I made a post on their forum. BTW, I couldn't open account on their repo, hcaptcha.com just refuses to work on Pale Moon, the dialog doesn't popup. Anyone here wants to check https://accounts.hcaptcha.com/demo? Of course it works on Edge. But they left GitHub because of issues like that.

Edited by UCyborg
Link to comment
Share on other sites

1 hour ago, UCyborg said:

I made a post on their forum

https://forum.palemoon.org/viewtopic.php?f=62&t=30769&p=247631

Many thanks! :thumbup

1 hour ago, UCyborg said:

hcaptcha.com just refuses to work on Pale Moon, the dialog doesn't popup.
Anyone here wants to check https://accounts.hcaptcha.com/demo?

Sure thing ;) ; I had no issue passing the hcaptcha challenges (two successive sets of images) in St52 (dirty profile):

CCQYnQq.png

The text next to the green tick reads "I'm human", the text in green letters reads "Challenge Success!"

Have you tried in a fresh PM profile? I can report similar success when using latest NM28, but won't post a screengrab, due to late-night laziness :whistle: ...

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
  • Recently Browsing   0 members

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