My Browser Builds (Part 5)

[roytam1]

New build of Serpent/UXP for XP!

Test binary:
Win32 https://o.rthost.win/basilisk/basilisk52-g4.8.win32-git-20250329-3219d2d-uxp-79614a4216-xpmod.7z
Win64 https://o.rthost.win/basilisk/basilisk52-g4.8.win64-git-20250329-3219d2d-uxp-79614a4216-xpmod.7z

source code that is comparable to my current working tree is available here: https://github.com/roytam1/UXP/commits/custom

IA32 Win32 https://o.rthost.win/basilisk/basilisk52-g4.8.win32-git-20250329-3219d2d-uxp-79614a4216-xpmod-ia32.7z

source code that is comparable to my current working tree is available here: https://github.com/roytam1/UXP/commits/ia32

NM28XP build:
Win32 https://o.rthost.win/palemoon/palemoon-28.10.7a1.win32-git-20250329-d849524bd-uxp-79614a4216-xpmod.7z
Win32 IA32 https://o.rthost.win/palemoon/palemoon-28.10.7a1.win32-git-20250329-d849524bd-uxp-79614a4216-xpmod-ia32.7z
Win32 SSE https://o.rthost.win/palemoon/palemoon-28.10.7a1.win32-git-20250329-d849524bd-uxp-79614a4216-xpmod-sse.7z
Win64 https://o.rthost.win/palemoon/palemoon-28.10.7a1.win64-git-20250329-d849524bd-uxp-79614a4216-xpmod.7z
Win7+ x64 AVX2 https://o.rthost.win/palemoon/palemoon-28.10.7a1.win64-git-20250329-d849524bd-uxp-79614a4216-w7plus-avx2.7z

Official UXP changes picked since my last build:
- No issue - Fix some deprot (6449b7514b)
- Issue #2711 - Apply pattern transforms when using text-rendering on canvas. (6411abd43c)
- Issue #2703 - Part 1: micro-optimize purging expired preflight cache entries. (d678f508de)
- Issue #2703 - Part 2: Don't repeatedly construct comparators when checking cache entries. (7854198085)
- Issue #2703 - Part 3: Make cache entry request checking more readable. (fdfe81fe54)
- Issue #2703 - Part 4: Fix incorrect clang warning in nsCORSListenerProxy. (8f2f4d8d3d)
- Issue #2713 - Check for NaN before std::min/max() in DOMQuad and DOMRect. (f1c456c625)
- No issue - Add range checking to SharedArrayBuffer references. (788cc4acb2)
- No issue - Fix incorrect js message ref in js shell. (c31cda6c74)
- Issue #2692 - Follow-Up: fix compilation in debug (7d3dc5a2f2)
- Issue #2715 - Compute Channel Binding Hashes using the certificate signature's hash algorithm as per spec (46fadb2d2b)

No official Pale-Moon changes picked since my last build.

No official Basilisk changes picked since my last build.

My changes since my last build:
- [Pale-Moon] bump general.useragent.app_version to 33.6.1 (71872cf0be)
- ported from mozilla: Add UnorderedRemoveElement{s}At to nsTArray, r=froydnj (3fa9e61e7e)
- js: use MOZ_MUST_USE instead of C++17 keyword (79614a4216)

Update Notice:
- You may delete file named icudt*.dat inside program folder when updating from old releases.

* Notice: From now on, UXP rev will point to `custom` branch of my UXP repo instead of MCP UXP repo, while "official UXP changes" shows only `tracking` branch changes.

Edited March 28 by roytam1

[roytam1]

New build of BOC/UXP for XP!

Test binary:
MailNews Win32 https://o.rthost.win/boc-uxp/mailnews.win32-20250329-7bda12e7-uxp-79614a4216-xpmod.7z
BNavigator Win32 https://o.rthost.win/boc-uxp/bnavigator.win32-20250329-7bda12e7-uxp-79614a4216-xpmod.7z

source repo (excluding UXP): https://github.com/roytam1/boc-uxp/tree/custom

* Notice: the profile prefix (i.e. parent folder names) are also changed since 2020-08-15 build, you may rename their names before using new binaries when updating from builds before 2020-08-15.

--

New build of HBL-UXP for XP!

Test binary:
IceDove-UXP(mail) https://o.rthost.win/hbl-uxp/icedove.win32-20250329-id-656ea98-uxp-79614a4216-xpmod.7z
IceApe-UXP(suite) https://o.rthost.win/hbl-uxp/iceape.win32-20250329-id-656ea98-ia-c642e3c-uxp-79614a4216-xpmod.7z

source repo (excluding UXP):
https://github.com/roytam1/icedove-uxp/tree/winbuild
https://github.com/roytam1/iceape-uxp/tree/winbuild

[roytam1]

New build of post-deprecated Serpent/moebius for XP!
* Notice: This repo will not be built on regular schedule, and changes are experimental as usual.
** Current moebius patch level should be on par with 52.9, but some security patches can not be applied/ported due to source milestone differences between versions.

Test binary:
Win32 https://o.rthost.win/basilisk/basilisk55-win32-git-20250329-fda70b1ae-xpmod.7z
Win64 https://o.rthost.win/basilisk/basilisk55-win64-git-20250329-fda70b1ae-xpmod.7z

repo: https://github.com/roytam1/basilisk55

Repo changes:
- import from `custom` branch of UXP: ported from mozilla: Add UnorderedRemoveElement{s}At to nsTArray, r=froydnj (3fa9e61e) (3da5d9cf8)
- ported from UXP: No issue - Fix some deprot (6449b751) (a846f6739)
- import from UXP: Issue #2711 - Apply pattern transforms when using text-rendering on canvas. (6411abd4) (eaa517905)
- import from UXP: Issue #2703 - Part 1: micro-optimize purging expired preflight cache entries. (d678f508) (703319d69)
- import from UXP: Issue #2703 - Part 2: Don't repeatedly construct comparators when checking cache entries. (78541980) (45bd57291)
- import from UXP: Issue #2703 - Part 3: Make cache entry request checking more readable. (fdfe81fe) (46bb59337)
- import from UXP: Issue #2703 - Part 4: Fix incorrect clang warning in nsCORSListenerProxy. (8f2f4d8d) (b5188f9a4)
- ported from UXP: Issue #2713 - Check for NaN before std::min/max() in DOMQuad and DOMRect. (f1c456c6) (8ca5b56e3)
- sync UXP changes: No issue - Add range checking to SharedArrayBuffer references. (788cc4ac) (1dc9aa75a)
- import from UXP: Issue #2692 - Follow-Up: fix compilation in debug (7d3dc5a2) (50c98682f)
- auth: SSPI: remove Telemetry (ccc66f744)
- import from UXP: Issue #2715 - Compute Channel Binding Hashes using the certificate signature's hash algorithm as per spec (46fadb2d) (fda70b1ae)

Edited March 28 by roytam1

[davidz]

13 hours ago, roytam1 said:

On 3/28/2025 at 2:16 PM, davidz said:

Hello roytam1!

In basilisk52-g4.8.win32-git-20250315 - when the multiprocess option is enabled, the 'Page Source Code' view doesn't work. (This only works in single process mode.)

The problem started with the version basilisk52-g4.8.win32-git-20241123!

Before that, on basilisk52-g4.8.win32-git-20241116 - everything worked fine.

Steps to reproduce the problem:

browser.tabs.remote.autostart", true
browser.tabs.remote.force-enable", true
dom.ipc.processCount", 6

When clicking on the 'View Page Source' item in the context menu, a blank page is displayed.
An error is displayed in the browser console(Ctrl+Shift+J):
TypeError: contentWindow is null  —  viewSource-content.js:226:11

Please fix this problem.

 

checked 20241123's changes but no idea what breaks it

So, it's useless to ask Moonchild(in upstream), considering that he is an ardent opponent of the multiprocess feature in UPX-browsers?
And we'll have to put up with this bug?

[j7n]

The arrogance of Discourse knows no bounds. Just as we learned to live with its quirks and hidden vertical scrolling, they announced a few days ago that even "recent" Chrome browsers will not be allowed. I saw this on Mp3Tag. They want even more "modern" features.

https://meta.discourse.org/t/dropping-ios-15-other-old-browsers-in-may-2025/358131

[VistaLover]

4 hours ago, j7n said:

They want even more "modern" features.

... Yes, this is really insane of them  ; it's not as if discourse-based forums, mostly containing text, need the top-of-the-line JS/CSS features released with "yesterday"'s Google Chrome  version ... 

Quote

The minimum Firefox version with support for these features is v128 (July 2024), and the minimum Chrome(ium) version is v119 (Oct 2023).

... Mozilla have extended Win7's support with Fx115esr until September, but discourse  will cut this support 4 months prior, on May 1st! I soon got bored reading the linked announcement and the comments that followed it, their basic reasoning couldn't hold water if it wanted to:

Quote

On the technical side, the features we’ll be requiring including relative color syntax, subgrid, and lookbehind regex, all of which will enable improved experiences for Discourse users.

(and, correct me if wrong, but "they" can't use English properly, I think "including" should've been "include"); what "improved experiences" for crying out loud?  People just post questions (in plain text), sometimes they attach a screengrab, and they expect a helpful reply (also mostly in text); why discriminate against older rendering engines?

https://meta.discourse.org/t/dropping-ios-15-other-old-browsers-in-may-2025/358131/33

Quote

I feel like people using old web browsers should still be able to use a basic gui framework, to comment, login, etc… as long as the web browser is still being worked on. There are a lot of web browsers that fit this qualification:

Badwolf, qupzilla, epiphany, basilisk-browser, palemoon, other uxp based web browsers etc…

Not everyone likes all this bloated eye candy that the web browsers embrace adding such as chromium based web browsers and firefox quantum and forks of that. (...) Each of the current modern web browsers, firefox, safari, opera, chromium web browsers, including edge and google chrome, all are infected with unneeded features, telemetry and, in many cases, data collection.

So… that’s how I feel about “modern”

Will have to wait and see how much "broken" UXP will be after May 1st on discourse-based forums/"communities" (quite a lot of them, actually) ...

(above is St52 with the 

! Discourse-based forums
||*/browser-detect-$script,important

custom uBO-legacy filter...)

EDIT: 

Discourse's browser-feature-checking script is, apparently:

https://d11a6trkgmumsb.cloudfront.net/assets/chunk.e772cb6376a12f35fc11.d41d8cd9.br.js:139:266343

Edited Sunday at 12:45 AM by VistaLover

[modnar]

While reading The Register about new Win11 "feature", I went to its Microsoft page and there (scroll down) linked_in, facebook and other socialist media icons are super-sized in 2025-03-27 version of Serpent 52.9.

Edited Wednesday at 11:34 AM by modnar
additional info

[nicolaasjan]

3 hours ago, modnar said:

While reading The Register about new Win11 "feature", I went to its Microsoft page and there (scroll down) linked_in, facebook and other socialist media icons are super-sized in 2025-03-27 version of Serpent 52.9.

Yes, the same issue can be seen in Pale Moon (official).

[Skorpios]

6 hours ago, modnar said:

While reading The Register about new Win11 "feature", I went to its Microsoft page and there (scroll down) linked_in, facebook and other socialist media icons are super-sized in 2025-03-27 version of Serpent 52.9.

Same ugly look with St55, NM28, MP68, lun3r, EE360/Chromium86 and MiniBrowser/Chromium87... all Experimental flags enabled in 360 and MB.

At least no warnings about "Outdated browser"...

---

Looks OK with Supermium/Thorium 122 (all Experimental flags enabled, might break other sites)

 

[NotHereToPlayGames]

1 hour ago, Skorpios said:

Same ugly look with St55, NM28, MP68, lun3r, EE360/Chromium86 and MiniBrowser/Chromium87

CSS

That's because the web site in question is using CSS NESTING (not all "failures" are javascript).

All of the browsers you listed can not use CSS Nesting.  Supermium/Chromium 122 CAN USE CSS Nesting.

The CSS being IGNORED/DROPPED by St55-Ch87 is this (note the indented &'s, this is "nesting", a css selector "inside" another css selector):
[in this case, it's selector #5 (img) inside selector #4 (a) inside selector #3 (li) inside selector #2 (...sharing-options...) inside selector #1 (...social-share...)]

 

CSS Nesting can not be rendered in anything below Chrome 120 and can not be rendered in anything below Firefox 117.

Edited Wednesday at 06:20 PM by NotHereToPlayGames

[AstroSkipper]

I would never have noticed such things. I would never think of reading something on a Microsoft page about Windows 11 in a legacy browser under a legacy OS. 

[VistaLover]

3 hours ago, NotHereToPlayGames said:

(note the indented &'s, this is "nesting", a css selector "inside" another css selector)

aka & nesting selector  ...

3 hours ago, NotHereToPlayGames said:

CSS Nesting can not be rendered in anything below Chrome 120 and can not be rendered in anything below Firefox 117

Partial (or full) support is behind custom (experimental) flags on earlier versions of both browsers: 

Probably useful info for people still on Win7 (where Cr109/Fx115esr are the last officially supported versions) ...

https://repo.palemoon.org/MoonchildProductions/UXP/issues?q=CSS&type=all&state=open

doesn't display an open issue for this CSS feature under UXP ...

Additionally, the footer on 

https://techcommunity.microsoft.com/

(as well as other content after the "Additional Help & Learning Resources" bit) isn't rendered properly (it renders as a vertical left-hand column) in UXP, possibly the same issue, too ...

FWIW, "Microsoft pages"  do not contain only stuff exclusive to Win11  , once the nesting selector makes its way onto other web "frameworks" and sites, the breakage on UXP browsers will become more common  ...

[AstroSkipper]

18 minutes ago, VistaLover said:

FWIW, "Microsoft pages"  do not contain only stuff exclusive to Win11 

FWIW, that's not what I said... 

[VistaLover]

26 minutes ago, AstroSkipper said:

that's not what I said

... I'm well aware , so that's why you weren't quoted ...

[NotHereToPlayGames]

6 hours ago, VistaLover said:

Partial (or full) support is behind custom (experimental) flags on earlier versions of both browsers

Agreed.  My intended point was more along the lines that these "experimental" flags effect both css and js.