My Browser Builds (Part 1)

[roytam1]

just have a deeper look into PM's webRTC code, its module is not updated to fx52 level and javascript module fails when loading https://test.webrtc.org/ demo.

so I decided to disable it again.

EDIT: pm28 builds updated

Edited January 14, 2019 by roytam1

[Gargam]

All right now with the new pm28 build : ENU -> FR ok !

Thank you a lot +++

[Tamris]

@roytam1 Thanks a lot, everything is back to normal again . Also thanks to @VistaLover , his detailed post saved us (again).

Edited January 14, 2019 by Tamris

[dencorso]

On 1/12/2019 at 3:14 AM, roytam1 said:

New build of basilisk/UXP for XP!

Test binary:
Win32 https://o.rths.cf/basilisk/basilisk52-g4.1.win32-git-20190112-f38edc94a-xpmod.7z

Something isn't right with this build. It's crashing randomly on opening a new tab. Nothing of the sort happens with v. 20181208, so it must be due to some change done recently. I imagined there'd be other reports about it, but since there are none, here's mine. Bug-hunting is on!

[TechnoRelic]

https://www.seamonkey-project.org/releases/#langpacks
For 'RT' Borealis XP Browser, have been able to successfully install two language packs from the above URL. Specifically, the "English (British)" and "Spanish (Spain)" ones. I used tips from (users here) 'RT' and 'VISTALOVER' in order to modify the INSTALL.RDF file to get this accomplished.

They show (list) as installed in BOREALIS Browser here:
ADD ONS MANAGER (Menu: Tools, Add Ons Manager) for LANGUAGES (Tab).
But I am NOT able to 'activate' either of them.

(Added a few days later) ==>
https://msfn.org/board/topic/177125-my-build-of-new-moon-temp-name-aka-pale-moon-for-xp/?do=findComment&comment=1158936
Okay, I got QUALITY feedback from RT (URL above, THANKS!). Who supplied a successful manner to install the SM-2.49 LANGPACKS in an apparently 'minimal' acceptable manner. The MENU Headings looked GOOD overall, when I did the install of the SPANISH Language PACK into the BorXP Browser.

 

Edited January 19, 2019 by TechnoRelic

[Mathwiz]

On 1/9/2019 at 2:29 PM, VistaLover said:

Hi ; this is a well known and documented limitation of said extension:

https://github.com/JustOff/ca-archive#compatibility-and-installation =>

The author (@JustOff) is a member of the Moonchild dev team, this addon was primarily conceived to be installed in Pale Moon and/or Basilisk; ... the latter does not officially support e10s (nor have any tests for it been performed with multiprocess mode turned on...).

However, the same author appears to be somehow affiliated with the Waterfox browser project, they are they ones kindly providing hosting space/bandwidth for all legacy extensions catalogued inside CAA.

Current Waterfox ... comes with e10s turned ON by default, hence CAA issue #2 was submitted; the extension author came up, in v1.2.2, with a "dirty hack" to deal with e10s in Waterfox:

What this hack in effect does is, instead of generating the

Multi-process mode is not supported now,
please disable it and restart Waterfox

pop up notification, upon clicking CAA's toolbar icon, it spawns a second non-e10s Waterfox window, where CAA can function properly (i.e. Waterfox can interpret the caa: protocol).

I have inspected the "new" code inside linked commit and, after local experimentation, have found out that implementing the same "dirty hack" for multiprocess Basilisk/Serpent is as simple as substituting the string "Waterfox"

(this is in file bootstrap.js) with the string "Basilisk":

- if (e10s && Services.appinfo.name != "Waterfox") {
+ if (e10s && Services.appinfo.name != "Basilisk") {

Hope it works for you as it did for me

Well, over the weekend CAA updated itself to version 1.2.3, overwriting the change you suggested. I reapplied the change and of course it works again, but obviously will have to do that every time CAA gets updated ... oh, well ....

[dencorso]

On 1/11/2019 at 8:28 AM, dencorso said:

On Basilisk/UXP, when I go to Help => About, what I get is this:

Which file should I edit to change the night-blue background or to add art to it? 

Reply to myself:
%PROGRAMFILES%\basilisk\browser\omni.ja\chrome\browser\content\branding\about-background.png     
 

 

[Mathwiz]

On 1/12/2019 at 8:13 AM, Sampei.Nihira said:

I have disabled WebRTC again.

The reason is a decrease in privacy:

https://browserleaks.com/webrtc

Specifically, WebRTC leaks two bits of info that may compromise your privacy:

1. Device ID hashes for your microphone & camera
2. The IP address on your local network

#1 can be used for browser fingerprinting, allowing the likes of Google and Facebook to secretly track your online activities; luckily it's mostly a problem for Chromium-family browsers, not for Firefox-family browsers like Basilisk. The latter browsers randomize the "salt" used in the hash whenever the browser is started, so unless you leave your browser session running for weeks at a time, the ability of anyone to use #1 as a secret tracking cookie is limited.

#2 is more problematic though. It can reveal your "real" address behind a VPN, which could be used for censorship, or alert the authorities that you're accessing "banned" material. More commonly, it will reveal one of those non-routeable local IP addresses starting with 10. or 192.168 assigned by your (real or virtual) router. That's less worrisome, but if it doesn't change often, it too can be used in conjunction with your public IP for browser fingerprinting.

If you don't need/use WebRTC, the website linked above contains instructions for disabling it and preventing those info leaks. But what if you do use it?

One solution might be the WebRTC Control add-on. This adds a toolbar button that simply toggles WebRTC on or off, a la the popular Flash Disable add-on. So you can leave it off for normal browsing, but turn it on before going to a site that requires it.

Edit: Should have checked first. Couldn't install WebRTC Control linked above. All three versions download OK but Basilisk reports that they all appear to be corrupt. Must be a bad hash somewhere Try the "classic" version of Disable WebRTC instead.

I think a better solution would be an add-on with a white-list, which would enable WebRTC automatically, but only for sites like Discord and Skype. But I haven't yet found a Basilisk-compatible add-on that has such a white-list 

Edited January 15, 2019 by Mathwiz

[Sampei.Nihira]

Hi

I use New Moon.

As Tobin also said:

https://forum.palemoon.org/viewtopic.php?f=4&t=21079&hilit=secure+baislisk+or+pale+moon

I think he is safer than Basilisk.

I prefer to disable WebRTC via about:config.

Thank you for your explanation.

Edited January 15, 2019 by Sampei.Nihira

[caliber]

watch later button (lower right corner) is missing on youtube under PM28.

 

 

[Mathwiz]

1 hour ago, Sampei.Nihira said:

I prefer to disable WebRTC via about:config.

That's fine, especially if you don't use it. Just turn it off and forget about it; there's no need to bother with add-ons.

The add-on I found just toggles the same preference in about:config. It's strictly a convenience for those who use WebRTC but want it off during general browsing for greater privacy protection

[404notfound]

The browser seems more aggressive in disabling incompatible extensions lately.

[Mcinwwl]

Hey, just a notice - I created a topic requesting dedicated board for roytam1 browsers https://msfn.org/board/topic/178117-roytam1s-browsers-dedicated-board-in-members-projects-sub-forum/  Take a look and vote in a poll if you agree

 

[Mathwiz]

On 6/21/2018 at 7:13 AM, TechnoRelic said:

These NewMoon (Basilisk) tweaks (about:config) were recommended by others to me.

canvas.poisondata ... true

I'd been using that myself, but decided to stop in favor of the "Canvas Defender" add-on.

The purpose of the canvas.poisondata pref is to randomize canvas data to avoid fingerprinting. However, there's a problem with this approach: most browsers don't have that feature. Even NM/Basilisk don't enable it by default. So all Facebook, Google, etc. have to do is to run the fingerprint twice, observe that the browser returns different results each time, and they'll know that the user belongs to the small subset who have browsers or add-ons that poison canvas data. Then other, less precise fingerprinting techniques can still uniquely identify users within that small subset.

Canvas Defender, in contrast, changes canvas data periodically rather than every time a fingerprint is attempted. The rate of the periodic changes can be set anywhere from once per minute to once per week (or turned off completely so you can change it manually as desired). The idea is to look like one of the "unwashed masses" who don't poison their canvas data (thus allowing tracking for a while), while in reality poisoning canvas data often enough to render said tracking useless.

[404notfound]

This video fails to play in latest Basilisk UXP

"Video format or MIME type is not supported"

Edited January 17, 2019 by 404notfound