My Browser Builds (Part 4)

[soggi]

The mirror of latest ArcticFox 40, BNavigator 0.9, Firefox 45ESR, IceApe 52, IceDove 52, K-Meleon 1.5.x/74/76, MailNews 52, New Moon 26.5/27/28, RetroZilla, RZ browser and Serpent 52/55 builds by @roytam1 has been updated -> soggi.org - tools.

changelog:
- added latest BNavigator 0.9 20221105 build
- added latest IceApe 52 20221105 build
- added latest IceDove 52 20221105 build
- added latest K-Meleon 76 20221105 build
- added latest MailNews 52 20221105 build
- added latest New Moon 27 20221105 builds
- added latest New Moon 28 20221105 builds
- added latest Serpent 52 20221105 builds

To don't lose track of things I want to update too someday...

todo:
- add various flash player versions
- add FlashFix for WinXP
- add VLC 2.2.8 (WinXP non-SSE2)
- add polyfill addons

-------------------------------------------

On 11/1/2022 at 9:42 AM, roytam1 said:

my ArcticFox 40 build is a "normal" build so it requires SSE2

my Firefox 28 build is plain VC2005 build so it doesn't require SSE/SSE2.

my RetroZilla 2.2 is plain VC6 build so it doesn't require SSE/SSE2.

my Palefill Polyfill 1.23 mod changes only install.rdf to allow SP55 to install. and the question of replacing GHLpf 1.2.19.3 is a "no", but I don't have decent idea on updating GHLpf for now.

Thank you for your fast answer! I noticed that I forgot to ask about your following builds:
- K-Meleon 1.5.4 doesn't require SSE?
- K-Meleon 74.0 requires SSE?
- K-Meleon 76.4.6 requires SSE/SSE2?
- New Moon 26.5.0 doesn't require SSE?
 

On 11/1/2022 at 1:15 PM, we3fan said:

Hi @soggi,

I can open ftp://ftp.dyu.edu.tw/pub/  with all my browsers.
But I can't open ftp://ftp.cc.ntut.edu.tw/pub/  with all browsers I tried, I even tried few countries (US Central, US East, US West, Canada East,
Canada West, France, Germany, Netherlands, Norway, Switzerland) with VPN but still nothing, I also get timeout.

It offered to download the file from ftp://ftp.cc.ntut.edu.tw/pub/Hardware/Diamond/display/fire-gl/fire-gl-1000-pro/b1kp154.exe  but after download
it had size 0 bytes, so it was not successful.

So you nearly have the same problem as I have.
 

On 11/3/2022 at 12:09 AM, VistaLover said:

The path that works for me is:

ftp://ftp.dyu.edu.tw/pub/Hardware/vendor/Diamond/display/fire-gl/fire-gl-1000-pro/b1kp154.exe

(206 KiB file...)

Can you, @AstroSkipper or someone else please look how large the whole /pub/ directory is and maybe even create a mirror FTP?

kind regards
soggi

Edited November 5, 2022 by soggi

[roytam1]

1 hour ago, soggi said:

Thank you for your fast answer! I noticed that I forgot to ask about your following builds:
- K-Meleon 1.5.4 doesn't require SSE?
- K-Meleon 74.0 requires SSE?
- K-Meleon 76.4.6 requires SSE/SSE2?
- New Moon 26.5.0 doesn't require SSE?

both of them are "Not require SSE/SSE2"

Edited November 5, 2022 by roytam1

[soggi]

1 hour ago, roytam1 said:

both of them are "Not require SSE/SSE2"

Both? You mean all four?

kind regards
soggi

[Mathwiz]

Quote

The settings should have a human-readable checkbox for "automatically play media".

No human-readable checkbox AFAIK, but there is a Boolean setting in about:config.

media.autoplay.enabled; defaults to false in St 55 (not sure of the default in St 52).

I suspect a good JavaScript programmer (@AstroSkipper?) could rig up a custom button to toggle that setting!

On 11/4/2022 at 2:13 AM, UCyborg said:

Why not just set browser.backspace_action in about:config according to your preference rather than mess with user scripts?

Probably because even if one were lucky enough to find it, it's an integer pref and it's not at all clear what 0 does, 1 does, 2 does, etc. Luckily we have @dmiranda:

On 11/4/2022 at 8:27 AM, dmiranda said:

0, which is the default = page back: a nightmare. In writing this post I had to redo it four or five times as I hit backspace to edit stuff. Dang.

1 takes you to the beginning of the page. At least I don't have to reload the page.

2 or any other number kills backspace.

The bug discussed above aside, I too have found the default behavior (backspace=page back) to be worse than useless, so I changed my setting to 2. It's way too easy to have the focus in the wrong place by accident, press that key, and lose everything you typed! If I wanted "page back" assigned to a key at all, it'd probably be F12, or maybe "Previous Track" on "multimedia" keyboards that have that key, but certainly not backspace!

[VistaLover]

14 minutes ago, Mathwiz said:

Probably because even if one were lucky enough to find it,
it's an integer pref and it's not at all clear what 0 does, 1 does, 2 does, etc.

... That's why the MozillaZine (not related to Mozilla) KB site was created in the first place :

http://kb.mozillazine.org/Browser.backspace_action

[XP2019]

Hello, thank you again for Basilisk.

I noticed it makes its own connections and data transfer to/from:
185.199.110.133
81.95.11.116

inetnum:        185.199.108.0 - 185.199.111.255
netname:        US-GITHUB-20170413

Abuse contact for '81.95.11.64 - 81.95.11.127' is 'abuse@afterburst.com'
inetnum:        81.95.11.64 - 81.95.11.127
netname:        OXYGEM
country:        DE

Is it possible to permanently eliminate these with a setting or source code change?

[Mathwiz]

34 minutes ago, VistaLover said:

That's why the MozillaZine (not related to Mozilla) KB site was created in the first place

Yes, and despite all my complaints about the company, Google still has a pretty good search engine (both unnecessary in this case since @dmiranda answered the question almost immediately after it was raised).

But a post answering a rhetorical question is usually meant to point out that the answer was not as obvious as the person asking it seemed to think it was! It's not as if we all read @UCyborg's post, slapped our foreheads, and thought, "Of course! Just set browser.backspace-action to 2! It was so obvious! Why didn't anyone think of that before, instead of suggesting all those Greasemonkey scripts?"

At any rate, another little browser annoyance gone. No more hitting backspace after an unintended mouse click outside the comment posting window and losing everything!

And while we're on the subject, what exactly did "issue 2019" fix? Is it related to the backspace bug, as @luweitest speculated? (And for the uninitiated - i.e., me - how does one search for "issue 2019" upstream so I don't have to keep asking @VistaLover for the answer?)

[AstroSkipper]

On 11/6/2022 at 2:53 AM, Mathwiz said:

No human-readable checkbox AFAIK, but there is a Boolean setting in about:config.

media.autoplay.enabled; defaults to false in St 55 (not sure of the default in St 52).

I suspect a good JavaScript programmer (@AstroSkipper?) could rig up a custom button to toggle that setting!

Thanks for the flowers!  The question mark though is a bit undefined and can be interpreted in different ways.  
Firstly, I am rather a batch than a JavaScript programmer if at all. Indeed, I created some custom buttons using JavaScript in the past. Reading reference manuals and using already existing code snippets was what I did to get the right code for a specific custom button. In some cases, I did it from scratch by myself, in others I used already existing code and modified it. One of my strengths!  But frankly, far away from a good JavaScript programmer. Otherwise, I wouldn't post websites here that no longer work properly due to JavaScript problems, which no one really seems to be interested in, though.

Secondly, as I posted in the previous browser thread https://msfn.org/board/topic/182647-my-browser-builds-part-3/?do=findComment&comment=1227973, I already created such a custom button. It toggles a boolean variable from true to false and vice versa in about:config. The status of this variable is shown by the button itself in the form of an image that changes its colour, i.e. its image accordingly. Actually, I created this button for a specific boolean variable, but it can work with any other if you change its code slightly. It works perfectly in New Moon 28 and probably also in other UXP browsers , but in Serpent 55, I really don't know. My custom buttons require the extension Custom Buttons, and I don't know if it is compatible with this browser. Truth be told, I never used Serpent 55 in the past.  So I have to test that. Anyway, I will write a post about this custom button in my thread "Extensions and custom buttons for UXP browsers - Corrections, modifications, adjustments, and special recommendations" soon. So, stay tuned! And, this is the mentioned thread: 

Kind regards, AstroSkipper 

Edited November 12, 2022 by AstroSkipper

[Nicholas McAnespy]

5 hours ago, roytam1 said:

both of them are "Not require SSE/SSE2"

Would doing a "plain" Firefox 28.0/K-Meleon 74.0, or a New Moon 26.5.0 build on Visual C++ 2008 change the CPU requirements due to enforcing MMX/SSE support? 

[roytam1]

6 hours ago, ClassicNick said:

Would doing a "plain" Firefox 28.0/K-Meleon 74.0, or a New Moon 26.5.0 build on Visual C++ 2008 change the CPU requirements due to enforcing MMX/SSE support? 

Quote

By default, all applications compiled against the Visual C++ 2008 Runtimes (static and dynamic linking) will only work on Windows 2000, Windows XP SP2 and later.

so no.

[dmiranda]

14 hours ago, Mathwiz said:

Google still has a pretty good search engine (both unnecessary in this case since @dmiranda

Almost there. I use now metager as main serach engine, google only when necessary. 

As per the latest sp52 update, it doesn't affect this bug -which, again seems to occur only with e10s on-.  I have noticed that backsapce also stops working (sometimes) in the address bar. 

 

[dmiranda]

16 hours ago, Mathwiz said:

But a post answering a rhetorical question is usually meant to point out that the answer was not as obvious as the person asking it seemed to think it was! It's not as if we all read @UCyborg's post, slapped our foreheads, and thought, "Of course! Just set browser.backspace-action to 2! It was so obvious! Why didn't anyone think of that before, instead of suggesting all those Greasemonkey scripts?"

 

What is interesting is that two builds prior this did not happen at all, at least on my set up.

[luweitest]

On 11/2/2022 at 5:40 PM, luweitest said:

After some tests I have pinned three preconditions for that bug to appear:

1, version: the latest basilisk52-g4.8.win32-git-20221029-3219d2d-uxp-a6338caef-xpmod.7z.

2, multi-process option opened, that is,  browser.tabs.remote.force-enable;true

3, my old profile

The 3rd condition is hard to debug. Given version 20221015 is fine at all scenes. I think maybe this changelog is relevant:

Issue #2019 - Do not dispatch keypress event for non-printable keys. (1392f5f22)

The 3rd condition could be deleted -- I was stupid that a new profile could not suffice the condition 2.

And I find a workaround.  After checking the Issue #2019 at https://github.com/roytam1/UXP/commit/1392f5f223b960d6332f9230687afcd3e8e5276f , I spot a line I think relative:

`+ pref("dom.keyboardevent.keypress.dispatch_non_printable_in_content", false);`

about:config do not have that pref, so I manually add

`dom.keyboardevent.keypress.dispatch_non_printable_in_content", true;`

Then the delete and backspace key function again.

p.s. Hope @roytam1 could make a fix according to my finding!

Edited November 7, 2022 by luweitest

[dmiranda]

2 hours ago, luweitest said:

`dom.keyboardevent.keypress.dispatch_non_printable_in_content", true;`

OMG. Thanks. Now we are back in business. Thank you @luweitest!

[luweitest]

I always have a lot of tabs kept open in Serpent52. Recently I found Serpent initiates a large bunch of connections to various address upon every start, presumably to fetch those tabs. It lags start of browser a lot. I hope it could make the connection when I click or select that tab. I searched but found no apparent solution. In Preference -> General -> startup -> when Serpent starts, select "Show a blank page" does not help, because it actually asks whether to save my tabs upon closing Serpent; The pref "browser.sessionstore.restore_on_demand" seems the right one from the description, but the default value true does not work;  add a pref "browser.sessionstore.restore_tabs_lazily, true" does not work either. Is there some other place to try?