My Browser Builds (Part 4)

[ClassicNick]

Another Update: I got Firefox 3.0b4 building on Visual C++ 6.0 SP5 with the "Windows" GFX toolkit. The display/rendering bugs are the same as in Firefox 3.0b1, but with text rendering properly this time. I know I want to build Firefox 3.0b5, but I don't know what version I actually want to settle on for a new fork. Since my ability to use GitHub (especially the desktop version) is as great as my ability to use Cygwin, @roytam1 I will eventually need you to upload source code for my browser so I can fork it on the GitHub website. Between Firefox 3.0b2 (Note: Search functionality doesn't seem to work when compiled on Visual C++ 2003 using the Thebes "Cairo-Windows" GFX toolkit, but I like it for being "lightweight" in comparison to the newer versions), Firefox 3.0b5 (I'll probably switch the unit measurement definitions to help with the display issues in that version), or Firefox 3.0.19, which one do you recommend I use for a new Firefox fork? In order to keep a mental note of what I want to do, and in the order I feel it should be done, I created a task list on GitHub. https://github.com/ClassicNick/Fx3.0a-VC6-mod/issues/4

[XPerceniol]

On 12/16/2022 at 6:39 PM, roytam1 said:

- [network] move some generic websocket code to the baseclass (6a1e54041)

Good grief; finally :)

Nice ... now I feel better to enable the websocket as I've been rolling without it and I hope this improves things a little more. Thank you again for your efforts and work on the projects.

network.websocket.enabled;false
network.websocket.max-connections;0
network.websocket.max-message-size;0
network.websocket.timeout.close;0
network.websocket.timeout.open;0
network.websocket.timeout.ping.response;0

 

[Mathwiz]

1 hour ago, XPerceniol said:

Good grief; finally

Nice ... now I feel better to enable the websocket as I've been rolling without it and I hope this improves things a little more. Thank you again for your efforts and work on the projects.

network.websocket.enabled;false
network.websocket.max-connections;0
network.websocket.max-message-size;0
network.websocket.timeout.close;0
network.websocket.timeout.open;0
network.websocket.timeout.ping.response;0

 

Can anyone explain what "websocket" is and whether the rest of us should care?

[Mathwiz]

On 12/10/2022 at 5:15 AM, AstroSkipper said:

Hello @Mathwiz! MailNews is based on Interlink Mail & News from Binary Outcast. I read on their website that they seem to have removed the protocol OAuth2 from their email client. Here is the respective link: https://binaryoutcast.com/projects/interlink/release-notes/

Some weeks ago, I was also looking for an email client and found the program OE Classic which is supposed to support the OAuth2 protocol as well: https://www.oeclassic.com/ This is the one I'll try the next time. Maybe, it is what you are looking for.

Kind regards, AstroSkipper 

I will give OE Classic a try; thanks. Looks like it supports everything I need. The only thing I couldn't find at the link was, which features are only available in the paid version? Guess I'll have to download it and try it to find that out. But to be fair, the paid version isn't very expensive.

On 12/10/2022 at 4:19 PM, Mathwiz said:

Quite by accident, I got it to work! I went back to IceDove and suddenly, it started working. I had the protocol just set to "Normal Password," so there must be something about the interaction with the Office 365 mail server that automatically triggers OAuth2.

Well, that was completely wrong. When I got back to the office last Monday, on a lark I decided to try my old email client, Windows Live Mail 2009, which does not support OAuth at all. Surprisingly, it worked! This showed me that, despite what I had read, "legacy" authentication was still working, and I concluded that IceDove/IceApe wasn't using OAuth at all. It's just sending my ID and password to the O365 servers, then those servers contact my workplace servers to validate my password the "old fashioned" way.

M$ was supposed to disable all authentication methods besides OAuth as of Oct. 1, but I learned that organizations using O365 can get a short "reprieve," allowing old authentication to work until Jan. 1. My workplace must have done this. So I have a couple more weeks before I need to find a new email client; time I'm using to upload my old emails to Micro$oft's servers so they won't be "trapped" in Windows Live Mail when it stops working with O365.

That still leaves the question of why OAuth in MailNews didn't work. I think the link above provides a clue: BOC removed OAuth support from Interlink because the "big boys," particularly Google and Micro$oft, started requiring that apps be registered with their services in order to use OAuth, and he felt Google's registration requirement, at least, was too onerous. (His post didn't mention Micro$oft.) Without registration, OAuth support won't work.

I don't know how far he got before he pulled the plug, though. If he got as far as obtaining an application ID and "secret," those are presumably still present in MailNews and in theory could be registered with O365, allowing MailNews to work.

(Personally, I think a better approach than taking his ball and going home would have been to make the ID and "secret" user preferences that could be set, say, to Mozilla Thunderbird's ID and "secret." Then the user could just pretend to be using Thunderbird. This is the approach used by an OAuth "proxy" I found at GitHub.)

[AstroSkipper]

7 hours ago, Mathwiz said:

I will give OE Classic a try; thanks. Looks like it supports everything I need. The only thing I couldn't find at the link was, which features are only available in the paid version? Guess I'll have to download it and try it to find that out. But to be fair, the paid version isn't very expensive.

At the bottom of the order page, you'll find a more detailed comparison between Free and Pro version with the headline "Free and Pro version comparison".  Here is the link: https://www.oeclassic.com/order

Cheers, AstroSkipper 

[AstroSkipper]

@Mathwiz! I changed my mind a bit. Due to some limitations of OE Classic, I will first try MailNews. Although Binary Outcast had removed the protocol OAuth2, @roytam1 confirmed to support this protocol in MailNews. Anyway! OE Classic Pro seems to be worth its money.

Greetings from Germany, AstroSkipper 

Edited December 18, 2022 by AstroSkipper
correction

[AstroSkipper]

9 hours ago, Mathwiz said:

Can anyone explain what "websocket" is and whether the rest of us should care?

The WebSocket protocol is a TCP-based network protocol designed to establish a bidirectional connection between a web application and a WebSocket server. From a technical point of view, with WebSocket, as with HTTP, the client starts a request, with the difference that after the transmission of the data to establish the connection, the underlying TCP connection remains in place and enables asynchronous transmissions in both directions. Source: Wikipedia

Cheers, AstroSkipper 

[Sampei.Nihira]

12 hours ago, Mathwiz said:

Can anyone explain what "websocket" is and whether the rest of us should care?

In addition to what Astroskypper said I add that active websocket is almost indispensable in some situations such as online games,chat,online trading......

Usually one should not worry about websockets because modern browsers do not allow connection to an unsecured server from a protected source.

Try checking in the first test for an insecure connection (ws).

If it is not allowed you should not worry about websocket:

Test:

https://www.piesocket.com/websocket-tester

Test1:

https://www.deviceinfo.me/

It is easier to use a rule in the adblocker if you want to block websocket:

||*$websocket

in this way it is possible to insert exceptions.

Edited December 18, 2022 by Sampei.Nihira

[VistaLover]

6 hours ago, Sampei.Nihira said:

Try checking in the first test for an insecure connection (ws).
If it is not allowed you should not worry about websocket:

Test:

https://www.piesocket.com/websocket-tester

Most sadly, that test can't be run under a UXP-based browser , because the website requires Javascript features not supported in UXP:

This is a well known villain when it comes to UXP (i.e. dynamic module import):

https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/import

First implemented in Firefox Quantum 67, UXP has a longstanding "open" issue about it:

https://repo.palemoon.org/MoonchildProductions/UXP/issues/1691

TL;DR: No light at the end of the tunnel any time soon, if ever...
People, "we" should accept UXP is on its deathbed, under life support ...

Edited December 18, 2022 by VistaLover
proper English

[XPerceniol]

19 minutes ago, VistaLover said:

TL;DR: No light at the end of the tunnel any time soon, if ever...
People, "we" should accept UXP is on its dying bed, under life support ...

....ugh

[Sampei.Nihira]

53 minutes ago, VistaLover said:

Most sadly, that test can't be run under a UXP-based browser , because the website requires Javascript features not supported in UXP:

This is a well known villain when it comes to UXP (i.e. dynamic module import):

https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/import

First implemented in Firefox Quantum 67, UXP has a longstanding "open" issue about it:

https://repo.palemoon.org/MoonchildProductions/UXP/issues/1691

TL;DR: No light at the end of the tunnel any time soon, if ever...
People, "we" should accept UXP is on its dying bed, under life support ...

Check with the second test:

[AstroSkipper]

2 hours ago, VistaLover said:

People, "we" should accept UXP is on its dying bed, under life support ...

This has been the case for years and is an inevitable development. I can live with that as long as the life-support machines are not completely switched off. 

Edited December 18, 2022 by AstroSkipper

[XPerceniol]

Just now, AstroSkipper said:

That's been the case for years. I can live with that as long as the life-support machines are not completely switched off. 

I agree 100%

[XPerceniol]

1 hour ago, Sampei.Nihira said:

Check with the second test:

Should we have it set to:

devtools.webide.monitorWebSocketURL;ws://localhost:9000

network.websocket.extensions.permessage-deflate;true

network.websocket.extensions.permessage-deflate;true

network.websocket.delay-failed-reconnects;true

Edited December 18, 2022 by XPerceniol

[Mathwiz]

7 hours ago, AstroSkipper said:

At the bottom of the order page, you'll find a more detailed comparison between Free and Pro version with the headline "Free and Pro version comparison".  Here is the link: https://www.oeclassic.com/order

Cheers, AstroSkipper 

I don't think they'll mind if I repost it here, at least temporarily:

This isn't the full list, but gives you the idea.

Any of those could be reasonable in isolation, but they do add up! In particular, it seems like a bit of overkill to add ads to outgoing messages and impose all the other restrictions! Is it adware or is it feature-limited? Shouldn't be both, IMO. Guess I'll pony up for the paid version.