Jump to content

My Browser Builds (Part 2)


Recommended Posts


1 hour ago, i430VX said:

On my K6 166MHz system, yes. That's the one. Haven't tried the Mendocino yet.

 

can you try redownloading the archive? new archive should have ucrtbase.dll dated in 2015.

according to this: https://stackoverflow.com/questions/52025833/crash-in-snprintf-s-in-ucrtbase-dll-with-status-illegal-instruction , newer ucrtbase relies to SSE(2).

Edited by roytam1
Link to comment
Share on other sites

Turns out it was because the AMD K6 seemingly doesnt have a full MMX implementation.
Serpent IA32 works fine on the Mendocino (full MMX)

Anyway:

Announcing browser installer version 3.1!

Change Log:

-Added Support for Serpent IA32
-added message to warn users to run as Admin
-updated serpent info in the info section

As always, get it here:
http://i430vx.strangled.net/files/XP/RoyTam%20Browser%20Installer.exe

Edited by i430VX
Link to comment
Share on other sites

12 minutes ago, roytam1 said:

can you try redownloading the archive? new archive should have ucrtbase.dll dated in 2015.

according to this: https://stackoverflow.com/questions/52025833/crash-in-snprintf-s-in-ucrtbase-dll-with-status-illegal-instruction , newer ucrtbase relies to SSE(2).

I can certainly try again on the K6, tomorrow. Just about to go to sleep here.

Edited by i430VX
Link to comment
Share on other sites

7 hours ago, Mathwiz said:

Interesting, since the regular (SSE2) build does support E10S. Not a big deal, but curious: did @roytam1 have to remove the e10s code to make his SSE build work?

Enabling E10S causes several rendering errors. But it's okay, I can do fine without. It's going to be a fun challenge to develop a UOC Patch for it.

Link to comment
Share on other sites

13 hours ago, ClassicNick said:

I have a Dell Inspiron 600m with a Pentium 4 Mobile (SSE2 compatible), and running Windows 2000 Pro USP 5.1 with Unofficial Update Rollup (Nov. 30th 2014). In My Experience, I can run most Windows XP only software, but applications that require SSE2 support don't work. I tested Serpent SSE, but I get an error message "The procedure entry point CoGetInterceptorFromTypeInfo could not be located in the dynamic link library ole32.dll". "Could't load XPCOM. I will post a screen shot shortly. I use K-Meleon 76.2 Goanna 20191019, and Newmoon 27.9.6 (20190223) and both work well. Congratulations on your browser builds!

A newer KernelEx from blackwingcat fixes that error but if you'd rather not update that then you can fix it manually.

Open a Hex Editor, I use HxD. Locate and open xul.dll then type in "CoGetInterceptor" in search and replace it with this string:

OleGetIconOfFile

Right under that you should see: "CoGetInterceptorFromTypeInfo". Replace it with this string and save:

StgCreateDocfileOnILockBytes

It should now open in Win 2000.

 

 

Edited by DanR20
Link to comment
Share on other sites

Unfortunately, SSErpent, being based on 52 ESR, carries over the broken APZ and Tiled compositing issues of other 52 ESR browsers. So while the UOC Patch I am developing does speed up scrolling, unfortunately it gets choppy at 1024x768 and beyond due to those issues.

@roytam1roytam1 Would it really be impossible to fix that? Or even porting the rendering engine used in FF45 ESR SSE to 52 ESR (and so, SSErpent)? That would make my life much easier, as I have to come up with creative ways to circumvent this problem, which unfortunately shows up if you increase the resolution (especially with ATI cards). 45 ESR has both APZ and Tiled Compositing working, and it really annoys me to know that are broken in the latter versions, because the Patch would speed them up massively if it wasn't for that.

Even implementing the semi-working Tiled compositing implementation of New Moon 27 would be fine, as the rendering errors (cosmetical only) don't hamper the performance, unlike 52 ESR browsers.

Edited by looking4awayout
Link to comment
Share on other sites

1 hour ago, ClassicNick said:

I now get a generic error message "basilisk.exe has generated errors and will be closed by Windows. You will need to restart the program. An error log is being created."

This is after making the change with the Hex Editor? I'm still doing it this way since I've got an older ole32.dll version of blackwingcat's kernelex and it 's always worked.

It might be related to USP 5.1 with Unofficial Update Rollup, I've never installed that so it's impossible for me to know. Installing a kernelex might conflict with that, not sure. Your best bet would be to go to the win 2000 thread and see if blackwingcat can help you directly:

https://msfn.org/board/topic/149233-kernelex-for-win2000/page/84/

 

Link to comment
Share on other sites

19 hours ago, looking4awayout said:

I have been testing SSErpent (just a nick I gave to the browser due to it working on SSE only systems) and while it does work, it's a little bit on the slow side.

I'm into this name- simple, catchy, and easy to remember what it refers to. Y'all who like it as well should upvote his port, so we know how popular it gets ;)

Link to comment
Share on other sites

16 hours ago, roytam1 said:

IIRC I didn't remove any e10s bits from my UXP repo.

10 hours ago, looking4awayout said:

Enabling E10S causes several rendering errors. But it's okay, I can do fine without. It's going to be a fun challenge to develop a UOC Patch for it.

OK, so the e10s code is still there; it just doesn't work properly in the SSE-only build. Weird, but it is what it is, I guess.

Wouldn't the NM28 version of the patch be your best bet, since the only difference between the 52 version and the NM28 version was the latter not enabling e10s?

2 hours ago, looking4awayout said:

Unfortunately, SSErpent, being based on 52 ESR, carries over the broken APZ and Tiled compositing issues of other 52 ESR browsers. So while the UOC Patch I am developing does speed up scrolling, unfortunately it gets choppy at 1024x768 and beyond due to those issues.

Yes, I would generally expect it to behave the same as the SSE2 build; bugs and all, unfortunately. Fixing it is certainly possible, but probably not a trivial task. There were many changes between FF 48 and FF 49, AIUI. @roytam1 would have to figure out which one(s) broke APZ and tiled compositing; then perhaps they could be reverted, but it depends on whether later changes depend on the ones needing reversion.

Link to comment
Share on other sites

I hope there is a way to port the rendering engine of FF45 ESR SSE to 52 ESR. Otherwise developing a UOC Patch for such a platform will be always, pardon my word, "half-assed". :(

I wish I was able to fix it, but I'm not a programmer and I don't know where to begin, alongside not having the time to do such a task (heck, I even have a new version of the 45 ESR UOC Patch and Enforcer, still unreleased as I'm not having the time to do that!), but I really hope this can be sorted out for good.

Those browsers can run really great on old hardware, if only those two things could be fixed.

Link to comment
Share on other sites

I'm sure Mozilla had (what they thought was) a good reason to redo the rendering engine in FF 49. It's probably tied into a thousand other things. So I'd expect putting the old engine into a new browser to be a bit like transplanting a brain....

I imagine there's some way to fix it; I just don't think that's the right approach. @VistaLover's probably our best expert on these questions, though, so I'll defer....

Link to comment
Share on other sites

On Serpent 55 / Moebius latest build:

After last click to sign in into MSFN board a few minutes ago, I've got this message page from msfn.org/board/login:

"Secure Connection Failed.

The connection to the server was reset while the page was loading.

The page you are trying to view cannot be shown because the authenticity of the received data could not be verified.

Please contact the website owners to inform them of this problem."

 

- but after "Click again" button, all is OK, I'm logged.

...

At the same time as I was logging in to MSFN via Serpent 55, I had an open Slimjet browser avec Zenmate enabled (but not functionning - why?), maybe there was any influence on the behavior of Serpent 55...

Edited by msfntor
Link to comment
Share on other sites

20 minutes ago, Mathwiz said:

I'm sure Mozilla had (what they thought was) a good reason to redo the rendering engine in FF 49. It's probably tied into a thousand other things. So I'd expect putting the old engine into a new browser to be a bit like transplanting a brain....

I imagine there's some way to fix it; I just don't think that's the right approach. @VistaLover's probably our best expert on these questions, though, so I'll defer....

Usually, everything Mozilla deems as "good" is the complete opposite, unfortunately...

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...