mjd79
Content Type
Profiles
Forums
Events
Posts posted by mjd79
-
-
Affinity, that I just didn't come across it
I remembered from the time of porting Chromium 110 to Win7 about the GetProcessAffinityMask function, so I used it, and now both versions 138 and 139 work. I'll test it for stability, codec and extension performance (initially YT and Ublock work) and write a proper tutorial. In general, this means that we will 99% be able to port the upcoming probably in a few months the next version of ESR.
I added the 3 missing imports as redirects in the original kernel32.dll, but you can simply swap in CFF Explorer
xul.dll
GetSystemCpuSetInformation -> GetProcessAffinityMask
GetSystemTimes -> GetSystemTime
mozglue.dll
api-ms-win-core-version-l1-1-1.dll -> version.dll
api-ms-win-core-realtime-l1-1-1.dll and function QueryUnbiasedInterruptTimePrecise -> api-ms-win-core-realtime-l1-1-0.dll and QueryUnbiasedInterruptTime
The rest of the missing .dll imports in mozglue.dll need to be replaced with kernel32.dll
And of course change MajorOperatingSystemVersion and MajorSubsystemVersion from A to 6 in firefox.exe.
Edit:
Default-browser-agent.exe does not work (problem in ucrtbase.dll) and crashreporter.exe (user32.dll GetDpiForWindow function is missing, it can be replaced, but then the window will be the smallest possible size and any content will be seen only after the window is maximized)
2 -
I'm very close to launch Firefox 138 beta and the latest nightly 139.0a1, I'm missing the GetSystemCpuSetInformation function. Without it, FF to 138 launches but does not load any page. From 139 it doesn't work after replacing it with others at all. It supposedly appears in the api-ms-win-core-processthreads-l1-1-3.dll file in Windows 10 10240. And indeed, I checked it, and the kernel32.dll from this compilation imports this function from that file. The problem is that I can't find that file anywhere, I only found the downlevel ones, which are the ones I already have in 8.1. I will immediately point out that version l1-1-2 does not include GetSystemCpuSetInformation, both in 8.1 and 10.
https://learn.microsoft.com/en-us/uwp/win32-and-com/win32-apis
In fact, I've already succeeded, except that in the original kernel32 under a different name, I added redirection of the missing function to a wrapper created by a certain cracker from Github.
However, I prefer to avoid this and solve it on my own, as this is the only missing dependency I have left.
1 -
12 minutes ago, VistaLover said:
... Please, STOP spreading untruths! The code is there on GitHub for those willing to read it; after all, Supermium is still OPEN source (minus the wrapper DLLs, that is):
https://github.com/win32ss/supermium/issues/1290#issuecomment-2764577016
Ok, you can find information about this in the source code, but there should be mentions with the releases themselves that Supermium 132 is much less secure than 126 with a working sandbox.
3 -
In any case, the disabled sandbox, without any information on github or in the browser itself (other than chrome://version or chrome://sandbox, which 99% of users will never look into), is simply a scandal!
5 -
Then the likelihood is that Supermium 124 and above are based on Chrome 111 or similar, and Supermium 122 and older are based on 110, 109 or even older.
Edit: As I see, I am not the only one who discovered this sandbox problem and it has already been reported on github. It will be interesting to see how the author responds to this.
5 -
I have reviewed Supermium versions 122 through 132 R1, and I believe that even if the theory about the older engine is true, it uses the same as Supermium 124, not 122. The 122 version differs too much in my opinion, if only in the imports and wrappers used (e.g. since version 124 uses APIs such as DiscardVirtualMemory) . Besides, going into chrome://versions on Supermium 122, 124, 126 R7 and 132 R1, I discovered that the latest version of Supermium probably uses the --no-sandbox flag by default under Windows 8.1, and in the version for older systems, as well as the version for Win 10 and 11, which interestingly continues to work after replacing DiscardVirtualMemory with VirtualAlloc.
6 -
I checked and it seems to be a problem with Supermium 132 R1. Even my custom 2830.0 doesn't work, which works on all other Chromium browsers under Win7, including Supermium 126 R7.
BTW I also discovered that the pwrp_k32.dll included with Supermium 126 R7 allows 2830.0 to run in any reasonably modern (not tested below v109) Chromium browser in Windows 7. Just open widevinecdm.dll in CFF Explorer and rename the import from kernel32.dll to pwrp_k32.dll
2 -
If you are using 10/11 it should work fine. Delete this manually copied Widevine and update in chrome://components
https://dl.google.com/widevine-cdm/4.10.2830.0-win-x64.zip
Alternatively, try version 2830.0.
1 -
Currently, the case is as follows:
Windows 8.1 and 8.0 - Widevine is running the latest version, 4.10.2891.0, in every Chromium-based browser as well as Firefox (although 115.21.0ESR is likely to download 4.10.2830.0, which also works)
Windows 7 - so far I have not been able to run Widevine newer than the aforementioned 4.10.2830.0, even in Supermium 126 after applying the included patch. I didn't respond to inquiring users earlier because the original method of running Widevine from Firefox on browsers like Chrome 109 or Opera 95 required the use of the kernel64 library, which is a wrapper of kernel32 created by a Russian cracker porting closed-source browsers to older systems. I will try to find another option which is legitimate and presentable on MSFN without breaking the rules.1 -
16 minutes ago, D.Draker said:
For example, in Opera there's a signature file for every important Opera's file inside the package. Sorry, they forgot to ask you.
Apropos of .sig files, an interesting fact is that modified Firefox is not blocked, and also has them.
1 -
13 minutes ago, D.Draker said:
I lost you! So you're not from Poland? Are you from United States of Armenia, then?
No, @NotHereToPlayGames has captcha on every browser, hence the suspicion that this is what happens when you enter from an IP outside Poland. I live in this country, have internet from a local provider, and the captcha doesn't show up at all, the site normally works unless I use modded Chromium browsers.
0 -
Perhaps this is the case with IPs from outside Poland. In any case, on any unmodified browser you don't have to pass the captcha, and even if you do, it succeeds.
0 -
Of course, I perform the tests on a completely fresh profile, without changing any flags or settings and without extensions.
0 -
2 minutes ago, NotHereToPlayGames said:
Define "doesn't work".
The 100% correct solution to the test always ends up blocked, only on modded Chrome. On unmodified versions or on Firefox at the same time the site runs without any verification, so the browser is blocked, not the IP or MAC address.
0 -
19 minutes ago, NotHereToPlayGames said:
I shall entertain followup's to you and your provided details.
Are you saying that allegro.pl *WORKS* with Chrome 122 and an UNTOUCHED kernel32? But does *NOT WORK* with the modifed kernel32?
From my understanding with this new info, it would be extended kernels that are being blocked, not Chrome 122 being blocked.
It doesn't work in both cases...
0 -
As for my Chrome 122 (122.0.6261.112), I tested two scenarios. One was changing the kernel32 import DiscardVirtualMemory to VirtualAlloc in chrome.dll and DWrite.dll to one from Windows 10 1511, the browser was then launched with --no-sandbox. The other is to use a custom kernel32 for chrome.dll, to which I can only provide a link on PM, because the author is generally a cracker, it fixes the sandbox, and a custom modified original kernel32.dll using all the virtual memory functions from the wrapper included in Supermium 132 R0 (which fixed the leaks of that memory) for chrome.dll. DWrite also with W10 1511.
In my case, the problem is with allegro.pl.
0 -
2 hours ago, D.Draker said:
The same result with either moi IP or VPN. I suspect the don't like my ported Chrome 125 as it's fingerprinted as "modified".
I have a similar problem with ported Chrome 122 on Windows 8.1 with allegro.pl website. I get blocked by by their counterpart of captcha every time. Flags etc don't help. Could you please check what will happen for you? Thanks.
0 -
On 3/1/2025 at 6:30 PM, NotHereToPlayGames said:
Disregard! That was short-lived! Videos are consistently stopping about three and half minutes in then the infamous swirling circle indicating buffering rears its ugly head!
IceCat remains fine - without VORAPIS.
This appears to be a problem with Ublock Origin MV2. I was able to reproduce this problem on Opera versions 95 and 117 with Ublock 1.62, but it works with Adguard MV2 4.4.22.. In any case, it works fine in Firefox with Ublock 1.62, oddly enough.
0 -
I'll check it out, but it seems that it will be hard for me to run anything above Chromium 110. D. Draker as far as I know ran Chromium up to version 125. BTW. that FF 137 has the same problem as FF 130+ on XP with One Core API, that is, any codec like VP9 or H264 does not work on YT. Perhaps someone can solve it, if not, I'll wait to see what the OCA developer does about it.
2 -
I don't have time for thorough testing, but FF 137 x64 compiled with patches by e3kskoy7wqk
https://github.com/e3kskoy7wqk/Firefox-for-windows-7worksruns on Vista with the extended kernel dated March 9, 2023. It only requires the api-ms-core-win-synch-l1-2-0.dll library from Windows 10 (I used from compilation 10240) and osver.ini, which in my case looks like this.[C:\Users\(username)\Desktop\firefox-137.0.en-US.win64\firefox\firefox.exe] enabled=1 Win7SuperVerFix=1 MajorVersion=6 MinorVersion=1 BuildNumber=7601 CSDVersion=Service Pack 1 PlatformID=2Of course type your path to the Firefox file, you can also type [global] instead
Without this dll library, the browser crashes when loading with 100% CPU usage, and without osver.ini it won't display any text correctly.
Edit: No playback of any videos works.
2 -
Ximonite hosted the win32ss site for a long time, so I thought it was safe....
0 -
1 hour ago, Klemper said:
I'm on 8.1, too. Could you please share this programme?
2 -
You guys misunderstood me lol. I only gave a link to the drivers, and above I wrote about the integration tools.
But you are right that it is best to use the Gigabyte tool.
0 -
Last Version of Software for Windows 8.1
in Windows 8
Posted · Edited by mjd79
I don't need it, I've already replaced them with empty files, I just posted about it informatively. https://github.com/e3kskoy7wqk/Firefox-for-windows-7 includes a working crashreporter, if anyone want a 100% complete browser (and actually with unnecessary additions).