>>78748378 (OP)
XP tan needs to drink more water! What a dehydrated girl!

>>78748405
She needs to become m o i s t.

how does it build registry hive files?

Winlogon Anon - check toward end of last thread please! - bump

MSVCRT Anon - publish changed files

Links for retards that still won't bother to read the info in OP:

Glowie Torrent (Repack):
magnet:?xt=urn:btih:889857B7D802AB0AF2526AF60FCB3D9750E503C2
Original nt5src.7z (NOTREPACKED):
magnet:?xt=urn:btih:1a4e5b67060ff2bc8fe2de36a6c265c77f392a0c

>>78748436
Just search for registry.c files and anything that references the use of a .ini file, they use those to build up registry hives for certain apps I thing, and the .HIV hives are most likely a set of pre-defined or system set reg entries (definitely don't quote me though I haven't properly looked)

>>78748667
Based

File: 1601503611967.png (241 KB, 600x578)

241 KB PNG

>>78748378 (OP)
>not the poop edit
shame...

>>78749206
Poop and blood edits are the best.

File: blood.png (250 KB, 600x578)

250 KB PNG

>>78749237

File: blood.png (252 KB, 600x578)

252 KB PNG

>>78749343
Improved version. Blood looks darker and more blood-like.

>>78749650
Nice.

File: VirtualBox_XP Xbox BLD_17(...).png (418 KB, 1455x801)

418 KB PNG

to the guy trying to compile the Xbox leak, make sure you have your xbox\private\developr\<username>\ folder setup correctly. That'll set all the necessary razzle scripts going.

> image related
this is what happens when you have that configured properly, and simply type 'build' in the command prompt, for the entire 'Xbox\private' tree. I haven't looked into the built files yet myself. I'm curious what 11 files are not compiled.

This is the first time I'm trying the May 2020 leak versus the old Barnabas leak. There's a lot more in this leak, so I anticipate there to be a bit more broken.

File: Screenshot 2020-11-16 202127.png (56 KB, 800x447)

56 KB PNG

>>78749866
I believe my profile folder is set up correct, I only have 1 error though and it bugs me I am struggling to figure it out

File: Screenshot 2020-11-16 202223.png (9 KB, 733x368)

9 KB PNG

>>78749866
I forgot to add
>>78749908
I did this building the whole xbox/private tree without realising and only got this error above

File: VirtualBox_XP Xbox BLD_17(...).png (423 KB, 1455x801)

423 KB PNG

>>78749922
strange, the only things that get that error are debug libs for the dashboards.

I am building on XP, so maybe it's a compatibility issue with the build tools on 10?

I have noticed I have started Razzle differently to you though

>>78749964
I copied the files from the xbox\private\developr\template folder into my user's folder inside the developr\ tree

then I just ran razzle.cmd, and it executed all the scripts from the template

>>78749963
And possibly Windows 10 X86 has some updated libs that razzle use, it relies heavy from what I can tell on built in VC runtimes

>>78749993
Ahh okay, I just followed the Readme after applying the update bundled with the leak lol

>>78749963
>strange, the only things that get that error are debug libs for the dashboards.

Okay yeah that is strange why I am getting that error then as I am building fre.. funny thing is the file does actually get built (even after deleting and rebuilding)

>>78750051
Looking at it, it looks to be building for a devkit by default not retail, that could be the issue to my error

will exinit and the other one be in the prepatch files

>>78748378 (OP)
>https://rentry.co/build-win2k3
Wow, I didn't know we have such a cool FAQ section.

https://youtu.be/cHkdQ5CG0PY?t=345 [Embed]
Windows NT running on a MIPS R4000 processor in '93

Alright think I've finished up systime.c now, everything seems to match up to .obj fine, and didn't have any problems doing quick setup/boot tests.

I haven't tried stress-testing it at all tho, would appreciate it if people here can try it out, make sure it doesn't have the time-slowdown issue that the earlier systime.c had under real hardware, etc.

>win2003_missing-ex-files_v1.zip
https://gofile.io/d/XX77DP

Contains latest exinit.c/systime.c/exp.h.

>>78750946
I don't mind it being added, but maybe it should only be optional and not replace the original .obj files, at least until we're sure there's no issues with them anyway.

>>78748870
That's naive to think there will be registry.c. Also registry hives are created from .inf files, not .ini.

>>78752936
>https://gofile.io/d/XX77DP
Why there is change also in .h file?

>>78756671
Well it was because some funcs needed to be forward declared, but now I just checked and I think since I reordered some things it should work with the original exp.h fine now.

(The exp.h is actually the same as one provided in prepatched v10, just included it here in case someone wasn't using that, but now I guess it's not needed at all, maybe prepatched can remove the exp.h change too)

>>78756957
yeah I guess is better to let the original ex.h file so we are more compatible overall!

>>78757602
*exp.h

File: Снимок.png (42 KB, 979x512)

42 KB PNG

Almost working with original exp.h!

>>78757848
Are there any build differences when you set machine name as winmain?

>>78757848
Oh yeah whoops, seems there are some things added in exp.h that systime.c needs. Maybe those could be moved to systime.c fine though, I'll look into it later.

>>78758109
winmain should work same way as srckit because winmain branch not existed in 2003

>>78758109
also it's not machine name it's _BuildBranch

So, i have know a bit about coding, like studied one C++ textbook (robert lafore objective C++ 5 edition or whatever), and remember most of things from it five years later. Also had some classes about C. I have two questions:
1) Can i help at this time, and what is not done at this point?
2) Do anon here know much about windows server 2003 and windows xp 64bit? Can i found some leaks on xp64, or make enough changes at windows 2003 server version, so it will be recognised by applications as windows 64bit? They have same core, after all.

>>78758285
We don't have the complete code for amd64. amd64 was shipped from MS staring from win server sp1, but if you mean 64bit than yes if you are able to find and emulator for ia64 than for that we have the complete code :)

>>78758678
We still can compile amd64 there is only around 150 files missing in postbuild.

>>78758678
>>78758757
Thanks a lot. So, if i have a normal, more modern intel processor, i can make a build without that IA-64 emulation or something?

Bump

>>78748667
>magnet:?xt=urn:btih:889857B7D802AB0AF2526AF60FCB3D9750E503C2
fucking glow nigger, why would you use torrent

>>78758285
>1) Can i help at this time, and what is not done at this point?
There's SO much to be done, of course.
But keep in mind the C++ standard MSFT used for XP is quite dated.
>>78757848
You rock!

>>78760991
I think it's fine with me, since i never was "used" to C++ 2011 standart which was latest. How may i assist? Assuming i vaguely remember object orienting part of my textbook, as well as being on a "uni" level of language knowledge. Also i want to start with something which anon mentioned, a working XP 64 bit copy, based on windows 2003 server core. Is it hard to at least "emulate"?

>>78761346
https://rentry.co/build-win2k3
Read "amd64 build support"

>>78761477
Okay, thanks. What about intel? I have Intel.

File: 1592599610335.jpg (5 KB, 251x251)

5 KB JPG

>>78761538

>>78761686
Sorry, i just didn't get it. I mean, i have real troubles understanding what will be with system having a few parts missed, if they are patched thoughtfully. Make it clear to me, anon, i think i had same issue with other anon and I-64, you can look upon it.

>>78761761
*IA-64, for the fuck sake, i look even dumber.

>>78760933
>why would you be a rational person

maybe because:

* MS-DOS 3.30 OEM Adaptation Kit (source code)
* MS-DOS 6.0 (source code)
* DDKs / WDKs stretching from Win 3.11 to Windows 7 (source code)
* Windows NT 3.5 (source code)
* Windows NT 4 (source code)
* Windows 2000 (source code)

* Windows XP SP1 (source code)
* Windows Server 2003 (build 3790) (source code)
(file name is 'nt5src.7z')

* Windows CE 3.0 Platform Builder (source code)
* Windows CE 4.2 Shared Source (source code)
* Windows CE 5.0 Shared Source (source code)
* Windows CE 6.0 R3 Shared Source (source code)
* Windows Embedded Compact 7.0 Shared Source (source code)
* Windows Embedded Compact 2013 (CE 8.0) Shared Source (source code)
* Windows 10 Shared Source Kit (source code)
* Windows Research Kernel 1.2 (source code)
* Xbox Live (source code)
(most recent copyright notice in the code says 2009)
* Xbox OS (source code)
(both the "Barnabas" release from 2002, and the leak that happened in May 2020)

...and also OpenNT.

NOTREPACKED faggot torrent:

* Windows XP SP1 (source code)
* Windows Server 2003 (build 3790) (source code)

oh, and by the way, with NOTREPACKED you get to wait twice as long for your archive to extract. who wouldn't want that?

stop being an insufferable dumbass who falls for M$ psy-ops and retarded memes.

>>78761761
We can build x86.
We can partially build amd64.

Build Machine can be x86 or x64.

Missing files are provided, check the guide..

Read the info that answers all this..Do a build... then ask stupid questions.

>>78761938
I believe the Xbox Live code has relations to the 360 from what I saw looking through it

>>78761968
makes sense if you think about the dates of the files. it's right around that era.

>>78761939
Yes, or course. Sorry for bothering you guys, i will do that. Is there something in thread that isn't in guide yet?

If you don't fuck know the diff between IA64 and amd64 you don't even deserve to working at anything with this, sorry we need good and people that know hardware not dump people. And this is exactly the problem of almost all bad programmers today they say I learnt C++ but they did not know anything about how a CPU work. We need good people

>>78762235
But i said myself about "read one text book" about language only + "studying in uni". I mean, i am not pushing my fixes on root in github, or something, there are people who can control my work when i'll submit it.

>>78761538
When we talk about AMD64, this is about a type of processor architecture, this is not limited to processors from AMD, Intel's i3/i5i/7/i9 range supports it too
AMD64 is also known as X86-64, AMD64 is not IA-64
see here https://en.wikipedia.org/wiki/X86-64

>>78763622
AMD64 called x86_64 as well is supported by Intel CPUs since Pentium 4! It was first found in Intel CPU in 2006 or 2007 and was called EM64T because Intel didn't want to call it AMD64.

All of them (AMD64, EM64T, X86_64, X64) means exactly the same - x86 architecture with long mode. It was advertised by AMD as easy upgrade to 32bit computing.

IA-64 is completely different architecture designed by Intel and HP and used only by Intel Itanium CPUs. This architecture is actually dead.

>>78763622
Thanks. I finally got an answer that stop me from being confused as fuck, i was just too young back then, actually.
>>78763856
And thanks again too. So, for winXP 64bit i also need an emulator of a dead architecture?

>>78763990
NO!

Windows XP is available for x86, ia64 and amd64.

>>78764028
No-no. I mean, i was using WinXP64 bit myself back in 2014, when i got my own new computer. I didn't want to get on win7 so much back then, so i find that option.

What i am asking about, is about build of that OS, what i need for it aside current problems of windows server 2003.

>>78764113
64bit xp is a mistake. It was unpopular and so there are problems with drivers. later Vista came out and vista drivers do not work on xp. Vista was first real and popular Windows on amd64.

Slightly off topic, Does anyone know if there has been attempts to make dirs, makefiles etc for the Win2000 source?

>>78764204
It was working fine in 2014 for me on my own hardware, even if time was lost already. Like, nvidia drivers, games, everything. While i heard 2003 server had problems with some programs for "general user", or "consumer" when it was launched, winXP 64bit can use all of 2003 server and XP stuff at the same time, what i also heard.

>>78764295
y was trying to compile the win2k source with the missing parts from opennt 4.5 but no luck yet.

>>78763990
If you wanna get the official build server yes, as AMD was only supported starting from SP1 and we have RTM code. I thin XP never shipped with IA64 as the it was only used in servers.

>>78764910
XP was shipped with IA64, I haven't checked it probs built on Server SP1 code

https://archive.org/details/WXPIFPPEN
https://archive.org/details/WXPIFPPEN

>>78765069
How do i search for leaks of a specific OS, like this?

Does IA-64 Windows launch x86 software?

>>78765714
yes with emulation layer but it was slow that was the big problem of IA64 that result in amd64 getting attention

>>78765787
>Emulation layer was slow
Understatement much? IIRC The x86 emulation on IA64 was so bad it was compared to a 100mhz pentium.

>>78765534
Google

bump

>>78765714
no ias64 is very different from x86

>>78767484
wow64 is used to launch 32bit software on both amd64 and ia64

>>78767749
The difference is wow64 acts like a pass-through on amd64, which simply switches CPU registers for an "emulated" process into 32-bit mode and redirects its path calls to 32-bit versions of system libraries. On ia64, it fancies an entire software x86 CPU emulator. Problem is it's slow as fuck and never meant to support "modern" instruction sets like mmx/sse family, which says a lot.

>>78767991
The IA64/Merced ISA was supposed to power the heavy iron mainframes of the era.

In the end, Itanium ended up being a kludge. On paper, it's far more powerful than x86 at the time, but once it was finally deployed -- as Itanium 2, because the first issue CPU's were bugged and plagued with issues -- Microsoft had already spent the last 2.5 years getting it supported.

The x86 emulator, provided by wow64, was crucial to getting IA64 working on NT, but wasn't intended for use by the end-user server admins for anything outside of built-in applications.

They spent most of 1998 through 2000 getting IA64 implemented, thinking it was going to be _the_ 64-bit platform to back for the next 20 years.

EMT64 and AMD64 were competing platforms that didn't start showing any silicon until around 2002-03. Intel's implementation wasn't as complete as AMD's - they ended up rolling AMD64's ISA into Intel's EMT64 ISA, to create the x86-64 spec that we all use today.

It's this ISA - the x86-64 spec, referred to as AMD64 within most OS's to this day, that we use on modern OS's. The ISA includes hardware support for earlier x86 ISA's, but not IA64, since it was never an x86-based ISA. It's also Intel only.

In the end Microsoft had to adapt NT to support AMD64, which came AFTER XP was released - this is why Windows XP Pro 64-bit for AMD64 came out based on Windows Server 2003 code.

It's interesting to see in this leak, the Simics tools used to simulate AMD64 hardware on top of x86-32. There's a lot of neat tidbits hidden in the leak, as it comes from a time before SP1 - which boasted full support for AMD64 on client PC's (xp 64-bit).

File: file.png (171 KB, 1626x956)

171 KB PNG

>>78748495
I am checking this out later to make Windows XP run Windows 10 software: https://github.com/tumagonx/XomPie
>Basically XomPie forwards missing function/API into backported ones from various newer freely available MS runtime dll or simply reuse (reroute) into existing system dlls. When both options not possible XomPie borrow implementation from Wine or make them dummies (if preferable).
>Example of "running" apps: Python 3.5, Adobe DNG Converter 9.6, MS Visual True Type
I almost got it working for a handful Win32 functions, however KERNEL32.DLL is not linking yet.
It would be useful for add a more complete set of functions later on.
I am considering making a guide after my PoC works, as it is more involved than replacing a couple of files. These newer .c/.h will need to go into an separate extras folders to generate a .lib which is linked with the .DLL in question. Later on one could add it could be moved to the main folder.
Already removed the PE header version in KERNEL32.DLL check, added _except_handler4_common in MSVCRT.DLL & MSVCRT version check in MSVCIRT.DLL.
The system boots stills fine after these replaced .DLL's and can already run write.exe (the launcher for Wordpad only as was found out) from Windows 10 on Windows XP

>>78768353
> the Simics tools used to simulate AMD64 hardware on top of x86-32
Interesting, I saw the simics crap in here but never bothered to look it up, seems simics is publicly released, maybe it could help the amd64 anons out

>>78770015
Stuff happening after ntldr debugger goes, but before windows windbg support gets enabled can only be debugged with Bochs/IDA, or something similar to it, like AMD Simnow or Simics.

>>78770280
i wish someday we could debug it in an IDE step by step with f10

File: windbgx-main-menu.png (99 KB, 1115x738)

99 KB PNG

>>78770524
>debug it in an IDE step by step with f10
You kinda can with windbg, if you have everything setup right it'll read symbols from .pdb, and can even let you do source-level debugging in some cases.

Mostly meant for kernel-mode things though, but there's some tricks for enabling user-mode debugging through it too: https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/controlling-the-user-mode-debugger-from-the-kernel-debugger

The new modern windbg barely even needs you to learn any kd commands neither, though I've heard it can have some issues with older kernels, I didn't have any problems with 3790 with it tho

>>78770816
It'd be really nice if we could get KD working in our VMs over something other than serial though, not sure if any network KD support is in this code, IIRC they mostly used firewire at this point which I'm not sure if any VMs can emulate/passthru at all...

>>78770880
>Kernel Debugging over LAN cable for Windows XP/2003 x32
https://github.com/MovAX0xDEAD/KDNET
Neat, seems win 8.1 kdnet file works on XP/2003 with some patching. Maybe we can edit the XP/2003 kd code so that patching kdnet isn't needed, could just be something like the MSVCRT where some export is missing.

>>78770880
network was introduced in win8.... I think add it require a reasonable kernel change ): yeah com with pipes are slow

>>78770955
Sucks they took so long to add network support, Xbox360 had kdnet since it launched IIRC, didn't see anything in xbox OG trunk about it though sadly.
I guess it probably took so long because of the different network adapters or something, seems that github only supports a few types, hopefully some VM emulates one of them

>>78771003
well they started on xbox because it is a console, almost all devkit use ethernet for connection and game deployment during development. At least XBOX360/PS3 do this

Where are the boot screen images kept in the src?

>>78748378 (OP)
What's the state of documentation? Are patches kept in one place?
Should I bother to compile it with the instructions from the guide?

>win2003_missing-ex-files_v1a.zip
https://gofile.io/d/7CFt8U

Edited systime.c to include the two externs it needs, and reverted exp.h back to the original, should build fine now hopefully.
If you're using the prepatched pack from build-win2k3 you should probably extract the exp.h from here over it, so you can revert your copy to original too.

>>78770524
This still can't debug early stages. That's, as I said, after NTLDR/BootMgr debug goes down, but before KD kernel support gets enabled. I've heard an anon here had problems debugging such stage in a past thread.

You can't debug bootsector and early Bootmgr bootstages this way too. UEFI doesn't have such problem if mobo vendor built in EFI debug support.

>>78770955
See >>78770953

Anyone here know if minlogon has a chk build + symbols available?
Which editions include it anyway?

>>78771552
Tiny update for systime.c, makes symbols section in obj match up when viewed with cvdump.
Code itself hasn't changed at all though, just .obj file metadata shit.
There's still a ~8KB size difference between the .obj this builds and the original, can't explain that yet, maybe they included some extra header files or some dumb crap.

>win2003_missing-ex-files_v1b.zip
https://gofile.io/d/Wr5oC8

Unfortunately exinit.obj symbols section still doesn't match up, that section in the original .obj is kinda weird, not sure if I could ever get it matching with that, oh well.

>>78772779
Oh damn, seems amd64 had some pointer-size issues breaking the build, fixed them here. x86 build should still be identical.
(this reminds me, I should probably try comparing IA64 version...)

>win2003_missing-ex-files_v1c.zip
https://gofile.io/d/X34wNH

>>78773300
Compared against IA64 + chk x86, found a couple tiny differences, and some asserts that were missing, need to do a full chk build and compare something first tho
(couldn't properly compare IA64 code though since nothing seems to handle IA64 .obj files ;_;, I'll keep looking around though)

>>78771413
>Are patches kept in one place?
No. You have to look through the threads to find things. The systime/exinit things have been linked in this thread, and the winlogon stuff was link in previous threads. If you want to find something, find some sort of 4chan archive and search yourself.
>Should I bother to compile it with the instructions from the guide?
Yes. It is up to date. Not much has changed recently, we've got it under control, so there's no need to change.

>>78773745
Alright got debug shit sorted (extra asserts + debug crap in IsRegistryKeyLocked)
Also found what the size difference was about, seems building them as dirty-builds makes the .obj files almost exactly match size with the original ones (systime.obj is ~11 bytes bigger than original, exinit.obj is ~345 bytes smaller)

>win2003_missing-ex-files_v2.zip
https://gofile.io/d/mxc9aR

Last update for a little while, maybe tomorrow I'll check out IA64 & AMD64-SP1 ntoskrnl.exe to see if the funcs are any different there.

>>78774225
Updated exinit.c/systime.c section in the guide: https://rentry.co/build-win2k3/#systimecexinitc

Winlogon link for those that don't feel like going through old threads: https://gofile.io/d/LwFdnI

So I read over the wiki that >>78763622 posted, I'm totally unfamiliar with architecture and shit. Any good reads on the general information on it or does fucking wikipedia have the best somehow?
Also, as a neet, I'm willing to learn what's needed to help in this project.

>>78774772
AMD64 = x86_64, 64 bit. Any modern desktop or laptop computer probably uses this. Exceptions being old (before 2006) PowerPC Macs and ARM Chromebooks.
IA-32 = x86 = i386 = i486 = i586 = i686 = i786, 32 bit. Used to be the standard before AMD64/x86_64.
IA-64, also 64 bit. It's its own thing, separate from i386 and AMD64. It was originally meant to be for servers, which is why old versions of Windows NT support it. It ended up being shit, so Intel scrapped it.

we should drop the IA-64 support shit and all the if statements related to NEC_98 too

>>78775633
Not a great idea. That would result in noisy diffs. On change stuff that needs to be changed.

Bump

If I want to build just ntoskrnl and hal and nothing else how can I do that?

>>78777148
base\ntos
base\hals
run bcz or bz in these folders if you get errors build these dirs:
published \
dload \
dloadhandler \
strsafe \
wdmlib \
seaudit \
crts \
urtl \
ntos \
hals

>>78777368
Thanks

bump

>>78774772
>wikipedia
how-stuff-works actually has introductory information on these topics. Check these out on (micro)processors and their internal workings - they feature also an introductory article on C programming. Most of Windows XP's source code is written in the C language.
https://computer.howstuffworks.com/microprocessor.htm
https://computer.howstuffworks.com/c.htm
And here is a nice overview about the internals of Windows NT/XP/2003
http://www.cs.sjtu.edu.cn/~kzhu/cs490/3/3_Win-Structuring.pdf
This is useful as you find a lot of it referenced in the source code

>>78773965
Thanks for info.

>>78748378 (OP)
Will Microsoft notice the increased number of NT specialists on the market?
Or will there be just more forks and Microsoft will get nothing?

>>78778688
Anyone working on this will never be able to work for Microsoft or any official business in NT/Windows development

>>78778688
Microsoft will always get nothing lol

>>78748378 (OP)
So what are y'all doing with the leak? Can it compile and run yet? Will this help make Wine/ReactOS work better?

>>78778778
We can compile and run fully working Windows XP/2003 from first weeks after leak. It won't help Wine and ReactOS any way...

>>78778778
Yes, ReactOS will be better thanks to that!

>>78778925
lol

>>78778778
This leak is only bad for reactos, it just spreads more poison around. If there was no windows source code leaks reactos would probably be complete already.

>>78778948
There was a lot of leaked code in ReactOS before this leak. It can only help them.

>>78778948
>>78779011
I wonder how it really will be with reactos. I personally hope for a big leap in development.

>>78777695
>http://www.cs.sjtu.edu.cn/~kzhu/cs490/3/3_Win-Structuring.pdf
It is material from a course about Windows internals: development / debugging / architecture.
>Introduction, OS Evolution and Reviews
>Windows Concepts and Tools
>Windows Structuring
>Core System Mechanisms and Windows API
>Concurrency: Windows Traps
>Concurrency: Windows Synchronization
>Windows Processes and Threads
>Thread Scheduling
>Memory Management
>I/O System
>Windows Security
Index is available here:
https://web.archive.org/web/20130127005412/
http://www.cs.sjtu.edu.cn/~kzhu/cs490/

>>78778764
Nope. This is only poisonous for MS competitors, FOSS or closed sourced.

>>78778777
>>78778764
Dumb Microsoft. NT experts train themselves for free, but they're going to waste it all.
They should seriously switch to Linux.

>>78779360
will never happen

>>78779360
>Linux

Not this again..

>>78779619
Yes apparently shit maroon are back trolling use, fuck you linux fuck you all linux fun boy. With love

>>78779554
Maybe, meanwhile Google owns the second half of the global OS market share using the Linux kernel, yet devices are locked down and Google keeps control over the ecosystem.
It's dumb not to use the cheaper kernel.
If maintaining becomes more expensive than porting their shit to a new kernel, they will switch.

>>78779554
they don't listen broh it's like speaking at the wall. You know 90% of software run on windows and you have to get 100% compatibility. MS cannot even guarantee compatibility when they do updates....

>>78779673
stop this bullshit, will never gonna happen

>>78779673
>meanwhile Google owns the second half of the global OS market share using the Linux kernel...
Yet they want to move out of it with Fuchsia. And not only them...

>devices are locked down and Google keeps control over the ecosystem.
But they don't have all the control they want, while their competitors can take advantage in the code Google develops. They, and their partners/investors really don't like this.

You have to see really dumbassed if you can't see it happening. Google already have what they wanted from Loonix. They don't need it anymore.

>>78779777 (trips)
>>78779668
>>78779619
>>78779619
I'm not going to argue if you don't want to, but keep in mind that with this kind of attitude you're more likely to start flame wars and trolling.
I'm especially talking to the "fuck you maroon" anon.

>>78779867
Now your arguments sound reasonable.
Do you think Microsoft will also make their own fancy microkernel or are they going to keep NT forever?

>>78779949
They tried already. They will stick with NT as it is really good architecture.

>>78779673
>If maintaining becomes more expensive than porting their shit to a new kernel, they will switch.

MS will never switch to Loonix because they can't completely control the shit in the ecosystem. All the foundation of MS relies in this simple fact.
And their partners/investors trust in the capability of MS to do it. When move and when stop moving. Partners dictate the way NT and Windows moves, and MS just obeys them. They can't do that with Loonix, because there are many 3rd party interests involved. What MS and their partners want wouldn't be what their competitors want.

So they will not be moving... To Loonix at least.

My build got stuck just after wizard.cpp - any help?

>>78779949
NT is adaptable enough. Unless there is a really big departure in general PC hardware architecture, they will keep bringing NT as far as they can.

>>78779898
I can agree but you realized that a lot of people are writing this stuff because the wanna troll right? If you look back all the threads than you realized why I said these words. We spoke of this stuff since weeks and yet randomly someone appear and said what this mean for ReactOs? And than they begin with the linux stuff. it's always the same so it is natural to think that it is just trolling, so fuck to all of the trollers than.

>>78779949
You realized MS spend millions with Midori and than give up? They had already tried to get ride of NT but when it come back to compatibility that is the hard part. Midori was a managed kernel, and yes that already was a bad idea to start with.

>>78780305
>You realized MS spend millions with Midori and than give up?
That abomination written in C#? I'm not surprised.
Even Google writes their kernel in C.

>>78780305
Isn't only compatibility. NT is high level and standarized enough to support Win32 developments, but also low level enough Hardware Vendors can do their nasty voodoo with the OS for their hardware development. And blindminded don't realize that, with UEFI/NT, actually MS controls the whole personal computer ecosystem. Actually, if competitors want to play with personal computer platform, they have to adapt to the UEFI/ACPI/WHQL trinity plus DirectX, not the inverse.

Microsoft is more likely to keep developing the replacement for Win32 than replace NT entirely.

The first iteration of non-win32 Windows was Windows RT, which then became a library of its own - the Windows RunTime. This eventually became the basis of the Universal Windows Platform.

WinRT is the managed code replacement for Win32, and is the primary focus of the Windows devs at this point in time, because it is supported on all Windows platforms, provides the same functionality across mobile, desktop, and game console platforms, and is full of buzzwords.

We might want to look into supporting UWP binaries and/or WinRT as a separate NT personality.

>>78780441
To be fair, it actually was good in paper. But when it got presented to IHV/ISVs for review, they just said "ewwww, nop". No sane vendor likes sandboxes they aren't allowed to manipulate to their will.

>>78780467
>We might want to look into supporting UWP binaries and/or WinRT as a separate NT personality.

They can try, but it will keep flocking as many intents they do. ISVs choose MS because Windows ability to be manipulated as their heart content, while keeping stable API across 3rd party software. ISVs who choose Windows don't like sandboxes and stores. When they want something like an store and sandboxed "simple" apps, Google is the king, without question.

>>78780441
yeah that think.... it was not C# but still was a bad idea from day0. Because you guess what it one the most good think ot windows, the kernel and that is the thing still write in C. :)

>>78780454
also people don't realize that all the WSL thing is the MS move to kill linux (:

>>78780573
Yes, that's how it is. That's the truth behind WSL. Dreamers want to think is a move to Loonix, but actually is the contrary... Is a move against Loonix. Sooner or later they will realize their fault, but then will be too late.

>>78780597
Yeah and that was a smart move from MS to be honest. Linux will be only used in embedded application where it really shine because there basically nothing else. For desktop it will become MS and apple but eventually also Apple will fall if they continue to be soo shitty at developers.

My build is stuck just after wizard.cpp - any help?

>>78779011
>>78779088
I have worked a little bit on reactos, and what I saw disgusted me. The thing is full of race conditions all over the place, much of the code is half baked. You would not need access to the source code to fix reactos, much of the code in windows is trivial and your could pretty much complete reactos from dissasembly. The source code leaks make reactos suspect, if some code looks the same as in windows then maybe it was just copied. Reactos seems to have to go out of its way to make the code different and it just ends up a half baked broken mess.

>>78781032
>The source code leaks make reactos suspect, if some code looks the same as in windows then maybe it was just copied.
Script kiddies keep chanting this forever, yet they don't offer any substantial proof. Just suppositions and conjectures. Do you know every ReactOS line of code has an specific dev name attached to it? If you have any proofs of their cheating, you can always indicate the exact block, with all the devs involved in making it together, with a simple git grep. This simple fact is the reason they don't allow anonymous commits.

>much of the code in windows is trivial
Lolz, yet people here is having a hard way to reconstruct... Let alone modules and complete files, but just object files. With the code in hand and all...

>>78781360
Most Alex code is illegal, so yeah they are cheating.

>>78748378 (OP)
>>78749343
>>78749650
why does she have orange juice and strawberry jelly on her legs?

>>78781631
You still have not provide any proof nor evidence, broh

>>78781677
Think anon, think!

>>78781677
https://github.com/reactos/reactos/blame/44c6df4b2769b2b2d7aa62b589625e1a5b87338d/ntoskrnl/ob/oblife.c#L31

>>78781360
I may have poked around the win2k source code leak and compared it to reactos code, some code matches line for line. This is not proof that it was copied from the source, but since the source is available to anyone with a simple google search that makes such code suspect and there is no way to ever prove that no copying was done.

Windows code really is trivial to reverse engineer, for christ sake it even comes with symbols and checked builds are available, its all there on a plate. Remember windows is modular, you do not have to reverse engineer the many gigabytes of worthless garbage, only the core system files are important.

>>78774225
Is it okay to include this in my Easy-Build source patcher?

>>78782599
You are absolutely right, look at example above. Why did Alex introduce that variable? It's never used except it is incremented - same as in leaked Windows code. No purpose for it. Well, I can image the purpose Microsoft had, but why did Alex done the same, even copying its name? ... Just look at his presentation about reversing without reversing, just think he worked on Tinykrnl as well and Tinykrnl is illegal. What did he tell at ReactOS interview and why this interview has been removed from reactos website? I will quote it: "It’s expected that TinyKRNL will share a lot of its kernel and HAL code with ReactOS". I think no comment is needed here, just WAKE UP anons!

>>78782599
>but since the source is available to anyone with a simple google search that makes such code suspect and there is no way to ever prove that no copying was done.

That's a conjecture dude. Unless you can prove it, you can't make it fly, and probably you may be demanded by defamation

>>78782880
If you feel so sure about all what you are talking about, then go and fight legally reactos. I'm sure MS would pay you greatly for your efforts.

>>78782599
>Windows code really is trivial to reverse engineer, for christ sake it even comes with symbols and checked builds are available, its all there on a plate.
Yeah yeah... Go and tell it to our anons who are working hard in reversing winlogon and systime.c/exinit.c, straight to their face. I'm they will have a few words for you...

>>78781631
Then fight Alex. Prove his code is illegal, and make your case fly. Until you can do it, STFU.

>>78781030
started again from scratch, seems to have gotten further this time.

File: mspaint.png (73 KB, 1008x878)

73 KB PNG

how do i compile ms paint

>>78783774
not using VS, razzle command line open the folder bcz.

>>78783845
can it be done in windows 10 or do I have to install XP?

>>78783859
Read the fucking guide

>>78784154
I apologize for my retardation

>>78783119
I don't care about ReactOS. It's simple piece of shit.

>>78784556
Eh I like ReactOS as a project. Said that they're gonna get no where ever without a big amount of money going their way.

>>78784591
Sad*

>>78784591
Its foundations written among others by Alex are illegal and current devs can't fix MM for last 10 years. buahahahaha

>>78782675
Prepatched ZIP will be updated with it soon so there might not be any need, it's going to be optional inside there though, so if you like maybe you could add an option for people to use it or not.
To enable using it you'd just need to delete the systime.obj/exinit.obj files like the missing-ex-files readme says, and then rename base\ntos\ex\exinit.c.dec & base\ntos\ex\systime.c.dec to remove the .dec extension.

The current prepatched ZIP has exinit.c.dec files based on the older exinit recreations, can probably test against those until the ZIP gets updated.

File: 1600765934522.png (543 KB, 809x867)

543 KB PNG

>>78784642
I still like it, so even if there's stolen code from years back, I still like what they're doing. All I want is that they get a good amount of money at some point so they can be able to get it to be somewhat stable and functional.
Also, why are you complaining about illegal code in a thread about fucking with Microsoft's copyrighted code that's illegal to have as well, are you retarded or can you not read?

>>78784642
>Its foundations written among others...
Blah Blah Blah. buahahahaha

>>78784746
Don't waste your time in trolls like him... He seems to have something against ReactOS... Who knows, maybe he got kicked by Alex or even a Mod there because something stupid he did there, and don't want to accept it.

>>78784741
Right okay, I think I will wait for when it's in the prepatched zip incase there are any last minute fixes that, and to cause less confusion to some people as its such a small patch in retrospect, and if its optional in the prepatched I will most likely follow that and make it optional for ease


I haven't actually got round to testing exinit/systine myself yet I was going to tonight

In the XP leaks there are some Microsoft Certificates valid until 2042

Anyone knows if microsoft revocated them?
Also was it before the /g/ leak, right? because they where already circulating

>>78784741
>it's going to be optional inside there though
I'm in two minds about this though... right now if we decide to keep using the original .obj files that came with it, those .obj files are from a fre build, so if you build a chk kernel it'll be using fre exinit/systime
Seems the actual official RTM chk kernel actually used chk versions of those though (since I guess they never had to bother with pre-built exinit/systime), so any chk kernel we build isn't going to match.
Our systime/exinit recreations do have all the chk stuff included now though, so using those in a chk build should allow it to match with official RTM chk (well, the systime/exinit parts at least)

AFAIK with our new recreations both fre & chk should match fine when inside ntoskrnl.exe, so maybe it is time we get rid of the .obj files it came with...

>>78784874
Not sure about the certs, but the MS zone code from in this leak was circulating long before the leak came about, so I wouldn't be surprised if some other things were too.

>>78784906
Exactly

I'll search for the revocation certificates

>>78784746
I tell this to you anons, because you you are half-brained and you believe it was written using clean room reverse engineering :D hahah

I don't care about this shit, it is not working. They are a bunch of idiots who cannot make it stable. From two illegal things I prefer our XP because it just works, what can't be said about reactos.

>>78784886
I think it would be best to make it a part of the prepatch zip then, less barriers with issues for users wanting to chk build.. maybe after a thorough check that at least x86 chk and fre can be built with the systime/exinit no issues then it should be safe to commit

>>78784948
Of genius make the same think yourself and let the world know how good you are at doing stuff the right way. Otherwise shut up and don't fucking bother us you are an idiot who criticize the others and do absolutely do nothing, better shut up if you have nothing to show. At least they did something ok it crash but is better than a fucking hobby OS that is not even multithreaded.

>>78785009
In theory a chk x86 build should go smooth as fre right? I've not even done a chk build myself so I will test that later when I get round to it

>>78785040
When I tried chk it built fine, but BSOD'd on boot, dunno if it's related to using fre systime/exinit or not though, one anon mentioned you needed a debugger attached to skip some assert, but that's strange to me since official RTM chk never needed that

>>78785040
yep you only need a debugger attached because after login an assert is firing that if not ignored in debugger make a bsod

>>78785011
Yeah this hobby os is at least stable!
What did they do? I can copy-paste leaked code as well and they couldn't even do that properly.

Just stop talking about it. This is Windows XP/2003 Leak thread, not reactos, so don't mention this piece of shit here.

>>78785070
to be honest we do not know if the code we have is rtm or not. That assert happen when I build a xp professional build from 2003 could not happen in server sku.

File: lord.jpg (22 KB, 200x200)

22 KB JPG

>>78784948
No, I don't think that every single line of code of ReactOS is clean room reverse engineered, it is almost impossible to find out what is or isn't stolen code without a full audit by a lawyer and a software developer who knows the NT kernel. You're acting like a retarded monkey by going back and forth and pumping your chest. Fork ReactOS and do what you want with it (removing the illegal code you hate) or use your own built XP image. You're making it seem like ReactOS fucked your mother and left her high and dry, fucksakes man, get a grip on yourself.

>>78785070
>but that's strange to me since official RTM chk never needed that
Maybe it's because our missing.7z files are all fre versions, fre winlogon etc. I hope someone makes a missing.7z chk variant some day

>>78785084
Stable but unusable lol. The problem is that to make an os you have to have a lot of money as it is an hard job to do. What the run a console?

>>78785084
I stop talking when people stop come here and ask for rectos because they are trolling.

>>78785092
That is also a good job, I would do that I have no idea where to start we have checked rtm iso indeed.

>>78785070
Right okay I'll setup windbg later and see if it does get through it.. if this is the case, it's a minor setback as most users building chk will be mainly debugging anyway I'd presume

Regarding Windows XP's Windows 10 software compatibility, still a work in progress to get the code to work functions to KERNEL32.DLL. Working on that tonight.

One thing already works now: C:\Windows\SysWOW64\write.exe from Windows 10, by applying the changes in the first link below.

While I am working on an entry page for the guide.. these pages/links already exists:
To enable Windows 10 software on Windows XP/2003: https://rentry.co/kernel32-version-unlock
To add structured error handling (SEH) using the _common_handler4_common function in MSVCRT.DLL that Windows 8/2012 started using: https://rentry.co/msvcrt-excepthandler4common
To be updated when I get around adding KERNEL32.DLL functions through a .lib file: Encode/DecodePointer functions used by XP2SP2+ applications:https://rentry.co/kernel32-obfuscated-pointers

>>78785084
>What did they do? I can copy-paste leaked code as well and they couldn't even do that properly.
Unfounded claims... Unfounded claims everywhere.

>>78785117
Stop answering them. Don't feed trolls.

File: file.png (57 KB, 959x365)

57 KB PNG

>>78785143
missing.cmd was used to make the missing.7z for fre, or at least a part of it, that only has files that were missing in fre though, maybe chk has some different files missing

>>78785143
Best bet is to find some chk XP and Server ISOs (the first result on a quick archive.org search was an XP sp1 chk iso, not sure of the build number)

Then mount the ISOs one by one, running the missing.cmd as a basis to see what they pull off

>>78785167
just post a zip with changed files

>>78785214
Forgot to add link to my example
https://archive.org/details/1849.0-os-2002-11

>>78785086
we should check what we got

>>78785203
Yeah I should try maybe I can try to dest a diff folder so I can zip it later. We know if this missing is the last version? I know we went through a lot of changes to this archive

>>78785266
Last version that was released afaik, but the missing.7z we use now had a lot of things added manually by some chat-anon, not sure if they still come to this thread now tho...

I think missing.cmd gets maybe 80% of things it needs, the rest can maybe be tracked down from error log messages.

>>78785220
Sure. First I'll add that handful KERNEL32.DLL functions too, so that it can run more applications, than just write.exe

>>78785307
Yeah that was what I was worried about. Do you know from which iso they based the file on?

>>78785347
IIRC they came from all the ISOs listed in the guide, except the amd64 one: https://rentry.co/build-win2k3/#original-cd-filenames

I'm not sure if chk builds exist for every edition though, might only be enterprise/professional, at least the only chk one I got for 2003 is
>5.2.3790.0.srv03_rtm.030324-2048_x86chk_server-enterpriseserver_retail_en-us-NRMECHK_EN.iso
Maybe there's more out there I didn't find though.

>>78785456
If we have other people to get banned at winbuild for sure we will have all the needed iso ahahah :)

>>78785491
I mean betaarchive of course

Who knew Windows Defender 'Real Time Protection' slows down my build time by 2 hours

File: file.png (40 KB, 1686x184)

40 KB PNG

>>78785803
we tried to warn you bro

>>78785830
Oh I know I was just saying it's crazy how much it does lol

>>78785803
And you can even get an extra if you disable Spectre bandaid.

>>78786062
What's the equivalent for AMD or am I wrong in thinking that was an Intel only 'fix'

>>78777695
>>78779208
Thx, based knowledge anons

>>78786114
InSpectre tool will tell you if your processor has spectre/meltdown bandaid enabled, amd or intel.

>>78786298
Cool thanks I will give it a look

File: file.png (23 KB, 408x124)

23 KB PNG

>ntvdm64.dll
What's this?

God damnit, trying to check our IA64 build against the released one to see if there's any difference in exinit/systime, but I can't get hold of the symbols pack
http://web.archive.org/web/20051020023630if_/http://msdl.microsoft.com/download/symbols/packages/windows2003/windows2003.ia64.fre.rtm.symbols.exe

Just keeps cutting out at 100MB, should be 105MB, incomplete one seems to corrupt the symbols when I try extracting ;_;
Is anyone able to download the full pack and upload it somewhere?

>>78786763
nvm, seemed to work after a few tries
IDA has issues loading the pdb though it seems

>>78786719
Ah lame, seems it's just for redirecting certain 16-bit apps to ported versions

>>78785803
Windows Defender is just cancer, it automatically removes allowed files after a week taking for granted all users are retarded and don't know what they are doing. Slowing your PC to a crawl is just added creme.

I've been searching old threads and can't find any info on it. Does anyone have a link to the encdecptr.7z that was mentioned in the guide?

>>78787292
should be the one linked at >>78111719

>>78787346
thanks

Bump

bomp

.pmuB

Anyone looking into the chk missing.7z atm?

I'm like 99% sure I saw a VS addin posted by a MS employee that let it open build.exe sources files, or convert them, no luck finding it now though

File: Converting_Fig2.png (34 KB, 588x363)

34 KB PNG

>>78790080
Closest I can find now is
https://docs.microsoft.com/en-us/windows-hardware/drivers/devtest/converting-a-wdk-sources-file-to-a-visual-studio-project
>Converting a WDK sources file to a Visual Studio project
Talks about using NMake2MsBuild to do the conversion, not the VS addin I saw though...

Also found https://www.osr.com/nt-insider/2011-issue3/converting-sources-based-projects-vcxproj/
which mentions
>In Visual Studio 11 you can skip doing this conversion from the command line and do the conversion inside Visual Studio via the “Convert Sources/Dirs” option found via the File Open menu item as shown in Figure 2.
pic related, I don't see that option in VS2019 though, maybe WDK would install the addin for it or something? Or does anyone here have "Visual Studio 11" installed and can check for that option? (looks like VS2010 to me)

>>78790080
Well alot of filesnare stamped Microsoft Developer Studio or something like that.. I this VS 6 had references to that but don't quote me

>>78789730
Seems not, I'm 1/2 way through an x86chk build now, I'll try postbuilding when it's done and see what missing files it complains about, if anyone likes I can upload the lists afterward

Has anyone tried replacing compilers/libcp.lib etc with ones from SP1 DDK?

bump

>>78789730
>>78790191
Seems the missing.cmd already mostly works for chk, just missing the following files
>srv_info.chm
>admtoolP.chm
>admtoolW.chm
>bootconP.chm

I'll add those in with next prepatched update, along with the new exinit stuff

>>78791263
If anyone wants to make a missing 7z for chk, just add these to missing.cmd:

if not exist "%_NTPOSTBLD%\srv_info.chm" expand -r "%_RTLISO%\I386\srv_info.ch_" "%_NTPOSTBLD%"
if not exist "%_NTPOSTBLD%\srv_info.chm" expand -r "%_RTLISO%\AMD64\srv_info.ch_" "%_NTPOSTBLD%"
if not exist "%_NTPOSTBLD%\srv_info.chm" expand -r "%_RTLISO%\IA64\srv_info.ch_" "%_NTPOSTBLD%"
if not exist "%_NTPOSTBLD%\admtoolP.chm" expand -r "%_RTLISO%\I386\admtoolP.ch_" "%_NTPOSTBLD%"
if not exist "%_NTPOSTBLD%\admtoolW.chm" expand -r "%_RTLISO%\I386\admtoolW.ch_" "%_NTPOSTBLD%"
if not exist "%_NTPOSTBLD%\bootconP.chm" expand -r "%_RTLISO%\I386\bootconP.ch_" "%_NTPOSTBLD%"

Then run missing.cmd on (in order)
>the x86chk server2003 + x86chk XP pro SP1
>the x86fre server2003 non-enterprise editions
>the x86fre XP SP1 home edition
>the amd64fre XP x64 edition iso (for srv_info.chm)

If you did that with an empty binaries.x86chk folder you should get 4333 files inside, delete the UDDI.cat as chk build seems to create its own one fine, leaving 4332 files, then you can 7z the contents and post somewhere.
(I'd do it myself but it'd take me hours to upload, really hope someone here is able to...)

This'll only fix the filechk errors though (i.e. layout.inf files that are missing), other postbuild stuff that still need missing files will still fail, gonna look into those later.