TuMaGoNx

My browser is Opera 12.18, use opera turbo, rarely turning on javascript and image. So fasttttttttt But hey back then I use palemoon atom for a while, here is normal palemoon (patched using xompie) since I don't know how to tell the difference, i show the inspector which atom claimed missing? I think. good night. gotta sleep EDIT: note that this is just "run" app, no heavy test/benchmark been performed to claim fitness to run on XP, I'd advise to use Atom+Firebug, the proper way to run on XP.

sorry I didn't meant belittle your works and spirit, me myself have experience to keep webkitgtk port available on win32 but with every minor version come up with huge changes, new API and refactors. Something that one silly men like me very easy to get lost... and eventually given up. of course you might be good enough to make it. who knows. "Then create a automated patcher at last . I dont think people will be happy in carrying cff or hxd." I use can only open source apps with open source project. I use gsar to replace string in binary, at the moment the mass-patcher (without signature unsginer) and dlls should be under 64KB compressed. not yet make 7sfx installer.. https://en.wikipedia.org/wiki/Google_Chrome_version_history I can't say they into removing stuff than adding stuff, the fact that you're still at version 50 show that you do face difficulties. As part of development cycle of course they would remove "redundant" code, to keep on track and freed the roadblock (that's XP support) to implement something "better". That's normal, like the move on MediaFoundation have you thought about it? how you backport it? This is what I immediately see about the current of Chrome: they "really" are moving away, fast. It's true that new features doesn't always meant new requirement Firefox for ex. are hanging on XP partly because they have netscape legacy codebase. But chrome is exception, this thing is so wild to begin with. https://en.wikipedia.org/wiki/Backporting Just realize it now, gotta to change my project desc. Oops.. example of backport is: WIC, WSSAPI, RSAENH, PowerShell and so on. What we do is extending OS.

@Dibya: "it is easier to patch known dll system" I've read the registry, to me this is as hackish as hooksubsystem :) (with system-wide consequence) I don't want to touch the system as much as possible. It's more "normal" to me to patch to target apps as I didn't aim non-trivial purpose with the shims. Just for those apps that "miscompiled" without XP support enabled in VC++ and those that "artificially" drop XP by using few missing APIs. "Crome based browser are blocked in xp , I have run crome and opera with xp compability settings same error shown under 8.1." You should just download and read the chrome source code rather than "probing" what happened underneath. It's quite clear what's the new features and requirement. So what if you could run few version of it, as I said browser evolve quickly with hundreds of developer backing up.

@Dibya: I think targeting "fast paced" technology such Browser is impossible... it's always employ cutting-edge ones (virtualization, memory management, MT) thats beyond XP scope. If any it would be more and more "disabled features" if it allowed to run on XP. BTW I think I'd abandon patchless idea.. hooksubsystem itself is hackish and might be restrictive, I can't use redirection manifest either as having another "kernel32.dll" often trigger DEP. Not to mention I get chicken-egg situation as my dll (the xp kernel32 part) is just forwarder not even wrapper, the redirection manifest would meant any dependencies (dll) would call the fake kernel instead of just the executable.

BTW Dibya, I think the redirection manifest may be better suit your case than mine, have you tried it on unpatched XP with your modded kernel put in the same directory of the executable? @FranceBB: if what you mean is Wine, I already knew... I would say even Wine don't claim any compliance yet it's still WIP. Have you test the conformance of WebGL?

"you can still make the project open source if you just post the code for the implemented functions and scripts that can patch in the functions on an existing XP system" tspatcher and uxtheme realtime patcher are open source too. I rather confused to read the topic title: compatibility *layer*, I thought we have same goal and way at first about opera/chrome: how you could be certain about "adding" unported feature such D3D11 and MediaFoundation? I'm no ASM guy though...

@Dibya For new Opera/Chrome/Blink/libCEF and others, isn't that thing rather not trivial :gulp:? d3d11, new MediaFoundation (delayload) no idea..

@Dibya Thanks a lot! I admit at first I'm baffled by this https://www.ethicalhacker.net/columns/heffner/intercepted-windows-hacking-via-dll-redirection Make me wondering why XP attempts of that route (wrapper) has been few (sorry if I missed) considering it is much more popular than win2k, is there any particular roadblock other than "there would never be a perfect functional wrapper" in the sense that even wine get crashed a lot? though I never envisioned to see it works beyond the trivial cases. Whats your reason not to choose that route?

@Dibya In my case, GetVersion issue will meant it need intercepted instead of forwarded directly. Then we could lookup says registry settings whether an apps need to be tricked or not, thus conditionally set version. Because the function is not always a blocker, for ex. it could determine if DWMAPI enabled by default or not (windows > 8) and other case.

Ah so this is what I need: https://code.google.com/archive/p/ollytlscatch/downloads?page=3 with full source code as well sweet!. With this and dll redirection manifest maybe possible to make dll shims without tampering system files and target executable..

Yeah I'm don't mean to endorse (just saying fact). XP of course less secure, every newer kernel is bound to be better. XP is the era when connectivity not (yet) assumed something that I appreciate (e.g. XP have offline documentation). Moreover with internet getting wilder, security expert even make Browser+VM bundle or at minimum Sandboxed, isn't that like not using OS to go online... make the value of secure OS and AV getting less and lesser. Once everything that run must be signed and certified and Win32 abandoned for walled garden framework, we can talk about secure Windows.

For me it doesn't matter "how to do it" but everything should be transparent and documented (including codecave), reviewable by peers and users. It's all about trust. In my way user had to know that they agree to take action and consequence (they actually the one who "patch" it). Me too literally grow with XP (heck, even internet cafe here still use XP+deepfreeze as defacto standard). XP also the only major OS that will operate happily being completely offline.

Ah that's too bad, might be hard to gain trust with just binaries. Guess i will go my way then.

@Dibya Do you plan to put your source code somewhere (like github)?

@FranceBB at least Win2000 src leaked, in my disk most of source code are in july 2000. But for project like this I prefer released as open source (can't trust anyone who modify system functionality) so Wine is good candidate. MSVCRT. those *_s are secure API came from VC 2005 I think and since then Vista and later have by default (in their msvcrt.dll). If you see how WDK7 works this special VC++ compiler only linked to system CRT and support multiple version of Windows by adding missing API via appended obj files (msvcrt_winxp.obj) during linking. XP of course based around version 7 (at least that's what MinGW says) BTW I found "missing" XP API update which I posted here http://opensourcepack.blogspot.com/2016/10/where-to-find-wssapi-backport-for-xp.html I think many overlooked it in 2009 (at least no updatepack contain it, maybe for legal reason) In Wine case, thing like ETW*, Event* or telemetry (those diagnostic stuff) simply dummied as their are not part of Apps functionality (just declare function as is and return "done")

Man... I wish I've read this first than trying myself... https://github.com/tumagonx/xompie/ What about using wine source code there a lot can be transplanted.

I think we should organize goals first (long term ones), not "let's do this because I think I can do this". Even BWC's long private research isn't that popular to end users. Four things that could made XP gain some popularity (or otherwise kill XP): 0. Security Issue Online XP Machine marketshare is about 10 percent, with POSReady workaround there is 3 years left. Consider that as deadline as McAfee said "Security is an Illusion" 1. RAM limitation There several decent patcher already, possible issue remained on third party drivers though. Personally this is the only part of system that I want to patch :) (since it can be load on demand by ntldr) anything else should be pristine from MS ( I use in-memory patcher for custom theme support / TCPIP Half-open) 2. Backported Apps Once RAM issues squashed, we can talk about bring up modern app (which have memory hog tendency) into XP If it is OpenSource, fork it, patch the source, rebuild (VC++ 2015 still support XP, GCC can be moded to use newer CRT) If it is closed source, avoid tampering the binary (think of digital sinature, false positive AV, or worse getting reported) Maybe in a form of _per apps_ shims dll which contain backported functions or else dynamically LoadLibrary() newer runtime and/or system runtime. 3. Backported DX10/11 (Games/3d apps) First, how many people want to put high-end machine with XP-32? Will these machine support XP? see issue #4 Wined3d (dx to gl), Angle (gles to dx), MesaGL (gl to llvm). All of these really good enhancement for XP but don't get hope too high 4. USB3 driver (esp. Intel) and overall driver support Is there USB3 backport? Unofficial? which means digital signature/WHQL issue. While desktop mobo used to offer 1.1 2.0 alongside, how about laptop? How about other hardware? I don't think screaming at vendor for unsupported OS is good ideas. Thats not how business work, after all they had to maintain "good" relationship with MS too. Consolidate the effort is one centralized site/forum, with one dedicated signer for public downloads Greetings!, a long-time MSFN lurker :p

Agreed, you should try it Sure your winnt.sif just fine (quotes make sense on value with space), I once have a trouble with mkisofs (totally different from cdimage). When I use ISO level 4 (wo any other extension joliet or r2) the Filesystem become case sensitive (big problem at DOS part but no problem at GUI part) and $Oem$ is not equal to $OEM$ so on. Shouldn't be a problem if you use the same image though.

Sorry Phaolo, i have just read post#9.. 100% my fault. So lets me say that you already know: BagMRU/1 which in my system (hex:14 00 1F 60 40 f0 5f 64 81 50 1b 10 9f 08 00 AA 00 2F 95 4E 00 00) point to bin's bag or more correctly system seek which BagMRU contained that hex, otherwise it will create new one orderly, read its nodeslot and get the shell. And we're agree that hex is persistent as i found that it also exist inside shell32.dll file And the only left problem is that weird sort? Then, nothing more i could say now(days ago actually) (oh, and i didn't set any stream) but I have follow your instructions exactly, more I even set date and time format to match yours, and add more deleted file in various time and date, but it sort perfectly. maybe it just a Bug for you Windows release

I customize it a bit more with nLite, but i have rushed to reinstall it (since it's in my real PC not VM) so I have deleted the master folder including settings file. does incorrect "ComputerType" in winnt.sif may cause it? and what about ntoskrnl? ntkrnlpa? I think this is serious wrong setting that get unnoticed by nLite and my me

Phaolo You really misunderstood me! After step 3 here is>> After I open and customize some (including "windows\Downloaded Program Files" and other) folder's order here is>> After I OPEN Recycle Bin here is>> note: i don't set the order yet, and voila BagMRU/1 pop in. After I restore my reg which is: (note: nodeslot 4d equal to 77 in decimal) Windows Registry Editor Version 5.00 [HKEY_CURRENT_USER\Software\Microsoft\Windows\ShellNoRoam\BagMRU\1] "NodeSlot"=dword:0000004d [HKEY_CURRENT_USER\Software\Microsoft\Windows\ShellNoRoam\Bags\77\Shell] "FolderType"="MyDocuments" "MinPos1280x1024(1).x"=dword:ffffffff "MinPos1280x1024(1).y"=dword:ffffffff "MaxPos1280x1024(1).x"=dword:ffffffff "MaxPos1280x1024(1).y"=dword:ffffffff "WinPos1280x1024(1).left"=dword:000000f2 "WinPos1280x1024(1).top"=dword:000000f2 "WinPos1280x1024(1).right"=dword:00000412 "WinPos1280x1024(1).bottom"=dword:0000034a "Rev"=dword:00000002 "WFlags"=dword:00000000 "ShowCmd"=dword:00000001 "FFlags"=dword:00000001 "HotKey"=dword:00000000 "Buttons"=dword:ffffffff "Links"=dword:00000000 "Address"=dword:ffffffff "Vid"="{137E7700-3573-11CF-AE69-08002B2E1262}" "Mode"=dword:00000004 "ScrollPos1280x1024(1).x"=dword:00000000 "ScrollPos1280x1024(1).y"=dword:00000000 "Sort"=dword:00000002 "SortDir"=dword:ffffffff "Col"=dword:00000002 "ColInfo"=hex:00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,fd,df,df,fd,0f,\ 00,06,00,28,00,10,00,34,00,50,00,00,00,00,00,01,00,00,00,02,00,00,00,03,00,\ 00,00,04,00,00,00,05,00,00,00,b4,00,b4,00,78,00,60,00,78,00,78,00,00,00,00,\ 00,01,00,00,00,02,00,00,00,03,00,00,00,04,00,00,00,05,00,00,00,ff,ff,ff,ff,\ 00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,00,\ 00,00,00,00,00,00,00,00,00,00,00 here is>> after logoff and open bin again >> still the same Now here is the quirks no.1: I should revise my advice to set deny permission and it should be mandatory! Yes, after unknown period (20 minutes in my test) BagMRU/1's nodeslot changed/deleted mysteriously. that's why you need to set deny permission (for users and ?system? ) for that key and MAYBE for bags/77 too like this>> quirks no.2: I'm fully realised that we are editing HKCU so it never universal to every user...that why some say its impossible/impratical Some say you could change permission from regini.exe (I haven't tried it) I hope you don't feel this is not achievable.... Bye

ummm, I think here at msfn someone has made a batch called "Silent .NET Maker synthesized" Are you sure havent see it? I tried myself. it similar to AIO but use different hacked 7z sfx that switchless.

It happen to me only once, but sadly dont know what cause it.. my PC is E2160 which is dualcore duh LOL but after install it only use single processor as shown in task manager, but devman tells it has 2 processor. I've change number of proc=2 or computer type=multiprocessor free but still revert back upon restart.. after rebuilt the image again from scratch and reinstall it works normal Its Unattended WinXP SP3 Pro with all postSP3 2008 updates any idea if there is a setting that can cause that?

In CDRTFE (x-cdroast equivalent), the order file is in pathlist.txt and seek time is consuming "I doubt this. Which Iso does force the file sort order?" i think what jaclaz mean is the LBA order not TOC order, in TOC file mapped in LBA (a sector or cluster?) which is physical cd position in bytes. "You may speed up if you put all needed files in an zip archive and extract this to your hard disk and run your installations from the hard disk" Is that mean change the "SourceDisksNames"?

"I found that not all programs use the global paths for them" Well, that the purpose of junction right? put it as windows default path for compatibility. The physically path (and registered) still C:\Programs but it have C:\Program Files Junction thanx leass, I'll sure try diskpart.. But do you have a clue about when the script will run? or where to put it? C:\Programs will be the registered path, D:\ linked to C:\Programs (when it executed?) then C:\Programs linked to C:\Program Files (executed at svcpack) at last D:\ set to not mounted with driveletter just mountpoint (not shown in explorer, just a tought) I always use C:\Program Files Juctioned to D:\Programs for 5 years, and never got BSOD, probably because i never use AV. And the one i found incompatible with junction was CDRTFE which its underlying mkisofs, and friends are in Cygwin emulation mode. and probably any cygwin ported