
user57
MemberContent Type
Profiles
Forums
Events
Everything posted by user57
-
SSE dont sound to bad for me to be honest, at some point a CPU would probaly just be to slow to handle the job - here it might could be solved but is it that SSL that was written in SSE ? if so it is certainly possible to write that one to normal opcodes then you probaly need to understand that XMM logic and transfer it to normal opcodes (emulators of elder retro games do so for example). since it would be a hash then nobody would reconize a differenz
-
"ImportError: DLL load failed while importing win32api: The specified module could not be found." that one sounds of interests because if its right what it says it has searched for a module (aka a dll or similiar) not a missing function (that may resides in that youtube downloader code) as i said i dont know python so i cant say exactly - but if its the other way around (for example it did not find the function) that description message is wrong but as i know python is a challenger to java - and java is rather a script/engine like language that means if a certain version build in some new functions or changed things up you bond to that engine - therefore it breaks up it got an advantage however since its a script "that should solve a piece of code" that code can get a different interpretation then that function can just be replaced with one function that works with the backwards compatibility i have a bit of a different meaning to vistalover you can write a code that works for all os (in this case just normal opcodes) then both works - and its "like" a end of story the way this gotten is a bit weird it sounds like 1 guy tryed to go against many people code (what he probaly has hard to win over) while if all just made normal opcodes that would not even have been a problem it makes it sound like its hard to make it backwards compatible - but rather its oposite in this case it might be right if you make the right "if´s" but if it gone the other way around (just making normal opcodes) neither future nor the backwardsguy "diskf" would have any problems and that way it would not be harder for anyone either in sence of a company backwards compatibility is a good thing - because it means more buyers, as it has a wide range of users that can buy that product
-
i do not know python however if i interpret it right it says it cant install that "packet" called "pypiwin32" version "219" ? if so not the compiler told you a problem, then the installer of that pypiwin32 failed whatever name it can be given "packet" "plugin" "addon" "extra piece to pyhton software" "pypiwin32 v219" https://pypi.org/project/pywin32/ only in case if i suspected right, if not im happy to see the comments and being correct to the right thing
-
a fingerprint ? its just a infrared light (flash) that red light is used as flash on your finger - then a sensor record that image that image is then compared with all "finger-pictures" it has in the databank like being a criminal :-) a few also look for the finger temperator with a second sensor (but why, then you hold that guys cutted finger under your arm - and that finger has the right temperature) the iris scanner is working similiar - the iris is even a little more unique - im not certain what light is used - but also it is a picture
-
i get it, its that python compiler that use these sse commands then that vc14+ are just different writings that the compiler has to translate, the compiler probaly didnt care using normal opcodes instead it just included sse commands the compiler then see what you wrote and translate them into opcodes, link them together ect. that c-runtime made us problems, in the past we could just edit the c++ files that generate that c-runtime that c-runtime is a .obj file (that one contains the functions the compiler links against/to) microsoft then hidden that .obj file somewhere (actually we dont know where maybe as resource in the vs2019 executable or that hidden .obj file somewhere in its temp files) - but no matter we know its a .obj file people wondered why they see functions in vs2019 they never used in their entire code - the c-runtime ".obj file" thats why it get executed before your executable is running -> after that it executes your code i did not find a good method that gives that vs2019 a good buildchain (so our functions are overruled) , and somehow you cant override these functions , if you try to trick that c-runtime it come up with an error i still wonder why a modern vs2019 compiler dont have a such thing, there seems to be no real override method vs2019 offers then it would be about to edit the python compiler and replace the instructions if python is open source - so you right to choose 2 executables - the other way around makes to much work
-
the same downloader has 2 versions for SSE2 and SSE ? why it didnt go the common way like check if mmx -> use mmx , if sse -> use see, if sse2 -> use sse2, ongoging you can definatly make it so, the compiler like vs2019 use sse if not turned off but if you have a routine that use sse you can skip it if you need, therefore the opcodes are then skipped if you dont have sse you can use normal opcodes or a different mmx-avx routine to have the right functional code one way to emulate a 64 bit command with normal opcodes is doing it multiple times , and instead of a register you can use the stack or some memory space you can control that piece as 64 bit i think dietmar did a such thing regarding replacing the 32 bit command CMPXCH8B command, it is a command that can change 64 bits in 32 bit operating systems using two 32 bit registers if its the same downloader it might solve this issue
-
XP/Vista-compatible clients for modern email services?
user57 replied to Mathwiz's topic in Windows XP
i think i have to understand the problem regarding oauth he wanted to get his emails with outlock - what is normal and common then microsoft aka hotmail.com added that oauth now it raise questions a : it need a second email for oauth / or login and pw for that oauth server b : it is just a mechanism that connect to that oauth server - where outlock dont have the oauth code - so it cant make this part and microsoft email just stops doing its job having that what they called a "token" to me it seems to be some hash then it can call with this hash, the server of interests and gather the data of interests that rather sounds like you have a login to some kind of server that says "this IP has sended me the right code - let this guy on your server" this sound all very old like a handshake with TCP SSL or a server that gives out something like "let this guy in" then your email such as a hotmail email allows you to see facebook, paypal, youtube - without having entered the password for facebook,youtube or paypal it raise questions where this oauth has its code , but the part that is making the question to the oauth server has to be on the users computer if its that it might be a module, a internal function, a hash maker in firefox, a certain code that is being executed that sounds insecure to me if someone has the right conditions can probaly just enter your facebook , youtube or wrose your paypal account this not only goes for a hacker, that also means people of interests can just make this with your account (such as the right people who have that trusted status - and that will not only be the police - and if so it raise questions why the police can just enter and look around in your facebook,youtube or paypal without having anything going on ...) so its a spy mechanism for the state - the more they know about the people the better they can enslave/control them (because guess what these people will have that mechnism´s - one might claim "on no i would not do that" - nope he will do he get a letter from a lawers and at some point he collopse, or something like "we dont do it yet/now" -> "oh see there in 2025 the terms of use changed now every people i want i can give this") so we cleared 1 question, why the do this - it dont give any security questions, rather it opens security questions and we know how this ends in a change of the so called "terms of use" - and then its done - you can be spyed -
XP/Vista-compatible clients for modern email services?
user57 replied to Mathwiz's topic in Windows XP
to me OAuth looks it just being a extra mechanism something like "shake-hands", that xp certainly could do if the mechanism is known where and how the picture OAuth only also shows a picture where your computer is asking their "OAuth-server" if everything is ok - this "OAuth-server" then communicated with the target (such as youtube, microsoft and others) when the OAuth with your computer was ok and the others (yt,ms ect.), then it grants access to the wanted resources like pictures and video that dont sound so special to me SSL or TCP-handshake is doing a similiar thing to me then it sounds like they just added a next one doing the same thing, with only one difference that a such mechanism is used 2 times -
XP/Vista-compatible clients for modern email services?
user57 replied to Mathwiz's topic in Windows XP
i have made a nice visual for the PSE (Page Size Extension) - the memory limit in the right part of the picture is from a older list of memory limits - that may or may not also include the other method (PAE Physical Address Extension) or both PSE and PAE combined the calculators show the related bits in 1010 binary format and in DEC for both 36 bit or 40 bit (since amd athlon maybe ? but somewhere around 1999 that started - now we have 2024) as you can guess PSE is 1 of the methods to reach more then 4 GB ram the other is PAE, a third way would be a second/3/4/ect. application (that then can address or point to different physical memory) interesting i find that intel lacked behind with 36 bits (64 GB) while amd already had 40 bit (1024 GB) also we can see that windows 2000 can have more then 4 GB ram, in that list it has 32 GB of ram (that also fits to the release dates of the cpu´s) -
XP/Vista-compatible clients for modern email services?
user57 replied to Mathwiz's topic in Windows XP
that might be interesting most cpu´s had 64 bit support while everybody used up 32 bit the speed maybe ? hmm no mainly because 2 reasons first normal opcodes are not that fast (you lose vs 32 bit even on x64 bit if you use MMX-AVX (we talked about fast cpu and compatible cpus recently)) then also you have more then 32 bits on that MMX or AVX registers (in 32 bit mode) then the next question kicks in high languages are rather made for making the things simple - but they are not fast c++ found a good compromise however it still lose to a assembly implementation so if somebody say i want to use 64 bits because that is faster then i must say no rather changing your programming language would speed up your code a lot (and also significant lower the file size) there is a big downside to assembly and maybe c++ but lets talk about assembly first, you need to know like a lot more of math and logical reason to do this also you need to write the entire code yourself (not like for (x=1,x<3x++)) you have to write this code this make it a lot more work ... - definatly a downside a other problem are engines, engines are useally simplier to program but also they not very fast (so that have to leave to if you ask for speed question) i dont want to talk to much off topic now but we had a such discussion recently (LAV engine) - but that is not true so directly it already are 2 engines (that LAV engine + the d3d9 engine) (and thats only the ones we certainly know of - maybe there are more) (now we have 2 engines it goes through before it even reach anything ... while we figured out that we dont even need that engine nor a engine to do so ) a other discussion is the memory limit and no the thing can be a little harder here i wrote about that already so i try to make it short this time in the past segments was a word (in like 8 bit and 16 bit (65kb) or 20 bit (1mb)) so the idea was to have a segment that points to 16 bit memory (65 kb) that * 4 bits (256 * 65 = ~ 1 mb) (i think some should have heared about that 1 mb thing somehow or seen somewhat) (here one for later to have one from wiki) https://en.wikipedia.org/wiki/X86_memory_segmentation a segment is like a arrow -> segment -> 0 = 0-65 kb segment -> 1 = 65-131 kb so if one arrow points to a house and that arrow can aim for a different house you understand segmenting to call out this part, 32 bit has segment registers - but it would be rather a long story more into detail (rather talking about 4 kb or 4 mb pages (long mode) ect.) the most applications dont need 4 gb either, for sumatra pdf this is the case - you can set the compiler to x64 (chrome for example start up a an next application (always called chrome.exe) and having always new address room) but except that the file is more and that it dont start on 32 bits nothing is diffrent for sumatra pdf, and guess what you can run that 32 bit sumatra pdf on x64 too to me that raise questions why i would even compile a x64 version -
XP/Vista-compatible clients for modern email services?
user57 replied to Mathwiz's topic in Windows XP
can i ask why is having x64 important ? -
to be honest i do not know all the user/system settings that are possible however the restriction that the other "users" cant do anything was a common want there used to be like a father giving his son his PC, making him a such "user" account - restricted , not followed to see everything , not allowed to install everything also on the most installers you can select "for all users(aka system)" or just "this user"
-
i think neither where that leads to, he should try to be the administrator and see if that happens there
-
on your installation you didnt set a administrator pw ? if empty you can just click ok there are programs and methods to remove that administrator password - maybe someone point it out some of those, its been a while for me since i do this - i never had problems with this as im always logged in as admistrator maybe: ? https://www.partitionwizard.com/clone-disk/windows-xp-password-reset.html
-
well it says check to free disc space that´s an easy one start -> my computer -> (rightclick local disc C:)) -> properties (free disc space) for the permissions you look the folders if something is wrong check all permissions C:\Documents and Settings\ C:\Documents and Settings\All Users\ C:\Documents and Settings\All Users\Application Data\ you look there if you have the right permissions "All Users" describe you not logged as administrator you can try "right click and start as administrator on one of these installers" or you can change the user to administrator and see if the installation works there that is somewhere start->Log Off where you can switch the user
-
well if that is always the problem adding the algo why you not just write them to normal assembly commands ? that is possible making that hash/algo with normal commands dont cause more cpu intensity because they are only a small thing in the code MMX and all its followers are rather for cpu intense parts (such as a decoder or encoder, or a sort mechanism) - but just to make a hash ? that raise questions its going back to a old story - where the makers once had 2 cpu´s 1 with old but very compatible commands (that only got improved in speed) 1 with specific made for speed but no compatibility - almost nothing worked here - and if so they had to be entire rewritten only to do so - big cost = both where taken - the CPU can do normal commands and MMX+ commands now they already beginning to say something like hum now we have AVX we dont need SSE4 or SSE2 anymore to my opinion the answer is very clear - we can have both
-
you wrote "Problem is that when i tried to install Visual C++ 2010 Redist" but thats a own installer its called something like "vcredist_x86.exe" but then in the screenshots you tryed to install some kind of the "dot.net framework 4 installers" later on "adobe reader xi" that raise the question why you said the VC 2010 redist is the problem - while you tryed to install different things while framework said the disc is full or inaccessable adobe said something like it cant connect the update server (if i got this one it right) there are installers need internet access is it a windows 32 bit XP or a win 64 bit XP ? 32 bit framework 4 to try (KB982670): https://www.catalog.update.microsoft.com/ScopedViewInline.aspx?updateid=0d076fbf-33cd-4b86-9288-ab31899b74da 64 bit version (KB982670 x64): https://www.catalog.update.microsoft.com/ScopedViewInline.aspx?updateid=434aa8fa-b425-47a9-be5b-dd000b66c6ed one more thing about the x64 bit version of XP - that time everybody used a 32 bit system thats why there are 32 bit version of the next following operating systems that time we could not find any real advantage to a x64 bit system therefore x64 XP missing many fixes that XP got up to 2019 - the people didnt care about it much - seems microsoft neither you can run a chrome 32 bit version on a 64 bit machine np, but if you set the compiler to "x64" you cant run it in 32 bits - while giving litterally no advantage chrome is running new executables (aka starting up chrome.exe severial times) - and those all got 2-4 gb room (that is a lot for a app) 2 GB if you want just userspace , 3 GB if you use the large page version (then the userspace is 3 GB for each startet app), 4 GB if you include the entire address room directly however those executables can point to different memory - therefore it can pass the 32 bit limit of 4 GB , by starting up multiple executables (only 1 way of severial ways, as paging can point to different memory too(paging was a thing in the past - it seems it got forgotten there was for example a 64 kb vs 1 mb paging version, just to make an example)) let us know if these framework installers work or not ty
-
the lap-top "Lenovo G580" you described has sse but or we talking about the specs you posted below ? "AMD Athlon 600 MHz" that one only got MMX just to confirm are you talking about this one ? : https://support.lenovo.com/us/en/solutions/pd025312-overview-lenovo-g580
-
that one (sp1)? https://www.microsoft.com/en-us/download/details.aspx?id=26999 screenshots before (like before you clicked) and when the error happens (like at 10 % , like after the ending ect.) a smartphone picture is already enough too
-
its something for the 2000 forum, there is an extra entry for win2k
-
maybe someone should search where those images are hidden, maybe they hidden in resources that my documents is just a folder something like "C:\Documents and Settings\Administrator\My Documents" the themes are useally in "C:\WINDOWS\Media" you can change these sounds to something you like the basic background should be in "C:\WINDOWS\Web\Wallpaper", here you can choose the folder + file is readed from if i want to test the soundcard in windows xp i useally go "C:\Documents and Settings\All Users\Documents\My Music\Sample Music" and play beethoven if you want a certain ICON you may use exting extractors such as "iconsext"
-
it would maybe have exactly this in parameter "dwIoControlCode" https://learn.microsoft.com/en-us/windows/win32/api/ioapiset/nf-ioapiset-deviceiocontrol https://learn.microsoft.com/en-us/windows/win32/devio/device-input-and-output-control-ioctl-l the hardware itself has a control bus for this, a data and a address bus that goes back why every cpu and hardware (such as a hdd drive) use a buffer/cache and thats also how the CPU can skip 32 bit commands and use for example a quadpumped stream, then rather is limited to the BUS/device speed it longly dispatched the just "tick 32" tick/cycle to also say this that the FPU in the past already was 64 bit wide (while running on a 32 bit cpu), then mmx and the others apeared extending this up to 128, 256 and 512 (avx512) - that not include the cache tricks nor the logical hardware units
-
im not certain but controlment to a device in windows is very simple you open a handle to that device, and after that you can stream your data either as input or output the text field "hardware device to use" would be a indicator to have it that way that raise questions again , why we cant give that "device" it´s right I/O control code that is also done with the famous deviceiocontrol function (read write and control) then it should be doable
-
its nice to have something like that but also that depents on a grafic card, it dont give normal xp users what might have a 4 core cpu or maybe a fast 2 core cpu a way to look this it raise the question why certain LAV engines can do this and certain can not , it raise questions how they activate that or set the right parameters - it would be good to know, also if you want to use a different LAV version where it dont work to the second answer DXVA in XP-32 dont work at all - that sounds like a dead beef - that engine cant be used for XP - sorry not a big fan of the x64 version of XP, it had a lot less support therefore it is often a bit buggy. also you can set the compiler to compile a x32 program and then both works 32 and 64, but if you set 64 bit then 32 bit dont work then there is the 4 gb question we have talked about that and yes XP/or 32 bit can use more thats because PTE PDE are entrys to physical pages , its more then just a 32 bit wide offset - a ramdisc proofed that in 32 bit are im really the only person that could write out a h265 decoder with normal opcodes up to XMM (what are 64 bit by the way, running on 32 bit - later even 128 and 256 bit wide - and yes for 32 bit)? https://stackoverflow.com/questions/44299401/difference-between-mmx-and-xmm-register
-
interesting point so the LAV is a engine to simplify the dx11 video engine ? there are such engines for GDI, they really made a lot of engines rather then using GDI directly, same goes for dx , its getting "upscriped" and slower