Ninho

@Mathwiz, All : as the originator of this thread, just wanted to "plus-one" that the "ProxHTTPSProxyMII" (.exe compiled version) + Proxomitron is working here and has been easy to set-up and configure thanks to your added remarks and caveats, solving the original problem. Congratulations added for your rather remarkable searching... and finding skills !

Great find, @Mathwiz ! It's actually many many yrs since I've not used the Proxomitron but it's still present and runnable on this system. However after a glance at the proxhttps page it seems like it needs many elements in addition to the proxo (openssl, python) which altogether might hide a few roadblocks - or not. I'll report back when I've had time to actually install and try that thing on my system... Postscript : I'm seeing on their forum that I should meet (major?) complications when installing and configuring proxhttps owing to the fact that my Windows XP (nor Seven) "boot" disks are not "C:". Oh well... Life wouldn't be life without complications, would it ?

I don't like Mozilla FF, at all. I'm still keeping an FF 10 ESR for cases like this; the FF 3.5 I mentionned in O.P. is normally reserved for basic Tor browsing here. Re Opera 12\\ no worky in Win 98 even w/ KernelEx say you ? Surprises me, I think that's the versions I have (on another PC). (Edit) Strike that, that's Opera 10.63. Not too useful at many modern web sites... (/Edit) While at it, I tried navigating to https://aidanwoods.com/blog ... not unexpectedly, Opera 10 / Win 98 have been "unable to locate remote server".

@Mathwiz: thank you, very nice and clear explanation of the varieties and nuances of key-exchange and en/decryption algos. @Jacklaz, MathW, Heinoganda : - I am not running real-time AV scanners, much less TLS proxies (mitm things!) - Chrome's Error code : Inaccessible Web page ERR_SSL_VERSION_OR_CIPHER_MISMATCH. The Web page at address https://www.aidanwoods.com/blog/faulty-login-pages/ may be momentarily inaccessible or it might have been moved permanently.

@Jacklaz : I don't know where you got this idea that I ignored your suggestions nor thought I "knew more than you", God forbid ! But I was not going to (re)install the QTweb or Opera 12 just for a check at most peripheral to the question... Happy new year !

Not at all. The question is whether it's possible to fix XP's native crypto libraries in order to accept the ECC (or whatever fancy) certs, not whether non-native crypto such as in Qtweb, Opera-Presto support them. The OP even asserted Firefox on XP does support the otherwise offending certs - which you decided to ignore :=) In addition, what Heinoganda said and you quoted here is in doubt:«The actual problem is the Windows XP no ECC certificates can be administrated or not registered and these are recognized as invalid.» This was certainly true of XP until SP2, but... isn't native support for ECC crypto one of the touted addidtions in XP-SP3 ? I guess it's more subtle than support versus no-support, the issue may be one of incomplete/buggy support for some flavor of ECC... Off-topic rant : why are ECC-based certs seemingly gaining wider use ? Of course they sport smaller key-lengths, but ISTM their security against progress in cryptanalysis is even less assured than RSA, whose mathematics are much better understood.

I guess it depends on the version of the Google Chrome browser one is running. While researching this issue a little more in breadth (if not in depth) I read that Chrome has had its own crypto for TLS and so on, starting from Chrome v. 37. But in my case I have Chrome 34 and stuck with it (for it is the last Chrome version able to run on a processor without SSE2 like my AMD Athlon XP) - this explains why our respective findings may differ in this respect : my Chrome still relies on Windows for the TLS crypto stuff.

On my XP SP3, running with the dll's from ReactOS : similar to Heinoganda's report, Chrome works fine, including with https sites, but the problem with sites s/a the example given upthread using newer incompatible encryption and/or key exchange only, is not solved by the replacement yet :=( I am suspecting the schanell.dll is not the one (or not the only one) involved.

Interesting find, thanks. Is there even a slight chance for such substitutions to work ? I know I've tried subsituting dll's from win 7 - not that I really expected that to work - and indeed, XP just died (blue-screened) during the boot sequence. OK, if you could post the dll's somewhere, I'll try substituting them (I'd rather not download the whole ReactOS - beta, which I don't use).

I have XP SP3 updated under the guise of "faux POS" (not all updates applied though) Lately I find some cases - more and more often - of HTTPS sites that : - under XP, won't open in Chrome (nor IE), i.e. using Windows own crypto API - but that - - will open under Firefox - which has its own crypto stuff - even a very old version (3.5) works ! - under Seven, for compare, same sites do open even in Chrome (or, God forbid! IE). An example page there exhibiting the phenomenon : https://www.aidanwoods.com/blog/faulty-login-pages/ Is it expected with fully updated XP SP3 and/or "POS" windows ? Or have I missed a related update ? For reference I have : crypt32.dll version : 5.131.2600.6459 (xpsp_sp3_qfe.131005-0434) and rsaenh.dll : 5.1.2600.6924 (not that I'm half sure those particular dll's are the keys of the problem)...

Yet the Microsoft Disk Defragmenter, "defrag.exe" version 4.70,.belonging to Windows ME but which I use in 98SE, has a Symantec copyright that further references Intel application accelerator, iirc.

Why indeed... is more or less what I've been saying, for full disk defrag choose one defragger (choice not critical) and keep using one and the same (per disk partition). I like to use the Windows built-in ones. Very wrong ! Microsoft has provided different defraggers since Windows 95, rarely house-made, more often crippled versions, licensed from various 3rd parties (Symantec, Intel...). I don't consider all of them "suck", not that they excel, but they get the job done, and the perceived price is right :=)

Very nice ! Thank you, Bphlpt and regards, too !

Another incredible webkit-based browser, which has its own thread in this very forum : http://www.msfn.org/board/topic/170291-web-browsers/ I'm trying it now - didn't know it nefore - I must say I'm impressed ! Extremely slim, portable installationn 6 MB ! that's six megabytes, not gigs :-) Fast, and also consumes much less RAM than its better known rivals, and with all that it does what you expect from a modern browszer, and more... WOW ! I'm bluffed ! Works fine on the Athlon-XP, of course since that is the thread's motto !

There's no defragger which would be best in an absolute sense, esp. in multi-boot environments, as someone above correctly pointed out. Now to the poll's simple question "which defragger do you use (in practice)?", my simple answers : - for casual whole-disk defragging, Windows' built-in - more precisely, the defragger built-in to the "windoze" which that particular disk is most tied to. I do not defrag linux "ext" or "reiser" partitions. Not that Windows (whichever) defragger does the best possible job, but it's counter productive to run different defraggers on successive occasions, since each defragger has different "ideas" and algorithms and so, it will take a long time destroying the work of the previous defragger... The choice is not critical, but choosing and keeping to one and the same (at least, per disk partition) is more important that using the "best" for whatever measure of "best"... Oh, and free (no cost) ones are good enough, keep the money for other goodies. - for special tasks, viz when needed to defragment a single file or afew selected files : Sysinternals' contig.exe , defraggler... Those were my 2 cents, take or leave... :=)

Super-duper idea ! Now posting using Iron-portable ! Warning : Version numbers on the download site are mismatched ! I wanted, at most, Chrome.exe version 34, but Iron's "version 34" in fact has Chrome 35 (!) and, not unexpectedly, will NOT run here. For anybody having the same requirements as I had (no SSE2), their most recent compatible version will have to be Portable Iron 33, housing Chrome.exe v 33.0.1800.0. Now to try and get compatible extensions (I only really need Adblock Plus and Noscript/Scriptno - can't live without the latter !) [Edited : ... ] Google makes its best efforts to prevent us users of non-Google Chromium derivatives from downloading and installing extensions from their Store. Workaround : find and install extensions "unpacked" in "developer mode". Which lead me to this truly remarkable extension I'm now playing with in my Iron, "HTTPSwitchboard" [ https://github.com/gorhill/httpswitchboard] Please share comments of that extension, ladies & gents, and please keep adding to the subject issue of the thread - software which breaks on pre-SSE2 processors, and workarounds and replacements thereof.

Many of us are still using AMD Athlon XPs (or even older processors - that do not execute "SSE2" instructions). The problem is, more and more popular software and installers now routinely build expecting SSE2 to be present,and even oftentimes not warning nor checking properly the actual processor features present - therefore, crashing ungracefully while attempting to run/install on non-SSE2 capable systems One goal of this topic - which might be made a sticky thread, if eventually catches enough visits / replies - is to discuss the situation and possible workarounds : list last versions of software, esp. browsers, that can be installed and operated without requiring SSE2 (even if some functions will then be crippled or disabled). This will help people avoid the downloading of software versions either dysfunctional or even crashing their systems despite claimed compatibility to their OS versions (compatibility claims seldom take this particular aspect in account) ! The idea has come to me from losing my installed Chrome 34 (hardware crash, no backup of the particular software!). I was - wrongly - confident I knew how to download older versions of the Chrome browsers from Google, alas! they changed their minds, and now, apparently, it is no more possible ! I /love/ Chrome - though hating how they force "upgrades", especially as they fall into the aforementioned category of developers who couldn't even be bothered to properly /check/ for the presence of SSE2 and let their apps/installers just /crash/ under you. Oh , well ! Does anybody have a working link to Chrome 34 for Win32 ( many results by popular search engines are NO GOOD : they will invariably download the NEW Google installer which in turns will download the newest, crashing, version from Google's servers). What other precompiled browsers, including Chromium (last version to NOT require SSE2 ?) ? Other, possibly non webkit based, nice modern browsers ? Meanwhile I'm posting using Firefox 10 ESR, which I really hate - almost as much as I hate MSIE !

It does, or I haven't triggered the conditions for it to fail yet ;=)If it doesn't bother you to recall, what is / was the symptom which made people switch back to the older dll version ? {edit:} Oh, never mind then, I'll have a fresh look thru the thread again. I have the previous dll version saved just in case smth would break again, anyway.

I think you misunderstood the purpose slightly of the little experiment I've been suggestingfor Mcinwwl to try. It is not meant as a concrete or even practical workaround against the bug, but to show possible differences in the Explorer shell behaviours, between Seven and older versions. In practice :=) I'm just...deleting the obtruding Desktop.ini_s (both!); in addition, set a value : UseDesktopIniCache (Dword) = 0 under the following Registry key : "HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\policies\Explorer" ... as per one of Jaclaz-provided links above.

Yes but I think you can move the icon outside of the visible area and it'll stay there, in effect, hidden. Try this experiment ; first check that your settings are not to "align on grid" nor "reorganize automatically". Increase screen resolution, temporarily; now frag the offending element to the right border (of larger resolution screen), finally restore screen to its usual resolution : does thumbs.db not stay off screen ?

Very informative link, thanks ! How do you find all those ? Jaklaz, you must be aware the Japanese have a concept of "living treasures", people truly irreplaceable for their accumulated knowledge and/or experience. Well I vote you for MSFN's "living treasure" ;=)

Ninho wrote: " I didn't write "system.ini" anywhere did I ?" ... Bphlpt : Just in the title of the thread. Done title editing! Duh... I really need some rest. At least in previous versions of Windows you could put them beyond the limits of the viewport. Is MS trying to be /useful/ by bringing them back in full view against my will ! Really ? Cheers and Regards as well///

I didn't write "system.ini" anywhere did I ? In any case yes it is about the files containing Desktop customisation settings : desktop.ini. Yes I am aware their role is little more than cosmetic, but deleting them is but a poor work around in last recourse for a standing bug (imho).

Let me start with great news : everything's fixed; now ! The last problem which remained could be classified "PEBKAC", I suppose : I had recently updated SRP (software restrictions policies) so Internet Explorer was forced to drop rights even under my administrative account - this was preventing the MS Update site from loading and causing the perpetual redirection, although giving no hint of the true cause. Seems I'm getting too old to remember what I did just a few weeks ago... In my defence I'm also playing with setting up Seven in a VM at the moment, it does get confusing at times. The automatic fixit has restored the latest version (.257). Since it is now working, I'm leaving as is ...

Ran the MS Fix-it! tool. It reported all green, having fixed everything it found. However : this has changed /nothing/ :=) Internet explorer 8 can't go to MU/WU web sites, redirects to the same error page 2497281 as always... whether I try to go to the site directly from the browser, or using the link in the updates control panel applet. Giving up ! {Edit} It's OK, fixed now, see second post under this one.