siria

Sounds like a similar project as the 'UOC-patch' But makes me wonder, why edit so complicated inside omni.ja? greprefs.js only contains default prefs too, they can be overriden outside in normal ways (about:config, or prefs.js, or placing a modified js-file in the defaults/pref folder to get own default values) But perhaps works differently in CometBird, no idea. Something else I wondered for a long time: am I the only one who has (now had) major trouble seeing those embedded crosslink iframes to other msfn-postings, with older Mozilla browsers?? Finally got them visible now with a little css-tweak.

While researching the kinda weird useragent result produced by my own, never-used ancient IE-browser, came across this really interesting site. Highly recommand for everyone interested in the useragent structure in general, what those "Windows NT" numbers mean exactly, or those ominious "NET CLR" elements, and especially how MSIE-strings are getting customized, etc.: https://docs.microsoft.com/en-us/previous-versions/windows/internet-explorer/ie-developer/compatibility/ms537503(v=vs.85)

Yep that's right. All browser agents start with "Mozilla" because in the early days of the internet that was the main browser, and just like today, many web authors seem to code only for the currently biggest browser. So the other browsers started to claim they were that browser too. Same prob as years later with MS-IE. And today with Chrome. Most authors code only for the main browser and the users of smaller ones are forced to fake their "user agent string" and pretend to be that other browser, if they want to get the same code. Which in most cases works fine for them too - if they get it. Since the "Mozilla" part has become so useless, the browser names appear now in the middle or even end of the strings. So it's no wonder those test website know (?) that you're using IE6, simply because the useragent string you send them claims exactly this: it contains "MSIE 6.0" in the middle. They'd just as well believe you're using Firefox if you'd send them a fake firefox string. Here's a site with large collections of various UA strings, find such lists quite interesting: https://developers.whatismybrowser.com/useragents/explore/software_type_specific/web-browser/ Or categorized: https://developers.whatismybrowser.com/useragents/explore/ Can still remember how slightly shocked I was when finally realizing the UA is not just about the browser, but additionally sends info about the system, and sometimes the language too! When using such rare systems as Win98 today this info makes you much easier to track and recognize again.

@SH: can you open the directory? http://o.rths.cf/palemoon/

@Sampei.Nihira Your screenshot is for httpS rths also works with plain http (as in VistaLover's link) Is http broken too?

Just because the UA string is different does NOT mean that Maxthon were NOT using IE6. It's just a little piece of text - and faked by KernelEx.

Yep, that explains why you can read the KM-forum with it, but not with your IE6, which sends the default useragent with Win9x.. And running a browser with a KernelEx compat setting explains why the useragent it sends is faked, even if you reset it. The browsers do 'believe' the compat-system were the true system ;-) It's really that simple in the case of KM-forum. The SF server is only interested in the system part of the useragent, no deeper checks of any sort. Actually in my experience it happens hardly ever that a server does not believe in fake UAs. Of course, they just can't dig deeper while javascript is blocked. If JS is allowed, and perhaps even 3rd-party scripts and XHR, the servers can check themselves all sorts of details from your browser and settings.

No need to 'believe', it's easy to check ;-) What platform is this page reporting, when you open it with Maxthon? http://mybrowserinfo.com/detail.asp Or the Windows version or "operating system" here: http://useragentstring.com/

That means IE6 sends the real system (Win9x) in the useragent, and Maxthon a fake one (newer one). Either intentionally faked, or by a KernelEx compat setting - right...?

Uhm... just wondering (am not using it myself), but what does that mean? Does the installer delete the whole old program folder?? If yes, it's important to know that unlike other browsers, in KM some user customizations are stored outside the profile... Mainly toolbars.cfg (gets edited), usually all macros and extensions go in the app folders too (except if a user knows exactly what he's doing, does NOT follow the usual instructions and creates himself the matching folders manually in the profile)

Just as info, since K-Meleon is often recommended especially for users with old systems. Also KM1.5.4 for being the last version to run on *very* old systems natively, without needing KernelEx or other tricks, like Windows98. Yet weirdly, the K-Meleon FORUM is blocked for visitors with "Win98" in their UserAgent string! Now tested the above string too, because it contains "Win 9x": Mozilla/5.0 (Windows; U; Win 9x 4.90; en-US; rv:1.8.1.20) Gecko/20081217 Firefox/2.0.0.20 And yeah, the forum seems blocked too. On http://kmeleonbrowser.org/forum/ Error 403 We're sorry, but we could not fulfill your request for /forum/index.php on this server. You do not have permission to access this server. Your technical support key is: 02f7-f43f-17f4-e8c8 You can use this key to fix this problem yourself. FIX: use any younger system in the (fake) UA. If the browser is started with a KernelEx compat setting, like Win2000 or "Windows NT" , all works fine too, since the browser thinks the KernelEx compat system were the real system and sends it as Default UA. We suspect Sourceforge for that nonsense, where the KM homepage and forum are hosted :-( They do strange things sometimes. Admittedly the Phorum software is very ancient too. And to add insult to injury, users with this prob will get from the server no hint about the reason! The help link in the error message leads to an external page (www.ioerror.us/bb2-support-key?key=...) that says you're intentionally blocked due to supposedly "security" reasons with your system, BUT - on old browsers that page refuses to open too! Looks like it needs modern TLS. Result: even this info message is blocked too on old browsers! What an evil trick. Just posting this here as hint that it's a bad idea to use Win9x in a fake UserAgent string if anyone may want to visit the K-Meleon forum. Possibly some other sites too, no idea. Include any other system instead, for example XP as in this fake IE7 string: Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1)

YES - Thank you!! FINALLY, that link seems to work, and most important, it has also the archived Opera addons for download! :-) And finally some example addons, which may be used for future own tweaks, if needed. Major pity I didn't know about Opera12.02 some 5-8 years earlier, as fallback browser when a website wouldn't load. And meanwhile there are other alternatives available, with yet better TLS, thanks to roytam1. The mean thing about waybackarchive is that they may collect all that precious old stuff, which is all great, but then will NOT ALLOW users of old machines to FIND it. Those users who really need that old stuff. No chance. And google doesn't help either to find direct wayback links. Or if anyone knows a trick to FIND their archived sites with OLD browsers, please let me know how. At any rate those usual search links with the /*/ inside have stopped working years ago.

If an addon "installs" properly or not, that doesn't mean much, whether that addon really "works" or not. It merely means that the min-max version numbers which the addon author wrote manually inside the install.rdf file, are matching your current browser version. It does *not* mean the browser were automatically executing any feature compatibility tests before allowing installation. For trying Firefox addons in the earlier KM7x versions we had to manually edit and fake those numbers all the time, to allow installation, and then just look and see if it works or not, or partly. Since you realized already that an addon is causing your prob, disable half of them, restart and check again. If your bookmarks and history are blocked, that could point to a privacy related addon... It can also happen that two addons are just conflicting with each other. And, although forbidden by Mozilla, it happened in the past that an addon author (Maono, of NoScript) intentionally manipulated another addon somehow (ABP, from W.Palant), so that his own stuff wouldn't be blocked. It's probably rather easy for addons to manipulate other addons intentionally.

No idea if it's reset again, but on page 147 you reported that frame rate 60 fixed it on your machine: layout.frame_rate = 60 (old default / new=30 / auto= -1) Perhaps it's not 60 anymore? The other pref mentioned around that issue was: layout.interruptible-reflow.enabled = true/false Such stuff can differ on different machines and different browser versions. It would be useful to write comments in the UOC file, with little explanations, like the ghacks author does. In my private default pref files I even add comments that are visible on about:config itself :-) pref("layout.frame_rate__INFO", "INFO: -1=auto, default old 60, new 30 but choppy SCROLLING for FF45esr XP msfn 20190516");

Oh. TLS? THAT is the most effective killer of old browsers in today's web. Works 100% deadly. It has to do with encrypted transmission of httpS web pages. If a server admin has decided to send readers only highly 'secure', extra strong encrypted pages, using modern TLS1.2 encryption, people on older browsers will not even get to see a single line of the page. Let me say it that way: Did you ever see such a similar message when opening a page?? Secure Connection Failed An error occurred during a connection to github.com. Cannot communicate securely with peer: no common encryption algorithm(s). (Error code: ssl_error_no_cypher_overlap) This happens when the browser doesn't understand the encrypted garbled stuff it receives. It cannot de-crypt it. Older browsers understand only older, weaker encryption protocols, like SSL3. Server admins could just as well decide to send that extra strong encryption only to newer browsers, and lesser secure, old SSL3 encrypted pages to older browsers. At least for harmless public webpages, not banking. Would be easy for them, and some do it. But more and more don't care anymore. It was the Almighty google boss who decided he wants the whole web encrypted, and used his monopoly on search engines to force website authors to do what he wants, if they still wanted to get listed in search results. And the next version, TLS1.3 is already coming up, sigh. That Opera toolbar actually did not take much work, the prob was only figuring out HOW to do it. Luckily someone gave me a few tips, in another forum. And of course have it long since forgotten again... But could look it up again if you're interested? It was some mix of drag-dropping elements and of editing text files... But I think most was just stored in some user config file, could find and upload it if you want. Oh I just remembered something, since you mentioned Opera11: Customization is partly broken in Opera12 in old systems, only possible to customize Opera11. So I just customized Opera11 and then copied over its config files into Opera12 - and it worked! :-)

I think you nailed it now. That's the point I kept overlooking, due to not using those builds myself: in PURE Firefox site-specific useragents were re-enabled again after awhile - but NOT in regular releases, only in nightlies, betas etc.! Just guess they disabled it in FF21 already, that's also why K-Meleon7x, which started with gecko24, had it unintentionally disabled too all those years. With no one even aware this very handy function exists and could work - until finally roytam1 came and initialized it. And after KG76 also in KG74.

@Zort While I used Opera12.02, this link always worked for me from a bookmark, after changing and clicking SAVE: opera:config#UserPrefs|CustomUser-Agent Do you still have the link to the UA-switcher? That an IE string works well now is no surprise: the UA that I'm using the most in KM1.6 is IE7! Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1) Have the suspicion, from experience, that some old websites (incl. Google search) still have inherited some special rules for it, since at its era it was non-standard. And they never touched those anymore, or forgot to delete... But today those few css+js differences are a far lesser prob for old Geckos as getting just slammed the door in the face with an "Old browser, get lost" sign. But sadly more and more of those old rules vanish anyway. One of the worst desasters, just recently, the old classic no-script version of Google-translations for "selected text" by macro stopped working :-( Now the Kmacro only still helps to translate full page URLs - and even this works ONLY by macro now, sending the page-URL included in the google-URL already. Or perhaps would work if enabling JS, but that's causing other giant probs. For such cases Opera12.02 has already much better chances too. Or the ultimate solution, if nothing else works and important enough: firing up KMG74 and allowing scripts (just nearly freezes the machine, but otherwise works) KernelEx: I fully understand your prob with hesitating to update it. Such projects are a major undertaking on real ancient machines. I've also not managed yet to try a newer version since the last adventure, although would love to. Never touch a running system... But am watching and waiting for a new version which sounds a bit easier again as current modular updates. But if you have the old basic KernelEx version that's already the most important update by far! From the browsers I suggested above, probably roytam's Firefox 3.6 (fx36) is the best for this case, if TLS1.2 is required. Along with old Opera which can handle newer page elements. If you're occasionally using K-Meleon again, on windows98, I have a couple of little tricks up the sleeve... For example: absolutely install the stylekiller macro! And priv3buttons for quick toggling various JS-levels. And perhaps Blockeria... And a few other macros, and some special css-rules.... (have packed all my custom rules into adblock.css, for toggling during session) But most own macros I use for fixing sites are only drafts. PS: not sure if uploading works here, or need JS, but if there's an attachment it shows my customized Opera toolbars, after getting a little help from another forum member. No, of course uploading did not work (KM1.6, WITH JS it got even worse) So here's a temporary link, just a screenshot: http://s000.tinyupload.com/?file_id=69601804641062789660

. Good idea. What still keeps me riddling is that ancient defaults-pref path used in the non-KM browsers, why only KM uses Mozillas newer defaults path. And suspect this mystery (to me) is probably related to the other one, that you can't overwrite the useragents hardcoded in omni.ja by a simple defaults js-file. Also have once read that both paths can work simultaneously, depending from some settings (perhaps in mozilla.cfg, can't remember). Anyway, for the time being would suggest to not make that path product dependant, just have the script look if the file exists inside the newer path, and if not found, look it up inside the ancient one?

Ah! Thanks, great not needing to open the JS+XHR doors anymore to see the file list :-)

@winxpi If you don't have KernelEx, the choices get slimmer. In that case I'd probably try his special Firefox3.6 (Fx36) Awhile ago I posted some direct links to roytam1 browsers: https://msfn.org/board/topic/178351-what-software-developed-by-roytam1-do-you-use/

YIKES... I certainly understand how you feel: KM1.6 is still my main browser too, because I need the macros, but half the web would be blocked completely too since 1-2 years without roytam1's updated TLS1.2 browsers! And for several years before getting the first retrozilla, the situation was already much worse as necessary, just because I had no clue about Opera12 yet :( And when finally learning about it, it was almost too late, there were already ALL Opera addons deleted on the web :( But since a year I use almost exclusively - old K-Meleon1.6 (only TLS1.0, but need tons of macros, most own ones), - roytams special FF2 - FF3.6 builds with TLS1.2, running even without KernelEx (no macros, but FF addons), - and if all fails, can fire up the most modern: roytam1's K-Meleon-Goanna74 (build for Win2000, buggy on 98se) KMG74 has by far the BEST TLS and CSS of all alternative browsers, and most KM macros still. The engine is from PaleMoon26, era of gecko24. But on 98se it has 2 KILLER PROBS: 1) It needs an updated version of KERNELEX, which affects the whole system. At the time I updated it was KernelEx...16, which was very buggy. Some younger apps started working better with it, but it partly messed the system for other old apps, e.g. a prob with new browser profiles! But have read most of those probs are FIXED again with yet newer versions. 2) KG74 CRASHES very often, especially when right-clicking anywhere. This got a lot better after finally figuring out it happens mostly when one of my countless macros updates a menu. And just yesterday noticed KG74 gets especially crashy after the first local file was opened. Must investigate that further. (3: and it comes with some hardcoded FF addons which can't be removed, one of them prevents faking the global UA but found a trick to disable this one) Are you aware there exist such alternatives with updated TLS1.2? Have read that roytam's retrozilla versions of Firefox2 (rzbrowser) and equally old Seamonkey (rz2.2, rz-suite), and his TLS1.2 version of Firefox3.6 (fx36) even run on PURE 9x or even NT. There is also another version of Firefox2 with TLS1.2, by the original retrozilla author, with the same version number as roytam's parallel builds (rz2.2). I have no clue what ME can do, but if you can run KM1.6 already, you must have installed at least the goold old basic KernelEx already? Sigh, wish there were also a TLS1.2 version of KM1.6/1.7 with working macros (the engine already runs fine in Fx36)

Sorry for the super dummy question, just theoretical, not in-depth: What would happen when trying to restore an old 98se backup image from an old notebook (Pentium3, build 2000-2002) on a younger machine, not using any tricks, just from a USB stick somehow with an image tool on it? Provided the first partition is FAT32 and sufficiently small...? Machine dead if no revival experts around? How are the changes to get the system running "just so"...?

(test) posted from 98se but only got through after deleting quote at beginning:

Please elaborate, to figure out solutions? The prob is, K-Meleon may be missing some essential things, but it also has lots of hidden features which simply aren't visible to unexperienced users. For example most buttons have a handy right-click menu, which most people never discover just because they have no little-arrow as indication (example Home or Go-buttons) A killer feature is the hidden privbar (View > Toolbars) with buttons for 1-click toggle of javascript, cookies etc. Or the "about:about" page has lots of working links to more settings, incl. about:addons, which are not found anywhere in the menus, just because the GUI hasn't been updated in the last ten years or so, only the engine (install macro aboutabout to get at least a makeshift-menu) Copy+search: select a text in a webpage and hit the search button (or similarly: select a text LINK and hit the Go-button) Duplicate a tab: pull the tab into an empty space on tab bar (if any left ;-), or right-click on Go-button If you want a closing cross on tabs: guess for this there's actually a GUI somewhere, perhaps in F2... (would have to look it up, am myself stuck on old version) There are also macros for easy useragent-toggling, I recommend my "useragents2018" which also helps for easier managing site-UAs.

Yeah. Very bad idea. The STRONG drawback is that users cannot change those values anymore if they forget it. Awhile later they may toggle a setting in the menus or prefsheets or wherever, and think they've changed it. Or they toggle something on about:config and think they've changed it. That works as long as your preflist is only a default list. But in user.js it will override all individual user settings again at next start, with the users even unaware. If they do eventually notice their own settings were lost, they will try again, only to find their values "don't stick". Only a few experts will then figure out that mystery ;-) My hint was more meant for users who like to fiddle themselves, if they have trouble to ENFORCE something, so they may "append" a few lines. And also as general tip that this widely unknown, very handy startup-prefs file exists. But to enforce your complete profile settings at every startup again is a very bad idea... Looking at those 'stubborn' prefs again, they are probably hardware or firmware related, depending on what's available on individual computers?