siria

> BTW I've tried KM74 and it crashed on me trying to load Youtube page No crashes here on youtube. The usual bugs probably: did you copy over the 3 files needed with modern KernelEx for httpS domains from an older profile? Or the other bug: first loaded a local page, then a web page?

Bruninho said: > I'm amazingly p***ed at how hard it is to get codecs right on Windows 98. > Frankly, codec detection is buggy on Windows 98. Even HTML5test.com says so. For me it's not very urgent to play videos directly in the browser, as long as I can download youtube videos and play them with a media player. Which seems much lighter anyway. The last years felt not much need to get any other videos at all, but probably possible to download more with 3rd-party services, perhaps the one I posted above (didn't test that in a long while) Oops, almost forgot: of course I'm downloading other videos too! Just rarely, and not as easy as youtube, and often un-successful. Oh well. Haven't tried services for this task yet, perhaps will when getting urgent enough. Usually so far only try to find a full mp4-link in the page source, if lucky with the help of macros, and best if the url is simply in the meta "og:url" But if you CAN run Win2000 and are happy with it, I wouldn't waste so much time struggling with win98 either :-)

Bruninho said: > I don't see anything higher than Windows 2000 apps working for the foreseeable future on Windows 98 Since many years a lot of XP apps work (more or less) on Win98 with basic KernelEx. Especially those who are intentionally blocked by their authors for no other reason as having no opportunity to check compatibility themselves. And what's really annoying, again and again based merely on the wrong claim that Win98 couldn't handly ANY unicode! As if Win98 would have been stuck at very first edition forever, even without any MS updates. While funnily no one insist that apps were "not compatible with XP" just because they need SP2 or SP3. Of course some XP-apps really work only partly on Win98, but that's already a lot better as "not at all". Bruninho said: > K-Meleon 74 Goanna could also be an option, if it was a very stable browser without UI issues. > Actually, I could install K-Meleon and try more, if only the UI wasn't THAT ugly. > I can't understand why K-Meleon 1.6 (or 1.5.4?) has no UI problems but 74 does? KMG74 sure has huge GUI Bugs on Win98, causing lots of crashing, not suitable as main browser. But can you elaborate what parts you mean with "ugly"? Just wondering because this sounds like "skin", but not sure...? If you do mean skins: no prob. The default skin is the same in KM1.6 and 74. The old Phoenity in KM1.5 was replaced with Aura in KM1.6 already, and this new one was technically slightly downgraded, the variation system messed, because it wasn't done by the main dev but a community supporter who didn't care for it. And additionally all button transparency was now "partial", which doesn't work on Win98 anymore. Producing such nice effects as the PrivacyToolbar, one of KM's greatest features since forever, having black text on black background on Win98... Grmpf. Compeletely unnecessary. But that's no prob - just copy over the old Phoenity folders from KM1.5 into newer versions. The old skins are still fully compatible even in the newest KM-versions. Then restart and choose it in F2>GUI. Or install a different skin, or extra toolbars, or modify it yourself.

YOUTUBE Win98: Bruninho said: > Youtube support. Of course I doubt there will be any workaround to play them - > unless I manage to set Opera to open the videos in VLC No idea about Opera, but in KM (or other browsers which can inject scripts) I'm just taking the easy way out: first download the whole video, then play it from local disk with a mediaplayer (MPClassic). What I usually like better anyway in most cases, except if first wanting to take a peak inside without needing to downloading the whole giant file first. So far this live-streaming is a prob on my system. In general certainly possible too, tried it once with VLC and a mobile rtsp-link for a teeny 144px file, but all other attemps so far only ended with crashing my whole system! Am sure other people can do it, but for the time being have postponed all further own experiments, and just keep downloading the whole files. And another catch of course: downloading only works if the individual video and youtube allow it. Luckily so far almost all allow it, at least of those I'm interested in. For downloading there are several options, either 3rd party websites creating the download links, example: https://bitdownloader.com/download?video=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DTbiedguhyvM (scroll down to get all formats) or using userscripts or addons or bookmarklets to create download links on the youtube page itself In newer browsers everyone seems crazy about ViewTube script, no idea if it works here too. Am just using smaller scripts in KM1.6, like this one here: http://kmeleonbrowser.org/forum/read.php?9,142050 (attached in 2nd post, the zip contains a KM-macro but also a normal javascript file) (Win98 visitors: if forum refuses access, just remove "Win98" from your useragent string)

I had the same prob on a notebook a few years back, sound was far too low (although "only" low, not "weird" at all). Double-checked "all" settings, all looked fine. Triple-checked again... still no clue. What can this be??? Gave up... Some days later hit accidentally a wrong F-key, and a volume bar flickered across the screen... OOPS!! There are special hardware buttons for volume on the keyboard too! Had known this before of course too, but was so completely focussed on software settings, somehow completely forgot those hardware buttons.

Mcinwwl said: > Any voter can post ONE message in here regarding his choice and express his/her opinion Just for clarification for other voters: 1 post below per member, but above in the poll, multiple names can be chosen at once :-) After clicking on "submit" it's over, afterwards it's not possible anymore to change the vote.. Mathwiz said: > I discovered (too late for most polls) that these polls are set up to let you > choose more than one name! So if you like two or more names, go ahead > and vote for all the ones you like. (Name with most votes still wins, of course.) Yeah, not alone, when the first poll was reset, it had 14 votes from 13 members...

Bruninho said: > Fx36 does have the correct system UI font but it still shows in > some parts the "black cursor" thing as seen in SeaMonkey. Do you think that could be the 3dot sign "..." ? Guess this single-character is missing in older fonts.

Whoa... nice font! ;-) Wonder if perhaps it's looking for a non-existing one. You can save the time of struggling with an Opera12.02+18 mix, it had turned out a dead end and misunderstanding a few months later. Bruninho said: > When I try to load the main brazilian news site (https://www.globo.com) after it > loads most of the site, Opera crashes and shows an error report Aargh, what's that stupid new fashion lately of all the time crashing instead of simply showing a cipher-error?! That's an absolute killer bug. For continuing I'd recommand to first try roytam's fx36 (FF3.6) instead of KM, since Fx is already a complete build. KM will need some fiddling first. And it's engine is older too (FF3.5/3.6) But hope you'll somehow manage to get current modern kernelex running younger browsers again.

Bruninho said: > Menus on SeaMonkey 2.6.1 are displayed and working as usual, what I am saying is > that the font family used by the menus are wrong and different from the System wide > font family used by Windows. I can read and use them normally, just the font family is wrong My guess would be it's probably just a normal XUL css font-setting, like everything else in Firefox&Co GUI. Am definitely no expert for this, but to my knowledge most Mozilla browsers (except K-Meleon) use XUL for the GUI. That's by default completely independant from the system GUI, although it's possible to make the browser GUI inherit looks from the system GUI, for every single element (with {-moz-appearance: xxx}). Styling xul elements with CSS works very much like styling html elements with CSS. Somewhere inside the browser are the default styles defined (my guess would be: a file search for *.css inside the program folder, and if nothing found unzip omni.ja(r) and search again. Did this, just for curiosity, and files like browser.css or global.css look interesting, lots of definitions inside. I don't mean to change those, just could be helpful to get an overall idea and for figuring out element names) And those default styles are overruled by styles in the active skin (theme?) And those skin styles are overruled again when a user changes chrome styles in the profile. _Howto_ customize the menu font: this file must exist and be edited manually: ../currentprofile/chrome/userChrome.css By default gecko browsers come with a userChrome_example.css (just rename it), with such examples: menupopup > * { font-size: 15pt !important; } #urlbar {font-family: monospace !important;} THIS is probably exactly what you need! http://kb.mozillazine.org/Pane_and_menu_fonts /* Global UI font */ * { font-size: 11pt !important; font-family: Verdana !important; } Name of some main Firefox UI elements: http://kb.mozillazine.org/Chrome_element_names_and_IDs How to inherit SYSTEM styles: https://developer.mozilla.org/en-US/docs/Web/CSS/appearance wild guess on my part, only for 1 element type, not tested: menupopup > * {-moz-appearance: menuitem;} More examples with screenshots: https://www.userchrome.org/what-is-userchrome-css.html Yet more example links... https://www.howtogeek.com/334716/how-to-customize-firefoxs-user-interface-with-userchrome.css/

@Bruninho: can you post a screenshot of your probs, especially that menu font stuff? I keep suspecting that either your KernelEx compat setting on the browser exe may be culprit, or else some missing VisualC-VS-Msvc...thingee files. While you're at basic KernelEx again, would create a bunch of new browser profiles with it, for various browsers and engine generations, to later copy over to newer KernelEx again. Just in case. But first necessary to open httpS-sites with those browsers with OLD KernelEx, to create the necessary 3 new files (in gecko key3.db, cert8.db, secmod.db) When using the profilemanager it's quite easy to create several parallel profiles, at startup a menu will show up to choose one or create new ones. Not sure which commandline, but alternatively can be switched on in profiles.ini too, line "StartWithLastProfile=0" If Opera is missing ciphers, and FF9 also has probs, would perhaps try next roytam's fx36 - Firefox 3.6. It even comes with a build-in profile already, but this only gets used when started with the contained bat file. By the way roytams TLS for Netscape9 also works in K-Meleon1.6 (era FF3.5) KernelEx compat must be set to XP. Not sure if of interest to you, it's quite quirky and most people prefer Firefox. Just taking the opportunity for some general hints if anyone interested, now or later. One of the catches is that the last official version on the homepage is only beta2, which is even older as later community updates. They had updated the engine to its last gecko sub-version, and fixed some other things, but those downloads are now broken too. The most important (partial) update, still by main dev, and luckily still available, seems to have been posted ONLY here, very well hidden: http://kmeleonbrowser.org/forum/read.php?1,122806 (at the bottom of the page, last link) http://kmeleon.sourceforge.net/files/KMeleon1.6.7z And of course roytams new TLS1.2 ciphers recently! http://kmeleonbrowser.org/forum/read.php?2,151536,151817#msg-151817 http://o.rths.ml/gpc/files1.rt/ns9-nss-update.7z Attention: sqlite3.dll must be removed from this zip in K-Meleon, it must keep the old original file or crash! (almost forgot, but thinking about... probably in Firefox/SM too??) Hmm... and IIRC way back then KM1.6 was actually rather tricky to install too. Major struggle - or well possible I just confuse that now with the first alpha versions?? Not sure anymore. In the beginning was missing again some pesky msvc... file (and of course meanwhile its download link to MS on the KM-homepage is broken too, sigh. There are alternative smaller VisualC packages linked somewhere, would have to search around, but probably not necessary) And after the last community engine update my system had kept crashing the browser with an error message for imglib2.dll, had to replace that with an older version again (2011-06-24) Oh well. Another K-Meleon trap is that SourceForge (we suspect) is blocking KM-Forum access for all users with "Win98" in the useragent string!! To make things worse, they only show a completely cryptic error message, some nonsense with "pull a ticket to help yourself".... But it's really just the useragent, which can be faked.

Me neither. The right upper corner says "15 members have voted". That obviously means the first 13 are still counted. BTW, is there a deadline for those polls...?

@Mcinwwl: still technical probs it seems ;-) Just a thought: the header says "Pick a name", but the poll allows multiple checks. Something like "pick all names you like" would avoid possible misunderstandings.

Bruninho said: > I need to figure out how to use Dependency Walker and understand its functions. Well, personally my understanding is most simple: open any exe or dll file with it. If all dependencies are okay, lots of black text shows up. Great. If dependencies are broken, it will popup an error message and missing or faulty files are marked in red. Those may need a newer version, or are perhaps missing completely, but clicking them reveals more info. Bruninho said: > As for the browser, yes it's https pages that are crashing the browser. I remember that when > I managed to make it work, I had added some NSS files by roytam1 for Fx10 (since > SeaMonkey 2.7.2 is based on it). Now it just doesn't work anymore. Uh oh... sounds like major trouble, if the browser immediately crashes instead of just showing normal error messages. Seems highly unusual? Perhaps newer NSS files help, but even then you'll still run occasionally into an odd site with missing certs and ciphers. But yes, it's probably a good idea to continue testing with an updated build. Just remembered: the culprit may very well just be Firefox10 in Win98! I was assuming that Kex24 were now lightyears ahead of old basic Kex, but perhaps not in this case. Firefox9 is a lot more stable. What I'm especially curious about is how KM-Goanna74 (FF24) runs in your "most modern" kernelex... In my older v16 it's very buggy and crashy, for example when first loading a local page and then from web, or vice versa.

@Mcinwwl: Thanks for fixing the NM28 poll. As I understand the current polls, they are for finding out "which" names might be liked, which is the big "??", not about IF or if NOT rebranding. About this point some people have already expressed their opinion loud and clear and repeatedly, nothing new there. But if they insist to get a checkbox to cross too, here's a suggestion for a specific YES/NO poll. Just to show opinions of course, all decisions remain up to roytam1. Who did declare after Tobins deal offer that he'd be ready to accept and rename, just that others shall please do the tedious work of finding names etc. not sure about wordings, just a (slightly too verbose) draft... - YES! roytams forks are different enough to deserve own names long since - YES. Makes sense. I understand browser confusion happens too easily - Yes. Wouldn't need new names, but really like Tobin's offer - Yes. Okay with me, just please keep new names similar to old ones - (yes) Would have kept ori names, but helping each other is better than struggling - (???) Can't decide... - (no) Slight preference to keep old names, but won't boil with rage if renamed - No. Because modified names will cause more confusion as ori ones - NO. I see zero reason to change names, and expect no future probs - NO! It's not true old names cause anyone probs, and don't dare to threaten us!

@Bruninho: Dependency Walker is a great tool for checking which files may be missing or causing probs. Not likely, but who knows: are all httpS-pages crashing your browser or not?

YIKES!! Slightly shocked now (again)... Bruninho said: > When I try to load sites like http://lite.cnn.io, it loads quickly and as expected. > Probably SM 2.7.2 can't handle heavy sites... but google.com isn't a heavy site! Your example site is only "http", that reminds me that my current KernelEx made it impossible to create in new browser profiles the necessary files for https-access. Is it incidentally or are perhaps all httpS pages broken? Luckily I had various old profiles, and just copying over 3 old files helped (cert8, key3, secmod). Am just not sure about possible side-effects. But my Kex is v16/17, yours 24, so guess this prob was already fixed? Or is due to a prob with old browser version? Perhaps works better with retrozilla or roytam's special builds? Other usual suspects are missing MSVC... files, or lost KernelEx compat settings after renaming folders or files...

Uh oh.... talking of user agent strings, just had a thought that gives me cold shivers, if true - but is it?? Guess most of us here are aware that certain, extremely important websites are fully intentionally blocking certain users, for example those using old browsers or those blocking dangerous features, by either denying their pages completely or sending a crippled version only. Happy greetings for example from enforced TLS1.2 everywhere for even the most harmless public content like wikipedia etc., or crippled youtube, github and other sites. Now recently there was that scary announcement that all major browser creators, led by Google, had happily agreed with extremely short notice, to kill the traditional, easily tweakable method of useragent data transfer, which had allowed to get most of those blocked/crippled pages delivered anyway. And they will instead replace it with a new method which is not fakeable anymore for users. Plus (IIRC) this new method will secretly deliver much more user data, about our browsers and probably systems, and no one knows what more. And the best, those user data are guaranteed to be true values. And one can expect it will be near impossible for non-experts to even realize WHAT exactly their browsers are now revealing secretly to all servers and trackers. AND they said, this new method will actively PULL those data by "sending a request" to our browsers. (perhaps not by a few remaining friendly sites, but today's main sites as usual, all tracking) Now my question is: "Request"...? How exactly?? What does that mean... involve a method like XHR perhaps? (xmlhttprequest) Or not?? If yes, does that mean such evil sites can abuse this method now to FORCE all visitors to allow global javascript, even XHR, or otherwise their servers will just happily refuse to deliver anything, or at best just cripple pages? Finally the perfect roundabout killer for most effective security, safety and anti-tracking measures, kicking out all such 'rebellious' visitors? Not to mention freezing weaker computers? Yeah I'm aware many sites do cripple content for noscript visitors today already by simple traditional methods, because they want or just don't care. But when that UA-Pull-Method will become standard this will happen on a whole new scale. Almost everywhere, like TLS1.2 was enforced too, and considered perfectly normal again to block anyone with older browsers (=with not the latest state-of-the-art spying features built in), without any real need. If yes, if JS and perhaps XHR will be required (?), or perhaps an even more intrusive and dangerous new method (what a scary thought, tons of new backdoor chances!), will there be any new tricks possible to circumvent that?? Although those might be very hard to use for non-expert users? And even if there will be such new tricks possible, they will almost certainly only be possible in the most modern browser versions too? Not possible to add by a simple addon for older browsers? Anyone already working on that prob...?

Would like voting options for "RT-Moon" and "XMoon". The other names are all unique, were not intended to get any prefixes, that just wouldn't fit. The prefixes were meant to be followed by the original core names, like "-Fox" or "-Moon", at any rate their heritage clearly recognizable. So the other names are standing for themselves, but "Moon" alone makes no sense, it only works "with" a prefix.

thumb pics on youtube search page: No idea if that helps you anything, and which view-variant it needs, but in my old browser (KM1.6/era FF3.5) with old UA this works to load more thumbs: javascript:(function(){var LIST=document.querySelectorAll("img[src*='/yts/']"); for(var j=0; j < LIST.length; j++){LIST[j].hasAttribute('data-thumb') ? LIST[j].src=LIST[j].getAttribute('data-thumb'):0;}})()

Sorry but there seems to be a misunderstanding: KM-Goanna IS and has always been a unique name for roytam1 forks of K-Meleon with Goanna engine! Introduced since about 2 years. And I most definitely did not mean "RT-K-Meleon" as a name suggestion, was only meant as reference. Perhaps this topic should rather be deleted?

Mcinwwl said: > Around 40 to implement all sugestions from siria and msfntor. Please change mine to: RT-Fox (RT = Family Prefix, if ok for roytam1?) RT-Moon RT-xxxx (?) XFox (X = Family Prefix) XMoon Xxxxx (?) Moonix MoonRay (All my "fox" are only meant for direct Firefox forks, like Fx3.6) (All my "moon" are only meant for PaleMoon/NewMoon) Current list for NEWMOON only: Matt A. Tobin New Moon -> Whistler Siria RT-Moon (RT = Family Prefix, if ok for roytam1) XMoon (X = Family Prefix) Moonix MoonRay Mathwiz New Moon -> Triton dencorso New Moon -> Peach Moonshine msfntor New Moon -> Noctus, Nocturnus, StrongMoon, MoonDream, NextMoon, NeoMoon, MoonBird, MoonLight, LightMoon, LastMoon, LateMoon, WinMoon, FoxyMoon, Foxtrot, RexMoon, RexyMoon, ... TechnoRelic RT*[insert browser name] "RFox CORDELIA" = RT NM28 Browser Monroe New Moon (?) -> MoonBeam New Moon (?) -> DeathRay New Moon (?) -> MantaRay LoneCrusader New Moon (?) -> Lunaris Destro New Moon -> Wolf Moon Just thinking: if technically possible please allow multiple votes in 1 poll per user

Bruninho said: > KernelEx is updated to 4.5.2016 according to the verify.exe, although I got it from > here: https://retrosystemsrevival.blogspot.com/2018/05/kernelex-45201617-updates.html > which says a different version. Thanks for posting this! Although I usually prefer to download from original topics on msfn here, but long since am not sure anymore exactly howto install KernelEx cumulative updates. Admittedly also haven't really researched it yet nor studied the first page lately, since last time my update attempts were a nightmare, and probably not getting around anytime soon to such a major undertaking again. But now just took a quick look at this blog page, noticed it has detailed installation instructions, and an easily accessible download, which hopefully is a complete compilation of everything after basic 4.5.2? And it contains an update.bat file, which I'm not necessarily planning to use, but its content now has my hopes flying high: There's only a bunch of new files to drop into windows/kernelex Folder? Is that really all, no registry mess to fear or anything landing in the normal Windows system folders (neither manually nor automagically)? Which hopefully means if anything goes wrong or doesn't work, I could simply copy back the previous files again in DOS? Not needing to do a backup of whole Windows folder again?? If so, an update seems finally in reach again... Regarding this update.bat, have no clue about DOS and syntax, but wonder if there's a typo here? if exist %src%\*.exe xcopy %src%\*.dll %dst%\kernelex\ /-Y

roytam's NewMoon26 and KM-Goanna74 (same engine) are compiled for Win2000. NM doesn't run on my 98se, but KG74 looks fine if KernelEx compat is set to Win2000 too, as was already mentioned by others. My menu font looks normal, only KG74 menus in general have some killer bug on 98, under certain circumstances causing crashes (e.g. after deleting a menuline dynamically during session, or if a menuline contains forbidden(?) special words, etc). But that's only one of several killer bugs, and strongly suspect those are rather 'normal' and fixable KM bugs, not necessarily win98 related. UI-fonts: no clue, but if prob remains and no real fix or better suggestions appear: in Mozilla browsers with XUL-GUI (=not KM) adding a special css setting in profile/chrome/userChrome.css may help? Or try a tweak in "ui.css" in default stylesheets? (path see post above, details search web ;-) Tip for win98 fancy WEB fonts: to see most (not all) fancy modern characters and emojis on current webpages, mainly on instagram, the widely recommended Symbola font works wonders. If it's Latin characters wouldn't look so absolutely awful, may even choose it as main font, not just as automatic fallback font for missing characters. Must be copied into c:/windows/fonts folder, and then in K-Meleon chosen in F2 > PageDisplay > Fonts for several language groups in the dropdown menu, especially for Western, Other(UNICODE), Cyrillic, Central European or whatever is browsed. For each, have font Symbola chosen in Cursive+Fantasy fields, that seems to work as automatic fallback if the main font misses single characters. Not sure how this works exactly and if both Cursive+Fantasy are needed, just glad it works. At least in KM. Also found 2 more emoji fonts which work, but overall they have much less characters as Symbola, so sticking with it. Pref example to test in other Mozilla browsers: font.name.cursive.x-unicode = Symbola font.name.fantasy.x-unicode = Symbola etc Also saw this in my aboutconfig now, no idea if native or perhaps once added manually for testing, or perhaps THIS may be causing the automatic fallbacks? But can't harm anyway: font.name-list.serif.x-unicode = Times New Roman, Microsoft Sans Serif, Symbola Thanks for the Opera userstyle/userscript links! Will try to learn it when getting around. For the moment still a bit postponing since it's a lot of work and old Opera currently only my 3rd fallback browser, but userstyles (css, e.g. for msfn) and userscripts (js) are absolutely essential functions today. Always so much great info here and help stuff, very glad about it, but sadly getting around to learn it far less as I'd like too, sigh. And if not needing it in main browser then so quickly forgotten again. Always hope this forum will live forever and can look it up later again when some day needing it urgently ;-) Google SEARCH width: No idea about width on other computers, but just in case: that site depends strongly on useragents, at least the available options. After their recent server changes had to set ua FF10 to still get all menu options, older and younger didn't work, and fiddled with some minor css tweaks again. Still cursing though that the number of total HITS is now completely inaccessible, huge prob! Find absolutely no workaround - anyone a tip...? Regarding WIDTH, mine shows also about 736px. No prob since that looks fine on 1024, but of course, if people with GIANT screens get the same one, that may be a bit tiny... Just for curiosity did a quick test what affects width in my current browser, and found this to work: body, #main {width: 95% !important; max-width: none !important; }

Frankly I just tend to simply RT-xxxx lately, since that's what we all will keep thinking and remembering anyway. Or how many people will be able to memorize completely different names, and additionally will remember which is based on which, if new names don't even remotely remind the origin? Let alone the rest of the world, completely clueless about such history. RoyFox/RoyMoon etc. is out, and I agree that feels a bit like too much ego. "RT" on the other hand feels rather factual, more like a symbol as a real name.

Thank you for confirming that layout fix works in other browsers of that era too. Note: As mentioned in a previous post, this userstyle for browsers of Firefox2-3 era engines is adjusted to my screen width of 1024px, since I couldn't figure out how to make it use "100%" of screen. That means the width is hardcoded for 1024px, using fix numbers, but of course lucky people with wider screens can easily increase the numbers to their own screens. The CSS code can be pasted in the personal userstyle sheets, whichever anyone uses. Am only familiar with Mozilla/Gecko browsers, and as Wunderbar98 explained already, their default path is .../yourprofile/chrome/userContent.css and if not existing, the file can be created with an editor like Notepad++, or look for "userContent-example.css" and change the name. Am quite optimistic that userstyle also helps Opera12.02, since the forum looks exactly the same as in my old Geckos. Just have no clue how to handle that browser and couldn't figure out yet in a first attempt how to enable personal userstyles there, or what I may have done to prevent them. Funnily found lots of css-files and also checkboxes to enable them and set folder paths, and settings for authorstyle and userstyle and whatnot, but still nothing got injected. Surely my fault, but didn't try further since others know this better anyway :-) Browser's DEFAULT stylesheets Looks like those are very easily accessible in old gecko browsers! In FF2+3.6 here: ../programs/browsername/res/***.css There are several, the most fundamental one being "html.css" In younger geckos they are also contained, just unzip omni.ja(r) and do a file search for "html.css" The advantage of those "default" styles compared to "user" styles is that those are not overruling the original styles of websites, only provide initial settings if none exist at all yet in old browsers. So far in a quick first step have only compared "html.css" in K-Meleon74 (FF24) and KM1.6 (FF3.5), then copied over a few html5 tags into KM1.6, plus a few random younger tags seen on caniuse.com. But that's just a start, update from FF3.5 to FF24, and fixes hardly anything yet, since most new tags are simple "block" features, just with more or less margins. Some day will have to compare new tags in younger FF versions too. Hardly anything useful here yet, except mainly "mark", a new tag for highlighted inline text. And "time" is inline too. mark { display: inline; background: yellow; color: black; } time { display: inline; } article, aside, figcaption, footer, header, hgroup, main, nav, section { display: block; } /* added to existing blockquote */ figure { display: block; margin: 1em 40px; } /* hidden elements */ /*#### html5 / only added DATALIST+TEMPLATE from KM74 */ area, base, basefont, head, meta, script, style, title, noembed, param, datalist, template { display: none; } /*#### only table border color / copied from KM74 */ table[bordercolor] > tbody, table[bordercolor] > thead, table[bordercolor] > tfoot, table[bordercolor] > col, table[bordercolor] > colgroup, table[bordercolor] > tr, table[bordercolor] > * > tr, table[bordercolor] > tr > td, table[bordercolor] > * > tr > td, table[bordercolor] > tr > th, table[bordercolor] > * > tr > th { border-color: inherit; }