00:24:00 https://kb.vmware.com/s/article/1023856 yay VMWare 00:24:25 the KB renders blank pages except for a "Support Assistant"balloon 00:24:38 the console registers a bunch of noise from their framework, "Aura" 00:25:00 including a unhelpful "TypeError: header is null" (logged by them!) 00:25:32 also the actual page sourcecode contains NOTHING that could be extracted back to something readable, it's all JavaScript down there! 00:29:19 Many websites nowadays just refuse to work if the browser is not "whitelisted" 00:31:29 so this is why those VMWare KB pages actually work through the Google cache 00:31:46 except that these days, Google tends to hide the "view cached" button 00:40:29 Where did they move it? 00:42:34 it's hidden behind a dropdown arrow in the little info box that opens if you click the 3 vertically stacked dots next to the page URL in the results header 00:42:45 note, this only applies for this week 00:42:56 maybe by next week it's completely gone, or misplaced 00:51:05 google cached button has never worked for me 00:51:12 how did u guys make it work? 00:52:25 first you have to pretend you're "recent Firefox" 00:52:36 otherwise GOOG will serve you cellphone pages 00:52:37 i never see any "view cached" behind the 3 dots. only some more options panel which is meaningless tbh 00:52:52 franstam: on that panel there is a "V" in a circle 00:52:55 click that 00:52:59 just discovered it today 00:53:02 oh now i see it 00:53:03 thanks 00:53:08 bullshiet tbh 00:53:23 Move! Fast! And! Break! Things! 00:53:26 is goog scared that we find their sekret cache? 00:53:35 scared that we use up all their bandwidth? 00:54:11 the problem is that cached pages get special treatment from the source servers by dropping down a notch or then on the JavaScripts 00:54:24 for example the VMWare KB serves perfectly flat pages to Googlebot 00:54:35 and encrypted JS framework vomit to the rest of us 00:55:04 and of course, we MUST eat the JS framework vomit, NO EXCEPTIONS, hence why GOOG now pretends the cache is gone 00:55:15 oic 00:55:24 its just to hide their incompetence 00:55:44 makes purfect sense 00:56:11 i dont mind mobile pages tbh. less js vomit u say 00:56:26 give me mobile pages then 06:24:28 Sompi: I can read fingerpori, but hesari's front page indeed blanks to me if I have javascript enabled :-\ 06:25:13 Sompi: usually I think this kind of thing suggests a coding error or lack of error handling in the javascript code 06:26:31 "JavaScript error: https://www.hs.fi/_next/static/chunks/main-b8e2658307c68cb0.js, line 1: TypeError: e is undefined" for the front page. 06:26:44 all praise the minified JS gods, who brought us meaningful line numbers 06:29:12 tomman: lol. This makes VMWare KB serve content too: Mozilla/4.0 (compatible; MSIE 6.0b; Windows 98) 06:30:07 I'd love mobile pages, except google pushed somewhat to have "js vomit" and "css with lots of whitespace" be the new "mobile page" 06:30:33 now back to .fi - what else is broken besides hesari? All of Sanoma? More than Sanoma? 06:32:40 at least one site from Alma and one from Savon media appear to load fine, but maybe I'm just checking the wrong pages 06:34:19 Ilta-Sanomat (Sanoma) blanks, Ilta-Lehti (Alma) does not blank 06:34:34 in fact, ilta-sanomat blanked while displaying the cookie "popup" 06:34:48 as in, it still showed the popup, but the page was already blank 06:36:36 aamulehti.fi blanks, also with a Sanoma cookie "popup". 06:38:55 turun sanomat and kauppalehti load OK too 06:39:31 so far: seems to work without javascript; seems to happen only on sites managed directly by sanoma media, but it might be that I'm not looking widely enough 07:04:30 it seems to involve performance.measure, toggling the settings didn't change it, could it be that they have the same issue as Google Translate full-page translation (requiring the new performance.measure spec) *but* don't have a fallback at all if performance.measure is not present? 10:21:12 njsg: Those errors in the console also come up with Firefox but the page doesn't blank 10:49:46 Sompi: same typeerror on the same js file? can you get the column number in the devtools inspector for firefox, just to make sure it's the same? 10:49:53 * njsg is not sure he has firefox ready nearby for testing 10:50:25 the column number put me in a place where it's doing performance measurements, so that's why I assumed it was that 10:50:43 of course life would be no fun if issues were always that simple to identify 10:52:34 Does it work on Seamonkey if the user-agent string says that it is Firefox? 10:55:35 I didn't try with a "newest firefox" UA, just some firefox UA, possibly too old, that didn't work 10:56:18 "identify as firefox" didn't help 10:56:20 These are just news sites with mostly text content and some images. They should work on all browsers... 10:56:47 also, why is this not happening everywhere on the sites? 10:57:03 I think a fingerpori strip didn't get blanked 10:57:37 works too: https://www.hs.fi/viivijawagner/?ref=snavi/ 10:57:55 Some random minified javascript library probably blanks the page if some unknown-to-us conditions are true 10:58:15 And things like these are extremely hard to debug as a browser developed 10:58:17 developer* 10:58:51 I'm thinking of how I'd like to approve network interaction by network interaction in devtools, kind of like Windows 4.x interactive config loading 10:59:16 And the page may actually be working just like intended. They are breaking it on purpose if someone uses a browser that they haven't heard about 10:59:20 but to be fair it'd just be quite simpler if the javascript weren't minifed and so on 11:00:27 my guess is that this is more "but what do you mean browsers might not support [this] or [that]? it has been in [browser I use] since [X time ago], why would anyone use anything older than that?" "you know there are browsers beside [one] and [another], right?" "what, lynx?", but could as well be intended breakage too 11:00:57 Most websites nowadays have literally megabytes of minified javascript, and the page itself doesn't have any contents that couldn't be shown with static HTML and CSS... 11:01:54 It is crazy. I wrote an operating system that can go online and host a HTTP server, and it has much less code than an average website has nowadays. And the code isn't even minified 11:02:59 https://oma.sanoma.fi/aihe/yhteystiedot/ fails too, but the weather section of Hesari (which redirects to IS?) works. 11:11:58 njsg the site uses dynamic imports. I am on it but not around the corner. My updated SpiderMonkey dislikes the new regexp patches. 11:12:03 one of the pages that does not blank had a message about "ad slot cover not found", but maybe that's just a coincidence 16:11:30 well, it IS relevant - for the FOUNDATION (cause they have to do /something/ to support their existence), but there is no relevance, otherwise, to me or you or to anything browser related