[13:59:38] Why there are constant login timeout errors? [14:00:02] Is it related to some updates on CentralAuth system? [14:00:49] timeout? As in your being logged out more frequently than you expect to be? [14:01:32] Also log out where? IDM or SUL? [14:01:57] T372702? [14:02:16] I constantly cleanup my cookies, except the ones on login.wikimedia.org. As expected, other sites can detect my central login status (re @wmtelegram_bot: timeout? As in your being logged out more frequently than you expect to be?) [14:02:46] And automatically log back in after a refresh [14:04:12] Over time they can get very very large. I’m using chrome on MacBook Pro (re @cvictorovich: I constantly cleanup my cookies, except the ones on login.wikimedia.org. As expected, other sites can detect my central login st...) [14:05:28] SUL (re @Ladsgroup: Also log out where? IDM or SUL?) [14:11:42] Seemingly okay for now. Cannot figure out why [14:12:38] Seems that "Download as PDF" is broken on en.wp, de.wp, meta, and possibly more sites. Is this a well-known issue, and I'm just slow the realize this? [14:12:39] https://phabricator.wikimedia.org/T375521 [14:28:32] also needs a friendlier error presentation. (re @fuzheado: Seems that "Download as PDF" is broken on en.wp, de.wp, meta, and possibly more sites. Is this a well-known issue, and I'm just ...) [14:41:11] I now wonder how long this has been an error, and if there's no fix in sight, we should remove it from the standard interface for a top 10 most visited web site in the world. [14:41:12] Oddly, Google Gemini AI says exporting PDF from REST API has not been working since October 2023 but I cannot find a definitive link or explanation of how it arrived at that conclusion. (re @jeremy_b: also needs a friendlier error presentation.) [14:41:29] https://tools-static.wmflabs.org/bridgebot/a82d9f55/file_65125.jpg [14:42:01] what does UI to export mean? [14:42:18] I'm not sure :) I was puzzled as well (re @jeremy_b: what does UI to export mean?) [14:42:32] It seems a bit odd that it would've been broken nearly a year without any apparent bug reports [14:42:55] Agree that it's odd, but not completely crazy (re @wmtelegram_bot: It seems a bit odd that it would've been broken nearly a year without any apparent bug reports) [14:43:44] I do tend to just use printable version and the browser PDF functionality [14:44:49] The logs would suggest it was probably working yesterday [14:45:23] https://grafana.wikimedia.org/d/000000309/electronpdfservice?orgId=1 [15:06:20] Yeah, it seems it may have only been broken ~12H [15:10:26] Working again now [15:20:10] PDF was working fine, i test it quite regularly. Dont trust gemini, its answers are terrible. (re @fuzheado: I now wonder how long this has been an error, and if there's no fix in sight, we should remove it from the standard interface fo...) [15:22:26] Yeah, native print to pdf works just fine. Ppl just dont know how to use their devices ;) (re @wmtelegram_bot: I do tend to just use printable version and the browser PDF functionality) [15:24:28] It is amusingly inefficient when it's just running chromium in the cloud for them... [15:24:49] Ive suggested a print to pdf flag for JS print() a while ago, to either a working group or some of the browser vendors.. [15:24:50] that should mostly eradicate the need for separate PDF services completely. [15:25:30] But i never followed up or actively promoted it, so theres provably not been any progress [15:35:40] @Krinkle24 ^^^ interested in helping me push for that after I get back ? [15:40:44] Agree that Google putting their Gemini/AI answers at the top of the SERP page seems like desperate "metoo-ism" given how poor the results are! (re @djhartman: PDF was working fine, i test it quite regularly. Dont trust gemini, its answers are terrible.) [15:49:38] https://youtu.be/ffKEL1eXfTI (re @djhartman: PDF was working fine, i test it quite regularly. Dont trust gemini, its answers are terrible.) [16:22:24] "Phew! I quit smoking last year and will adjust accordingly!" (re @Ladsgroup: https://youtu.be/ffKEL1eXfTI) [16:52:42] Is anyone else experiencing this issue (re @cvictorovich: Why there are constant login timeout errors?) [16:53:27] I don't think a lot of people copy your specific practice of clear cookings expect X [16:54:14] Indeed; there's known logout issues... But randomly deleting cookies regularly is unlikely to be helping [16:55:24] Cookies indeed stack up a lot sometimes [16:56:21] Next time I may share a screencast if I encounter again [16:56:39] I don't think Wikipedia et al are creating MB of them, for example [16:57:20] But indeed it happens; sometimes 100+ MB (re @wmtelegram_bot: I don't think Wikipedia et al are creating MB of them, for example) [16:58:27] 100MiB of cookies attached to a Wikimedia wiki seems very, very, very unlikely. I guess maybe a user script or gadget could be causing that, but not the core software. [16:58:57] LocalStorage is a different thing than cookies and that will get used heavily [16:59:10] I can share details if you like [17:00:52] Maybe not publicly [17:01:41] Not now: the situation isn’t happening now [17:02:04] After sometime I may reproduce it