19:17 <jaeger> I'll revert the XDG stuff in /etc/profile today in favor of pam_xdg for 3.7, guessing by now nobody is going to object to that

19:18 <beerman> pam_xdg?

19:18 <beerman> i wasn't following along to closely the last days

19:18 <jaeger> That or manually, whichever users prefer

19:19 <jaeger> https://www.sdaoden.eu/code-pam_xdg.html

19:19 <beerman> yeah, i am not using that..

19:19 <jaeger> I didn't think you would, heh

19:19 <jaeger> But it's an option

19:20 <beerman> ¯\_(ツ)_/¯ i guess

19:20 <jaeger> If you want to read back in the logs for the discussion before I make the change back, feel free

19:20 <jaeger> I'm not in a rush, working today anyway

19:21 <beerman> i mean, pam_xdg also handles the critical /run/user/dir, same as core/dumb_runtime_dir in 3.7

19:21 <beerman> the rest is quite optional in 99% of cases, just nice to give a structure. we've been talking about "nice things to add which cost literally nothing"

19:22 <jaeger> yeah, the concern as I understand it is not about whether or not it's nice to have, but where to put it

19:23 <jaeger> /etc/profile not necessarily making sense in all cases (non-bash shells for example)

19:23 <beerman> wasn't somebody recommending pam_env.so?

19:23 <jaeger> Possibly, I don't recall. I have no doubt that would also work

19:24 <beerman> non bash is non standard and people going away from the standard might want to keep such cases in mind. but sure, fair enough

19:24 <beerman> i would prefer pam_env then, as it avoids pulling another package

19:24 <jaeger> Yeah, I probably would also use pam_env in this case since it's already provided by linux-pam

19:24 <beerman> also i fail to find how pam_xdg handles the full spec

19:25 <jaeger> No idea here, I've never used it or looked at its source, just saw that it was recommended

19:25 <jaeger> You can take it up with stenur if you want, heh

19:25 <beerman> if its about /run/user/$UID, i choose dumb_runtime_dir as its less loc and does just that, nothing more or less.

19:25 <beerman> i don't really feel like i need to take it up with him

19:26 <jaeger> pam_env also allows for individual user configs in ~/.pam_environment so everyone can customize as they see fit and not rely on defaults

19:26 <jaeger> Not surprised... I'm just saying I don't know pam_xdg's operations under the hood, so to speak

19:29 <beerman> i know, i know.. both modules should virtually do exactly the same, but ifreuds version is tidier, imo.

19:33 <jaeger> To be clear, I'm only talkinga bout removing the recently-added XDG_* vars from /etc/profile in bash, not forcing a choice of pam_env/pam_xdg on the user... each user is, as always, free to choose whichever way they want

19:34 <beerman> yeah i don't mind

19:34 <jaeger> ok

19:34 <jaeger> The subject might be worth a wiki article, in fact

19:35 <beerman> yeah i am not writing any wiki articles anymore, sorry, that pmwiki syntax is.. just horrible

19:36 <jaeger> I'm not particularly in love with any wiki software, if you want to try something different, fine by me. I think pmwiki was jue's choice, though I might be wrong

19:36 <jaeger> I do hate the double-click-to-edit behavior but as I learned recently that's a feature which could be disabled.

19:37 <beerman> heh

19:37 <beerman> yeah it was

19:37 <beerman> i mean, i kinda get it, and i kinda not get it. but i don't want to change everything just because

19:37 <farkuhar> I'd be willing to contribute some wiki writing, or help with updating the handbook. The pmwiki syntax looks daunting, but I've gotten accustomed to worse.

19:38 <jaeger> speaking of that, I said I'd look at sharing the handbook source and I haven't forgotten that... just keep getting busy with other crap

19:38 <beerman> i have asked for a markdown syntax addon a while ago, jue promised to look into it

19:38 <beerman> markdown is so much easier for everything

19:40 <jaeger> https://crux.nu/~jaeger/handbook/

19:45 <farkuhar> thanks, jaeger. Wow, does pmwiki actually go through the entire commit history to construct the source code that gets turned into html? That's a bizarre design decision, if true.

19:46 <jaeger> I think only the current and previous versions so it can show a diff... but I'm not sure, haven't looked closely

21:02 <farkuhar> on the subject of "nice things to add which cost literally nothing", one cost of these niceties is that users don't acquire the knowledge of how the upstream tools behave *without* distro customization, as frinnst had to educate someone here: https://lists.crux.nu/pipermail/crux/2019-September/006396.html

21:25 <beerman> well, i didn't exactly take this for a teaching distro either. placing them with a link to a spec however i felt was fair enough for CRUX to do

21:25 <beerman> ¯\_(ツ)_/¯

21:26 <beerman> jaeger: fwiw. XDG_RUNTIME_DIR we still might want to add by default

21:26 <beerman> one way or the other, i'd say pam_env then

22:15 <stenur> btw, .. glibc 2.36?

22:26 <stenur> pam_xdg injects all variables the XDG standard 0.8 defines; using pam_env i think will do the entire configuration file dance. But honestly i am more for peach than beer