| rwp | Xenguy, JFRT that X terminate key sequence has always been Control-Alt-***BACKSPACE*** not Control-Alt-Delete which is different key. | 00:47 |
|---|---|---|
| rwp | It's disabled by default now. Because... Pranks! But when enabled it will IMMEDIATELY terminate X and of course anything running in X. There is no dialog prompt asking if that is okay. It just drops immediately. | 00:48 |
| Xenguy | rwp, Aha, good point, I stand corrected (in the old daze on the old OS C-A-D used to reboot the system) | 01:11 |
| Xenguy | Do you happen to know the file that would get configured? | 01:11 |
| fsmithred | dpkg-reconfigure xserver-xorg I think | 01:12 |
| rwp | I put that in ~/.xsessionrc for use by slim, lightdm, and the other xdm DEs. | 01:12 |
| Xenguy | So you configure that file for C-A-B to restart Xorg then? | 01:13 |
| Xenguy | fsmithred, Always appreciate the tips, so much I don't know | 01:14 |
| rwp | Yes. Configuring X is running commands and "setxkbmap -option terminate:ctrl_alt_bksp -option compose:menu" is a command that can be run from the command line. | 01:14 |
| fsmithred | ^^^ is that what goes in ~/.xsessionrc? | 01:14 |
| rwp | I also run xmodmap which is the somewhat older way to configure the keyboard. setxkbmap resets xmodmap settings. So I first run setxkbmap (which might possibly do everything that is needed) and then run "xmodmap ~/.xmodmap" to finish up. Then additionally I put "xset r rate 275 45" to adjust the keyboard repeat rate to my liking. | 01:15 |
| Xenguy | Looks like I have no such file on this system, so I'd have to set it up afresh | 01:16 |
| rwp | fsmithred, When Xsession starts up it runs all of the parts in /etc/X11/Xsession.d/* which includes the /bin/sh sourcing ~/.xsessionrc if it exists. The syntax therefore must be /bin/sh syntax and not bash or csh or zsh or other shell. | 01:16 |
| rwp | It's the /etc/X11/Xsession.d/40x11-common_xsessionrc which does it if you want to look at the actual script that does the work. | 01:17 |
| rwp | Since we are talking about slim then ~/.xsessionrc is the shared file for the DEs to source. | 01:17 |
| Xenguy | Good to know, thanks kindly | 01:18 |
| rwp | And also for startx too. So people using startx will also have that work for them too. | 01:18 |
| fsmithred | thanks | 01:19 |
| rwp | I am not the typical person and I log into the vt console text mode first and then I run xinit which means I am actually using ~/.xinitrc file there. But anyone who is doing this is already old-school and going to know that they need to do things the old way. ~/.xsessionrc is the shared file for all of the newer X start-up processes. | 01:19 |
| rwp | I feel I should mention another file which has a similar name but a different purpose. I recommend not using it. I shouldn't mention it. But it has such a similar name that it gets confused with .xsessionrc because the name is .xsession which can be used in a roundabout way. I recommend to ignore it and stick with .xsessionrc which is well understood. | 01:21 |
| Xenguy | :nod: | 01:23 |
| rwp | Oh and I was wrong about startx which uses either .xinitrc or .xserverrc files but not .xsessionrc my bad for confusing those on startx. | 01:24 |
| freem | I asked here a while ago why I would not get coredumps with a program on my VPS. I suspected it was because of a crash handler. I just tried to nuke it (got busy in the meantime) and indeed, I was right. I do not remember who helped me back then, but maybe you'll recognize yourself. | 11:45 |
| hagbard | Q-Sig22rt | 13:08 |
| hagbard | wrong window | 13:09 |
| CueXXIII | freem: what does "ulimit -c" show? if you want core files and have the space, i suggest setting it to unlimited (ulimit -c unlimited) | 15:51 |
| freem | ulimit was correctly setup :) | 15:57 |
| freem | as said, the real problem was this signal handler which would prevent coredumps to happen | 15:58 |
| freem | because some daemon-engine contributor thought it was smart to put a signal handler even in the dedicated server binary | 15:58 |
| CueXXIII | ah, if your program has a handler for sigsegv then of course you get no coredump | 15:58 |
| freem | which is ... say politely, at least, quite dubious | 15:58 |
| freem | exactly | 15:59 |
| freem | I am trying to not say my impolite opinion, but I'm certain you can guess it | 15:59 |
| CueXXIII | yeah. but that's an issue with one particular program | 16:00 |
| freem | (and sorry for the probably incorrect word, I'm not sure I know the correct english one) | 16:00 |
| freem | yes | 16:00 |
| freem | which is open source, so I could patch it | 16:00 |
| freem | but someone here helped me on this, so I wanted them to have a chance to know the end of the story, even if after few weeks | 16:00 |
| freem | I kind of consider it a duty, that, when you asked for help in a place, when you get the final solution, you share it in the same place, even if unrelated to the direct topic | 16:01 |
| freem | this info might also be helpful to programmers idling around and who might create a daemon one day. Please: do not *not* catch signals without a good reason, and please give at least a build option to get rid of that, as it is directly harmful for debugging purposes. | 16:02 |
| freem | but really, my main intent was to thank the time someone spent to help me (including with the ulimit thing) by giving the full story | 16:03 |
| freem | I should have tried earlier, but got busy and distracted | 16:03 |
| CueXXIII | or maybe ehen have a runtime option/config to disable the sigsegv handler | 16:04 |
| CueXXIII | i can't think of a case where a segv would be expected in a well written program and not be the source of some security problem | 16:05 |
| * joerg absolutely agrees on >><freem> I kind of consider it a duty, that, when you asked for help in a place, when you get the final solution, you share it in the same place, even if unrelated to the direct topic<< | 16:51 | |
| freem | CueXXIII: yes, a runtime option would be ideal, but I do not have enough time to dive in the game engine's internals, I want to make a game out of it :) | 17:22 |
| freem | I believe the signal handler was placed for the *client* side binary of the game engine | 17:22 |
| freem | so that it is possible to show a dialog box to the user. This makes sense. Lot of. But... I'm rather annoyed by... bunch of things, including past drama between me and others | 17:24 |
| freem | let's be honest. | 17:24 |
| CueXXIII | freem: ah, a game engine could make sense. so the game can capture the coredump when it occurs and send it to the developer as bugreport | 17:24 |
| freem | and it is not drama in which I am innocent neither, I have had my misbehaviors | 17:24 |
| freem | no | 17:24 |
| freem | the signal handler is not made in a way that a coredump is done, otherwise I would not have had a problem with this | 17:25 |
| freem | it is done to only show a meaningless dialog box saying the game crashed with, usually, a completely meaningless message even for devs, of which I was in the past | 17:25 |
| cakebandit | I made a tutorial on installing Devuan on a Odroid, similar to a raspi4, its here https://forum.odroid.com/viewtopic.php?f=217&t=49070 | 17:26 |
| freem | anyway, I have resent and tend to be rather acidic, which is a stupid and childish behavior. | 17:26 |
| cakebandit | I was wondering why the urls in /etc/apt/sources.list are http and not https? | 17:26 |
| freem | cakebandit: what would be https interest in debian/devuan's sources.list? | 17:27 |
| CueXXIII | because anything is checked with signatures and hashes anyways | 17:27 |
| freem | the packages are each individually signed, in a much more robust and reliable way than the stupid chain of lies the interweb loves so much | 17:27 |
| cakebandit | oh, okay, I want sure | 17:27 |
| CueXXIII | and https used to be much more burden on the servers, and nobody changed it since then | 17:28 |
| cakebandit | *wasnt | 17:28 |
| cakebandit | thanks | 17:28 |
| freem | it is not unheard of TLS providers to be buggy or corrupted after all | 17:28 |
| freem | I personnally trust debian and devuan maintainers a lot more than those who maintain the TLS stuff | 17:29 |
| cakebandit | ok , I use letsencrypt for my web server | 17:30 |
| freem | the main problem with the web's "chain of trust" is that any part of the chain being corrupted corrupts the whole | 17:30 |
| freem | I exagerate a bit, ofc | 17:31 |
| freem | I still would pick TOFU over chain of trust any day | 17:31 |
| cakebandit | TOFU? | 17:32 |
| freem | because I have strictly no idea and no trust in who might eventually sign the upper certificate | 17:32 |
| freem | trust on first use | 17:32 |
| freem | basically, you query a resource. There is a cert, you accept it, or not. It is registered and remembered. If the same resource get a cert change, you are asked if that is normal. | 17:33 |
| CueXXIII | you could change the sources.list to https, that won't change the additional integrity checking of the archive anyways, it would just be another transport layer | 17:33 |
| freem | unlike the chain of trust, you actually *see* that something happenened, so have more control | 17:33 |
| cakebandit | Im no expert, thanks for the answer. cheers | 17:33 |
| freem | I'm no expert neither | 17:34 |
| CueXXIII | but it would prevent a mitm to know which packages you have installed (by knowing which packages you upgrade) | 17:34 |
| freem | technically, https does not prevent tofu, though, only main browsers actively fight that idea | 17:34 |
| freem | let's encrypt is just a workaround their mindset | 17:35 |
| Hurgotron | freem: "certificate patrol" for Firefox | 17:35 |
| cakebandit | lol | 17:35 |
| freem | :) | 17:35 |
| freem | the only thing I know about this is that if I had to have a real trust in interweb, I would patch my browser and OS to get rid of the certificates, so that I would validate them myself | 17:36 |
| freem | and it would still be weak because the chain *can* be corrupted. It happened in the past, in will happen anew. | 17:36 |
| freem | I believe more in showing things to users, even if I know that most users don't care | 17:37 |
| CueXXIII | freem: good luck trying to verify the certs yourself. try calling your bank and checking their cert fingerpring | 17:37 |
| freem | yes | 17:38 |
| freem | it is why I do not trust the web | 17:38 |
| freem | I rarely buy on internet because it is impossible to actually understand what happens when you click on a button on a website | 17:38 |
| freem | those days, web devs seems to even have forgotten about the <a href=...> thing | 17:39 |
| freem | they use javascript to do it | 17:39 |
| freem | either they're stupid, or they're malicious and do this to pwn you | 17:39 |
| freem | one way or another | 17:39 |
| freem | but I'm afraid I'm offtopic, my bad :) | 17:40 |
| user71 | It might be about the 'secure connection' symbols in their browsers. | 17:40 |
| cakebandit | passionate :-) | 17:40 |
| user71 | 'unsecure' gets a bad rap | 17:41 |
| freem | I forgot to go to bed last night, that helps a bit with lack of restraint :D | 17:41 |
| freem | user71: yes, and notably, firefox the "hero" of freedom contributed to make it hard to self-sign your own website | 17:42 |
| freem | in the meantime, they do sell your data to google, ofc | 17:42 |
| freem | IIRC even their add-on page would send data to google, ain't it a joke? | 17:43 |
| cakebandit | I'm using librewolf, but that has been controversial lately. There is another browser I may switch to that uses a file to disable various things, I forgot the name. | 17:44 |
| freem | the one browser I have some hope into is otter browser | 18:07 |
| freem | sadly not packaged, and have bad perfs on my debian for some reason I don't know | 18:07 |
| freem | it is one of the offspring of opera12 with vivaldi, but it is open source and supports several rendering engines | 18:08 |
| freem | unlike vivaldi, I feel it focuses more on the privacy features than eye candiness, but still have mouse gestures and other classical opera 12 features | 18:08 |
| freem | (not all, though, foss, indie project, you know) | 18:09 |
| cakebandit | ah, well I guess you can modify a file called user.js and use default firefox, here is one https://github.com/yokoffing/Betterfox ,but its not the one that was being talked about on 4chan | 18:09 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!