| * joerg just muses if #!/bin/sh is now #!/usr/bin/sh | 00:50 | |
| rrq | anyone knows how to stop bash from omitting failed commands from the history? | 05:50 |
|---|---|---|
| DelTomix | rrq: never noticed bash doing that here before ? maybe check .bashrc for a trap on ERR or maybe something buried in the prompt or $PS1 ?? unless maybe something new in $HISTIGNORE ? | 07:00 |
| rrq | probably me confusing myself :) | 07:01 |
| DelTomix | I can relate :) | 07:07 |
| friedhelm | join #openSUSE-buildservice | 10:49 |
| bb|hcb | rrq: bash omits commands that start with space | 11:27 |
| bb|hcb | And there is a setting about that, but I forgot the name | 11:27 |
| rrq | yeah I feel like I've had several cases where it omits a command that fails ... I can see it in the terminal but up-arrow doesn't recall it | 11:29 |
| rrq | and then, after 1-2 other commands it's again available with some up-arrows | 11:30 |
| rrq | but this happens so occasionally that I haven't really focussed on it; it may have been be jumping around to some ather terminal .. often some handful or two terminals in progress... | 11:32 |
| rrq | this time it bugged me enough to drop a line to see if anyone else here had a similar observation | 11:32 |
| joerg | I *think* the commandline content is added to the history buffer *before* it's getting executed, so bash would need to actively *remove* / delete that last record from the history in case of $failure (return code != 0 ?) | 12:36 |
| joerg | rrq: man 1P fc|less -j11 '+/shall be entered at the end of the history list and then re-executed by sh' ;# ;-) | 13:43 |
| DelTomix | Yeah the setting to omit from history is $HISTCONTROL - which can be "ignorespace" (commands with leading spaces), "ignoredups"(successive identical commands), or "ignoreboth" | 13:44 |
| DelTomix | with $HISTIGNORE you can ignore based on pattern matches | 13:50 |
| joerg | wow, TIL | 14:09 |
| joerg | leading whitespace tricked me a few times. it's not in history so you malso can't recall it with (usually) arrow-up to remove the leading space :-) | 14:10 |
| joerg | -m | 14:10 |
| joerg | needs literal c&p from the terminal | 14:11 |
| sachy | leading space is great so you dont have .history filled with everyday garbage - just important commands | 14:20 |
| mason | Randomly, even if, for example, dash installs /usr/bin/sh (or /usr/bin/dash with a /usr/bin/sh symlink) this will still break dpkg -S unless PATHs stop using /bin before /usr/bin, as an example of something packages themselves moving the files won't address. | 16:02 |
| joerg | the webpaqe https://subdivi.de/~helmut/dep17.html Helmut set up seems to very thoroughly cover all implications. | 16:20 |
| joerg | nice analysis of the situation and the plans to proceed | 16:22 |
| mason | He's doing quite a bit of work, yes. | 16:34 |
| mason | Note the first paragraph in "proposal" though. They're not getting buy-in from the dpkg folks because they note a distinct set of issues: https://wiki.debian.org/Teams/Dpkg/MergedUsr | 16:37 |
| joerg | >>This can reduce the overhead of distinct directories and where to put what<< *cough*. This is the justification for messing up FHS? Sounds like a poor deal | 17:01 |
| bb|hcb | It IS a poor deal. Some people pushed that in, leaving for others to clean the mess | 17:09 |
| bb|hcb | Last couple of months helmut is the one who is working hard to clean that mess. | 17:11 |
| bb|hcb | IMHO it would be much easier to revert the decision but politics+hubris are in the way | 17:11 |
| sachy | public? no. RH with its tremendous power to break things. | 17:15 |
| DelTomix | exactly bb|hcb | 17:20 |
| DelTomix | I'm pretty sure he also understands its a poor idea, in great detail judging by his work. | 17:24 |
| joerg | :nod: | 17:31 |
| DelTomix | onefang's earlier comment "onefang wonders why that would be a thing in the first place?" regarding bb|hcb's point "there are races when a package installs /bin/foo and another replaces it with /usr/bin/foo..." made me think maybe we are feeding the wrong end of the donkey, and rrq is absolutely correct about "as long as files are named via their installation path there's no issue." however, | 17:36 |
| DelTomix | there's going to be a transition where its a mess, and so maybe helmut's idea of having a tool to find and identify the issues can help save work wading through that mess | 17:36 |
| DelTomix | ...but most of that is still going to have to be on the Debian side | 17:41 |
| ksx4system | hello :) what happens if you try to replace sysvinit with something that's not (yet) available in Devuan repository? | 18:58 |
| ksx4system | as far as I can see Google revived upstart and it's consequently updated | 18:58 |
| ksx4system | I'm seriously considering porting it to Devuan | 18:58 |
| bgstack15 | ksx4system: You are welcome to experiment and let us know the results! I don't recall how upstart was better/different, but we do advertise init diversity! | 20:10 |
| bgstack15 | I didn't hate CentOS 6 because of its init, let me say. | 20:11 |
| bb|hcb | DelTomix: The transition is already in progress... :( | 20:47 |
| ksx4system | bgstack15, I liked it because it didn't mess with /etc/init.d scripts (same as for sysvinit) and OS booted a little faster on shit hardware | 21:19 |
| ksx4system | (not much of a difference on then decent hw) | 21:20 |
| sachy | anyone here planning to attend LinuxDays 2023 (Prague, 2023-10-08 https://www.linuxdays.cz/2023/) ? At our hackerspace boot Im goint to provide also some Devuan propaganda :) | 21:57 |
| sachy | *going | 21:57 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!