Re: [gentoo-user] problem installing confluent-kafka from guru
Il giorno mar 22 feb 2022 alle ore 08:02 Anatoly Oreshkin < anatoly.oresh...@gmail.com> ha scritto: > Hello, > > I am a newbie to Gentoo. > As end user I am using package confluent-kafka from guru repository. > To install it I have performed the following steps in command line: > [...] > 6. modify > /var/db/repos/guru/dev-python/confluent-kafka/confluent-kafka-1.7.0.ebuild > change PYTHON_COMPAT=( python3_8 ) -> PYTHON_COMPAT=( python3_9 ) > sed -i 's/python3_8/python3_9/g' > /var/db/repos/guru/dev-python/confluent-kafka/confluent-kafka-1.7.0.ebuild > sorry but you can't do this, you have to stick with python3.8 I plan to update the openstack packages next month after I graduate
Re: [gentoo-user] Piping "emerge -a" to some other command
On Fri, 25 Feb 2022 11:26:34 +0100, Dr Rainer Woitok wrote: > Nice idea. However, I'd like to keep things as simple as possible, be- > cause an important usecase for me is installing a new or fixing an ex- > isting system after booting from a USB stick. The "script" binary is in > package "sys-apps/util-linux" which belongs to "@system", so it SHOULD > be on my USB stick. Like "cmp" and "diff" from package "sys-apps/diff- > utils" which belong to "@system" too, but which were NOT on my USB stick > when I used it to install my current system about three years ago, or at > least they were not in "PATH" (I then didn't frantically need these, so > I didn't really search for them). Use a different live distro, one that includes all you need - such as System Rescue. I think Arve's suggestion of using screen (or tmux) is the simplest effective solution. I routinely run emerge -a @world in a tmux session, and scrolling back though long update lists is easy. -- Neil Bothwick I spilled Spot remover on my dog. Now he's gone. pgpfL5KNBc_pQ.pgp Description: OpenPGP digital signature
Re: [gentoo-user] Piping "emerge -a" to some other command
Arve, On Friday, 2022-02-25 10:15:18 +0100, you wrote: > ... > I'm guessing any proposed solution would fail when what you want to > capture has an interactive component like this. I think "fail" is the wrong word here. My own scripts send such quest- ions to the same device they retrieve the answer from ("/dev/tty" in my case). So both, question and answer would be missing from the log file produced by "tee", but they would appear on the screen, which seems more important to me. But you are right in that it would at least be difficult to have them in both places. Sincerely, Rainer
Re: [gentoo-user] Piping "emerge -a" to some other command
John, On Thursday, 2022-02-24 09:59:50 -0500, you wrote: > ... > I use script for that purpose and it works great, the output is > complete and I can just say no and its all there. Nice idea. However, I'd like to keep things as simple as possible, be- cause an important usecase for me is installing a new or fixing an ex- isting system after booting from a USB stick. The "script" binary is in package "sys-apps/util-linux" which belongs to "@system", so it SHOULD be on my USB stick. Like "cmp" and "diff" from package "sys-apps/diff- utils" which belong to "@system" too, but which were NOT on my USB stick when I used it to install my current system about three years ago, or at least they were not in "PATH" (I then didn't frantically need these, so I didn't really search for them). I do not know what "nano" (the only editor then available) will do with all the control characters in the "script" output file, and "scriptre- play" doesn't seem to offer any interactive control. I do have a "gawk" based shell function which turns a "script" output file into a normal text file containing command log + standard output + standard error sans any control characters, but this will not yet be available in that situ- ation. So I'm still preferring "tee" here which definitely WAS on my USB stick. Sincerely, Rainer
Re: [gentoo-user] Piping "emerge -a" to some other command
On Fri, 25 Feb 2022 at 10:00, Dr Rainer Woitok wrote: > A quick search in the "emerge" manual page for "pars" and "pip" did not > turn up anything I considered relevant. Can anyone give more hints? I'm guessing any proposed solution would fail when what you want to capture has an interactive component like this. My solution would rather be to run your emerge commands in a place that would give you the scroll-back that you lack now, such as in a screen session. It feels like something like a screen session would be potentially beneficial in more ways as well, such as being able to do this remotely. Regards, Arve
Re: [gentoo-user] Piping "emerge -a" to some other command
Rich, On Thursday, 2022-02-24 07:10:10 -0500, you wrote: > ... > Wouldn't it be more appropriate to use -p in this case, which will > require no input? Running "emerge -p ..." followed by "emerge ..." without "-p" would run the dependency analysis twice for EVERY package update. Using a single "emerge -a" call will run the dependency analysis only once for almost each package update. The option to say "N" and further inspect the "em- erge" output using an editor is meant as sort of a safety belt which is not intended to be used on a daily basis. > I haven't dug into the man page, but I'm sure there are also options > that are more intended to make the output more suitable for > parsing/etc. A quick search in the "emerge" manual page for "pars" and "pip" did not turn up anything I considered relevant. Can anyone give more hints? Sincerely, Rainer