Re: Trouble upgrading Debian
On Sun, Sep 05, 2021 at 09:01:22PM -0300, Dedeco Balaco wrote: [...] > I did not receive the Greg answer. The only 3 messages I have in my > folder now are the 2 you sent, and the one I receive for my message: Here's the start of the thread: https://lists.debian.org/debian-user/2021/09/threads.html#00154 Yes, it's quite possible that your mail takes a bit longer to appear if you are not subscribed. But it /will/ appear. This mailing list is open "by design". Cheers - t signature.asc Description: Digital signature
Re: Installing old/deprecated packages
On 9/5/21, Linux-Fan wrote:
> riveravaldez writes:
>
>> I have this `phwmon.py`[1] which I use with fluxbox to have a couple
>> of system monitors at hand. It depends on some python2 packages, so
>> stopped working some time ago.
>
> Any specific reason for preferring `phwmon.py` over a tool like `conky`?
Hi, Linux-Fan, thanks a lot for your answers.
`conky` is great, but you have to see the desktop to see `conky`, and I
tend to work with maximized windows.
Monitors like `phwmon.py` or the ones that come by default with IceWM
for instance are permanently visible in the sys-tray/taskbar (no matter
you're using fluxbox, openbox+tint2, etc.). That's the only reason:
minimal and visible.
>> I've just made it work, installing manually (# apt-get install
>> packages.deb)
>> this packages that I've downloaded from Debian OldStable official
>> archives:
>>
>> python-psutil
>> python-is-python2 (this is in fact in Testing)
>> python-numpy
>> python-pkg-resources
>> python-cairo
>> libffi6
>> python-gobject-2
>> python-gtk2
>>
>> Therefore, my questions:
>>
>> How safe is this?
>
> IMHO it's pretty OK because that is quite similar to having upgraded from an
> old system with the legacy packages installed to a new release where they
> are no longer part of.
Thanks!
>> Is it better to install them as I did, or adding the corresponding line
>> in
>> sources.list and pull them from there? Is there any difference?
>
> There are differences: Whenever you install packages, you may not notice
> that they are only avaliable in old releases because the output of
> `apt-cache search` and similar tools will include old packages. Also,
> running a release with stable+oldstable in sources.list is less common than
> the other case: stable in sources.list and some oldstable packages leftover
> from upgrades. In case bugs are fixed in the oldstable package, you will get
> them automatically if you have them in sources.list.
>
> My personal choice would be to install the packages without adding the
> oldstable repositories as to be reminded that they are obsolete and are
> likely to stop working in the future.
Thanks again. Very informative and educational.
When you say 'as to be reminded that they are obsolete', how/when/where
the system will remind me this?, will it be?
> Be aware that libraries like `python-psutil` may not work with newer
> kernels. Here (on oldstable with a backported kernel 5.10) the script would
>
> not run due to excess fields reported by the kernel for disk statistics:
>
> | $ ./phwmon.py
> | Traceback (most recent call last):
> | File "./phwmon.py", line 341, in
> | HardwareMonitor()
> | File "./phwmon.py", line 128, in __init__
> | self.initDiskIo()
> | File "./phwmon.py", line 274, in initDiskIo
> | v = psutil.disk_io_counters(perdisk=False)
> | File "/usr/lib/python2.7/dist-packages/psutil/__init__.py", line 2131,
> in disk_io_counters
> | rawdict = _psplatform.disk_io_counters(**kwargs)
> | File "/usr/lib/python2.7/dist-packages/psutil/_pslinux.py", line 1121,
> in disk_io_counters
> | for entry in gen:
> | File "/usr/lib/python2.7/dist-packages/psutil/_pslinux.py", line 1094,
> in read_procfs
> | raise ValueError("not sure how to interpret line %r" % line)
> | ValueError: not sure how to interpret line ' 259 0 nvme0n1 42428
> 17299 3905792 8439 49354 7425 3352623 15456 0 48512 26929 43429 11 476835656
> 3033 0 0\n'
Yes, indeed. I didn't mentioned it but I had to "fix" that as seen in:
https://gitlab.com/o9000/phwmon/-/issues/3#note_374558691
Essentially, convert:
`elif flen == 14 or flen == 18:`
to
`elif flen == 14 or flen == 18 or flen == 20:`
In /usr/lib/python2.7/dist-packages/psutil/_pslinux.py
Supposedly shouldn't be problematic, but I'm not sure. Any
comment on this?
> See also: https://forums.bunsenlabs.org/viewtopic.php?id=967
Thanks. Checked.
> [...]
>
>> [1] https://gitlab.com/o9000/phwmon
>
> Btw. it looks as if `python-is-python2` is not needed for this to run?
> `phwmon.py` states `python2` explicitly.
Probably, not sure. But if I'm remembering properly it was pulled as
a dependency's dependency.
> HTH
> Linux-Fan
>
> öö
>
> PS: If you are interested in my thoughts on status bars, see here:
> https://masysma.lima-city.de/32/i3bar.xhtml
Thanks a lot, LF!
I'm checking it right now. Very interesting.
Best regards.
'sudo apt-get update' stall/error
$ sudo apt-get update ... ? Nothing happens, when I do C-c C-c it says on stderr: Got SIGINT, quitting. Traceback (most recent call last): File "/usr/share/squid-deb-proxy-client/apt-avahi-discover", line 126, in address = get_proxy_host_port_from_avahi() File "/usr/share/squid-deb-proxy-client/apt-avahi-discover", line 79, in get_proxy_host_port_from_avahi for line in p.stdout: KeyboardInterrupt Ideas? -- underground experts united https://dataswamp.org/~incal
Re: 'sudo apt-get update' stall/error
Hi. On Mon, Sep 06, 2021 at 01:43:03AM +0200, Emanuel Berg wrote: > $ sudo apt-get update ... > address = get_proxy_host_port_from_avahi() > File "/usr/share/squid-deb-proxy-client/apt-avahi-discover", line 79, in > get_proxy_host_port_from_avahi ... > > Ideas? apt purge squid-deb-proxy-client, for starters. If you're using Squid as a proxy - just set the proxy in /etc/apt/apt.conf, like this: Acquire::http::Proxy "http://:3148"; If not - no further action is required. Reco
Re: Trouble upgrading Debian
Em 05-09-2021 19:10, piorunz escreveu: > On 05/09/2021 05:27, Dedeco Balaco wrote: > >> What should i do to solve this? > You reposted your question two hours later, so I repost the solution and > CC you as well. I am sorry it was reposted. I had trouble subscribing to the list, using the site, and I thought the first message would be rejected. After several tries without success, I decided to subscribe using the mail address command. I instantly received a confirmation of subscription, and this showed me I was not subscribed before. I also did not see my first message in the list (yes, I waited 20 minutes for the site refreshing time). So, I resent my question. > 1. Start by reading the error message. If you can't understand it, then > paste it here. As Greg Wooledge already said. I did not receive the Greg answer. The only 3 messages I have in my folder now are the 2 you sent, and the one I receive for my message: I will search the Greg message in the archives, since you mention it. > 2. Disable colouring e-mails in your Thunderbird. Or disable HTML > e-mails all together, use plain text e-mail. No. I do not want to edit emails with the traditional "black over white" area. And I have seen that the colors I send can be ignored. And Thunderbird also sends a pure text message together with the HTML messages I prefer to use. Thank you
Re: Trouble upgrading Debian
On 05/09/2021 05:27, Dedeco Balaco wrote: What should i do to solve this? You reposted your question two hours later, so I repost the solution and CC you as well. 1. Start by reading the error message. If you can't understand it, then paste it here. As Greg Wooledge already said. 2. Disable colouring e-mails in your Thunderbird. Or disable HTML e-mails all together, use plain text e-mail. -- With kindest regards, piorunz. ⢀⣴⠾⠻⢶⣦⠀ ⣾⠁⢠⠒⠀⣿⡁ Debian - The universal operating system ⢿⡄⠘⠷⠚⠋⠀ https://www.debian.org ⠈⠳⣄
Re: APT testing and unstabe Firefox: can't find newest version from unstable
On 9/5/21, Andrew M.A. Cater wrote: > > This is the problem with web browsers getting bigger, more complex > dependencies, more infrastructure complexities - and it has always > been so. Web browsers are also the go-to applications for stress > testing any machine once again. You nailed that! Mine keeps bogging down with 8GB ram and 2.7GHz dual core. It's running with a smaller Firefox session's worth of tabs than what ran for hours on 2GB ram and 1.66GHz dual core. I swear it feels like their browsers are trained to sniff around to see what power we've got then adjust their usage of our resources accordingly. As an afterthought, maybe it's the websites themselves doing the sniffing for available resources, too. Might not be like that, but it's how it feels based on how I can't seem to get ahead of that game here. I'm just so over it with respect to having to log out then log back in to clear out the cobwebs when it starts grinding to a halt. As a secondary afterthought turned heads up: In cleaning out my setup regularly, I one day noticed a BUNCH of cookies at the top of the last time used list when they should not have been. The relevance to browser resource usage is that I hadn't been on the affected websites' tabs in months. A lot of cookies are respectfully sitting silent and unused, but there are a few that are not. That's going to take an escalating toll on available computer resources, too. One obvious quick fix would be to manually block those cookies if they're not important and as they become apparent. Oh, and don't get me started griping about those websites that plant 70, 80, 100+ cookies per single or maybe two or three page turns on their sites. I've seen that happen in the past while deleting a site's entire cookie lineup because their site's not working properly. Cindy :) -- Cindy-Sue Causey Talking Rock, Pickens County, Georgia, USA * runs with birdseed *
Re: APT testing and unstabe Firefox: can't find newest version from unstable
On Sun 05 Sep 2021 at 19:31:32 +, Andrew M.A. Cater wrote: > On Sun, Sep 05, 2021 at 08:56:36PM +0200, Oliver Schoede wrote: > > On Fri, 03 Sep 2021 20:50:06 +0200 > > Sven Joachim wrote: > > > > > >Version 91 is only in experimental. > > > > > > > Probably blocked by some Rust stuff again. Anyone who's waiting and if > > possible please get a flatpak and get on with your life. Debian is > > providing that for a reason, too. We've been at the same point about a > > year ago when on some mailing list it was suggested Debian should just > > provide a flatpak. A joke of course, well I think it was. Still I > > decided to actually give it a try and have been happily using two of > > these since then, Firefox, and Chromium, which itself is too often > > vulnerable in Sid. Perhaps in the future distributions should really > > consider making do with, say, Firefox ESR and direct users who need > > "more" to something anyone can sort of agree on and flock together, > > that might well be avenues like Flatpak or AppImage. Container > > solutions are certainly not the be-all and end-all but I don't see much > > of a drawback for a case like this. You'll spend about a GiB extra, > > it's basically pulling its own small userland, once. Command line use > > needs some getting used to, kind of like systemd, hardly surprising if > > you know where it's from. But easy enough, same with desktop > > integration. There's no sane reason for using an outdated web browser > > today. If you want or need to stay purist, there is always ESR. > > > > Oliver > > > > This is the problem with web browsers getting bigger, more complex > dependencies, more infrastructure complexities - and it has always > been so. Web browsers are also the go-to applications for stress > testing any machine once again. There aren't any extensive problem with Firefox on buster and bullseye. Mine works, reliably. Maybe there are reports that say otherwise? > Flatpaks and appimages are fine if they can be built - there's every > chance that they, too witll be hit by this sort of thing at some > point. Why bother with what Debian does not provide by default. -- Brian.
Re: APT testing and unstabe Firefox: can't find newest version from unstable
On Sun, Sep 05, 2021 at 08:56:36PM +0200, Oliver Schoede wrote: > On Fri, 03 Sep 2021 20:50:06 +0200 > Sven Joachim wrote: > > > >Version 91 is only in experimental. > > > > Probably blocked by some Rust stuff again. Anyone who's waiting and if > possible please get a flatpak and get on with your life. Debian is > providing that for a reason, too. We've been at the same point about a > year ago when on some mailing list it was suggested Debian should just > provide a flatpak. A joke of course, well I think it was. Still I > decided to actually give it a try and have been happily using two of > these since then, Firefox, and Chromium, which itself is too often > vulnerable in Sid. Perhaps in the future distributions should really > consider making do with, say, Firefox ESR and direct users who need > "more" to something anyone can sort of agree on and flock together, > that might well be avenues like Flatpak or AppImage. Container > solutions are certainly not the be-all and end-all but I don't see much > of a drawback for a case like this. You'll spend about a GiB extra, > it's basically pulling its own small userland, once. Command line use > needs some getting used to, kind of like systemd, hardly surprising if > you know where it's from. But easy enough, same with desktop > integration. There's no sane reason for using an outdated web browser > today. If you want or need to stay purist, there is always ESR. > > Oliver > This is the problem with web browsers getting bigger, more complex dependencies, more infrastructure complexities - and it has always been so. Web browsers are also the go-to applications for stress testing any machine once again. Flatpaks and appimages are fine if they can be built - there's every chance that they, too witll be hit by this sort of thing at some point. Firefox ESR is actually releatively reasonable in terms of how fast it moves - it still isn't easy for anyone to build. [And upstream show no particular interest in Firefox for other architectures - so have fun if you're running arm]. At some point, bookworm will settle a little more and it will be feasible to start providing lots more in bullseye-backports. In the interim
Re: EFI boot installer
On Sun, Sep 05, 2021 at 09:21:38PM +0300, Gunnar Gervin wrote: > Hi all. > Where do I find a live Debian x86-64 iso image with EFI 'cd image' in it ? > (My 2007 ex-Mac computer now only runs Puppy x86-64 live dvd in RAM.) > My Macbook 64b computer won't open any iso without a cd image named EFI. > Alt key on boot up will show distro's ISO image, on the right side I need > an EFI cd. > Read that Debian supports EFI. So I'm here, asking for your help &/or a > link to ISO. > BR, > Gunnar Gervin. > PS. > Distro developers, if any here *:* > It's all right you dislike EFI bootloader after MS released it, but please > let me use it. > DS. We've been using EFI since 2008? As suggested before to you: https://cdimage.debian.org/debian-cd/current/amd64/iso-cd/debian-mac-11.0.0-amd64-netinst.iso is the _first_ one I'd try on your old "ex-Mac" (since that's around the time that some Macs of that vintage had difficulty reading certain CD formats). If that works - great. If not, I'd try either of the following on either machine: https://cdimage.debian.org/debian-cd/current/amd64/iso-dvd/debian-11.0.0-amd64-DVD-1.iso is what I'd try on the newer one (or possibly the non-free unofficial version which includes firmware at https://cdimage.debian.org/cdimage/unofficial/non-free/cd-including-firmware/11.0.0+nonfree/amd64/iso-dvd/firmware-11.0.0-amd64-DVD-1.iso ) You might also want to try the one of the Debian-live images at https://cdimage.debian.org/cdimage/unofficial/non-free/cd-including-firmware/11.0.0-live+nonfree/amd64/iso-hybrid/ but I would _NOT_ recommend using Calamares to install. Hope this helps, as ever, Andy Cater
Re: EFI boot installer
On 2021-09-05 at 14:21, Gunnar Gervin wrote: > Hi all. Where do I find a live Debian x86-64 iso image with EFI 'cd > image' in it ? As far as I understand, all of the Debian install ISOs will support both EFI and non-EFI-style boot. (At my workplace, we have Linux-based utility CDs which do support booting in both modes depending on which version of the optical drive is selected at the F12 "select a boot device" menu. In that instance it's easy to confirm that they really are booting to different configurations, because the visual experience is significantly different, but that doesn't always have to be the case.) Whether the resulting install is itself EFI-boot or not depends on other factors; I knew about them relatively recently, when I built my current computer and installed it for EFI boot, but just at the moment I don't remember the specifics. If memory serves, all I really did on that front was to boot into the installer via EFI boot mode, create and designate appropriate partitions to fit the requirements for EFI boot, and possibly select appropriate options during the install process; the installer did the rest. I suspect that the first step was the one which told the installer which boot mode to set up the new install for. > (My 2007 ex-Mac computer now only runs Puppy x86-64 live dvd in > RAM.) My Macbook 64b computer won't open any iso without a cd image > named EFI. Alt key on boot up will show distro's ISO image, on the > right side I need an EFI cd. Read that Debian supports EFI. So I'm > here, asking for your help &/or a link to ISO. I used the amd64 netinst ISO for Debian 11, via writing it to a USB drive; IIRC I performed the install during late June / early July, so assuming I downloaded right before using it, that wouldn't be far different from what became the release version. That's available at https://cdimage.debian.org/debian-cd/current/amd64/iso-cd/debian-11.0.0-amd64-netinst.iso which is linked to from https://www.debian.org/CD/netinst/ . It's not named as EFI, but naming is irrelevant to the contents. I can confirm that the bullseye amd64 netinst ISO which I have locally right now (which seems to be from April, not June) includes the EFI subdirectory which is probably what should make EFI boot possible. If your machine will actually refuse to use a given ISO based on its *name*, then you need to either figure out how to rename the ISO (which may or may not be as simple as changing the filename), or use a different machine; that's not a reasonable design, and there's probably not much we can do about it. -- The Wanderer The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man. -- George Bernard Shaw signature.asc Description: OpenPGP digital signature
Re: APT testing and unstabe Firefox: can't find newest version from unstable
On Fri, 03 Sep 2021 20:50:06 +0200 Sven Joachim wrote: > >Version 91 is only in experimental. > Probably blocked by some Rust stuff again. Anyone who's waiting and if possible please get a flatpak and get on with your life. Debian is providing that for a reason, too. We've been at the same point about a year ago when on some mailing list it was suggested Debian should just provide a flatpak. A joke of course, well I think it was. Still I decided to actually give it a try and have been happily using two of these since then, Firefox, and Chromium, which itself is too often vulnerable in Sid. Perhaps in the future distributions should really consider making do with, say, Firefox ESR and direct users who need "more" to something anyone can sort of agree on and flock together, that might well be avenues like Flatpak or AppImage. Container solutions are certainly not the be-all and end-all but I don't see much of a drawback for a case like this. You'll spend about a GiB extra, it's basically pulling its own small userland, once. Command line use needs some getting used to, kind of like systemd, hardly surprising if you know where it's from. But easy enough, same with desktop integration. There's no sane reason for using an outdated web browser today. If you want or need to stay purist, there is always ESR. Oliver
EFI boot installer
Hi all. Where do I find a live Debian x86-64 iso image with EFI 'cd image' in it ? (My 2007 ex-Mac computer now only runs Puppy x86-64 live dvd in RAM.) My Macbook 64b computer won't open any iso without a cd image named EFI. Alt key on boot up will show distro's ISO image, on the right side I need an EFI cd. Read that Debian supports EFI. So I'm here, asking for your help &/or a link to ISO. BR, Gunnar Gervin. PS. Distro developers, if any here *:* It's all right you dislike EFI bootloader after MS released it, but please let me use it. DS.
Re: Why ``color_prompt`` is only set for ``xterm``?
I may as well add to the fun here.
Some years back, probably the late '90s or so, there was a short lived
"Bash Prompt Project". I seem to recall it have a package in Debian
back around that time. I had previously done colorizing of the MS-DOS
prompt and had desired to do the same in Bash. Along the way I found
that project and I think made a contribution. It's dead now but my
colored prompts live on!
A couple of years ago I became fascinated with terminfo and so am using
it rather than specifying ANSI escape codes directly. Here is the
entirety of my ~/.bashrc that deals with setting the prompt:
-8<---
# Git repository status in the prompt
if [[ -e /usr/lib/git-core/git-sh-prompt ]]; then
# From https://gist.github.com/eliotsykes/47516b877f5a4f7cd52f
source /usr/lib/git-core/git-sh-prompt
export GIT_PS1_SHOWDIRTYSTATE=true # staged '+', unstaged '*'
export GIT_PS1_SHOWSTASHSTATE=true # '$' something is stashed
export GIT_PS1_SHOWUPSTREAM="auto" # '<' behind, '>' ahead, '<>'
diverged, '=' no difference
export GIT_PS1_SHOWUNTRACKEDFILES=true # '%' untracked files
fi
# set variable identifying the chroot you work in (used in the prompt below)
if [ -z "${debian_chroot:-}" ] && [ -r /etc/debian_chroot ]; then
debian_chroot=$(cat /etc/debian_chroot)
fi
# uncomment for a colored prompt, if the terminal has the capability; turned
# off by default to not distract the user: the focus in a terminal window
# should be on the output of commands, not on the prompt
force_color_prompt=yes
if [ -n "$force_color_prompt" ]; then
if [ -x /usr/bin/tput ] && tput setaf 1 >&/dev/null; then
# We have color support; assume it's compliant with Ecma-48
# (ISO/IEC-6429). (Lack of such support is extremely rare, and such
# a case would tend to support setf rather than setaf.)
color_prompt=yes
else
color_prompt=
fi
fi
# Use variables set from terminfo capabilities to make PS1 and LESS_TERMCAP_*
# variables less cryptic.
#
# Colors: 0, Black; 1, Red; 2, Green; 3, Yellow; 4, Blue; 5, Magenta; 6, Cyan;
7, White.
# Foreground (text) colors
BLK=$(tput setaf 0)
RED=$(tput setaf 1)
GRN=$(tput setaf 2)
YEL=$(tput setaf 3)
BLU=$(tput setaf 4)
MAG=$(tput setaf 5)
CYA=$(tput setaf 6)
WHT=$(tput setaf 7)
# Background colors
BLKB=$(tput setab 0)
REDB=$(tput setab 1)
GRNB=$(tput setab 2)
YELB=$(tput setab 3)
BLUB=$(tput setab 4)
MAGB=$(tput setab 5)
CYAB=$(tput setab 6)
WHTB=$(tput setab 7)
# Character attibutes
BLD=$(tput bold)
ITA=$(tput sitm)
NOR=$(tput sgr 0)
# More prompt madness. Show exit status if command returned non-zero result.
prompt_last_exit_status () {
PROMPT_LAST_EXIT_STATUS=$?
if [[ ${PROMPT_LAST_EXIT_STATUS} == "0" ]]; then
PROMPT_LAST_EXIT_STATUS=
else
PROMPT_LAST_EXIT_STATUS='{'$BLD$RED$PROMPT_LAST_EXIT_STATUS$NOR'}'
fi
}
prompt_command () {
prompt_last_exit_status
# From: https://superuser.com/a/180275
CurDir=$(pwd|sed -e "s!$HOME!~!"|sed -re "s!([^/])[^/]+/!\1/!g")
}
export PROMPT_COMMAND=prompt_command
if [ "$color_prompt" = yes ]; then
PS1='('
PS1+='\[$CYA\]$SHLVL\[$NOR\]'
PS1+=')'
PS1+='['
PS1+='\[$YEL\]\j\[$NOR\]'
PS1+=']'
PS1+='\[$GRN\]\u\[$NOR\]'
PS1+='@'
PS1+='\[$BLD$GRN\]\h\[$NOR\]'
PS1+=':'
PS1+='\[$BLD$BLU\]$CurDir\[$NOR\]'
PS1+=' '
PS1+='$(__git_ps1 "(\[$MAG\]\[$ITA\]%s\[$NOR\])")'
PS1+=' '
PS1+='$PROMPT_LAST_EXIT_STATUS'
PS1+='\n'
PS1+='\[$BLD$WHT\]\$\[$NOR\] '
else
PS1='${debian_chroot:+($debian_chroot)}\u@\h:\w\$ '
fi
unset color_prompt force_color_prompt
->8---
Generally, for root accounts I change GRN to RED and on virtual machines
to MAG. Of course quite a range is available but the most limiting is
the Linux virtual console driver so I keep things relatively simple.
I also created some posts on my blog about tput[1] and how to reset the
color palettes of the various terminals[2].
- Nate
[1] https://www.n0nb.us/blog/2020/02/tinkering-in-the-terminal-with-tput/
[2]
https://www.n0nb.us/blog/2020/02/coping-with-color-on-the-linux-console-and-xterm-and-friends/
--
"The optimist proclaims that we live in the best of all
possible worlds. The pessimist fears this is true."
Web: https://www.n0nb.us
Projects: https://github.com/N0NB
GPG fingerprint: 82D6 4F6B 0E67 CD41 F689 BBA6 FB2C 5130 D55A 8819
signature.asc
Description: PGP signature
Re: Internet diagnosing
On Sun, Sep 05, 2021 at 04:57:44PM +0200, Julius Hamilton wrote: > Hey, > > I was wondering if anyone knew a similar user email group like this one for > questions related to internet connectivity. Maybe there is a widely used > open source network software tool where people discuss network questions. > > My question is about there being some relationship between certain WiFi > networks I use and my machine's inability to access the internet in some > ways. For example, on one WiFi network, Google works fine in a browser, but > Stack Overflow returns "This site can’t be reached; https://superuser.com/ > is unreachable. > ERR_ADDRESS_UNREACHABLE". On another, I can use my Android browser > normally, but nothing through Andronix (Android Linux). > > So I am curious where in the process the issue is, probably the WiFi > network's settings? But, what kinds of settings are there? What would it be > checking for to block Andronix, for example, or Stack Overflow? > > These are private networks so no one is blocking specific categories of > sites, for example. The reason seems to be something more technical and > accidental. > > Thank you, > Julius Hi Julius, There are all sorts of reasons for your IP not to be trusted / not to be reachable. There's no guarantee that someone hasn't put in a policy somewhere saying "we don't trust cafe networks provided by $ISP / we don't accept email from $PROVIDER wherever it appears to come from" and they're on a deny list somewhere. Bad DNS / bad trust relationships for SSL certs can crop up anywhere. Within Debian, see this quite often, for example, for some would be contributors to wiki.debian.org who can't get in to register - because their network has previously been blocked at some point for spam/DDOS or whatever. The people concerned may be entirely innocent but can only be unblocked on a case by case basis. This would be off-topic on this list - which normally deals only with issues arising from Debian - but it may be that someone else has general, "best endeavours" advice which may help. With every good wish, as ever, Andy Cater
Re: masked service file
On Sun, Sep 05, 2021 at 01:56:30PM +, Andrew M.A. Cater wrote: [...] > * play the ball not the man - don't engage in disguised personal attacks Yes, pretty please. Thank you, Andrew. Cheers - t signature.asc Description: Digital signature
Internet diagnosing
Hey, I was wondering if anyone knew a similar user email group like this one for questions related to internet connectivity. Maybe there is a widely used open source network software tool where people discuss network questions. My question is about there being some relationship between certain WiFi networks I use and my machine's inability to access the internet in some ways. For example, on one WiFi network, Google works fine in a browser, but Stack Overflow returns "This site can’t be reached; https://superuser.com/ is unreachable. ERR_ADDRESS_UNREACHABLE". On another, I can use my Android browser normally, but nothing through Andronix (Android Linux). So I am curious where in the process the issue is, probably the WiFi network's settings? But, what kinds of settings are there? What would it be checking for to block Andronix, for example, or Stack Overflow? These are private networks so no one is blocking specific categories of sites, for example. The reason seems to be something more technical and accidental. Thank you, Julius
Re: masked service file
On Wed, Sep 01, 2021 at 03:09:59PM -0400, Greg Wooledge wrote: > On Wed, Sep 01, 2021 at 07:54:02PM +0100, Brian wrote: > > As a non-professional sysadmin, this "Brian person" is wondering why > > the OP's question (clearly relating to saned) is sidelined in favour > > of other concerns. > > I'm not sure *what* the OP's question really is. They're on an X-Y > tangent about masking, which is extremely irritating to me, since they > aren't giving enough information even to be able to answer the tangential > question, let alone to solve the original problem. > > We can't do *anything* with the lack of information we're being given. > > > Who the hell knows? Maybe if we had some *details* we might be able to > formulate a guess. > > > > Can I find out why it's masked ? > > But sure, OP, keep on not providing information. It's so useful. It's > getting you such high quality answers. > People in general == all the contributors to this thread. I can sense the exasperation: I can feel the tension. Believe me, I understand where you are all coming from BUT You've prompted my (almost weekly) reminder to * adhere to the code of conduct. * play the ball not the man - don't engage in disguised personal attacks * step aside from emotion and ask technical questions dispassionately and clearly * make this mailing list a worthwhile place to read There is a space for frustration and annoyance that somehow, despite your very best efforts, the person typing at the other end doesn't "get it" no matter how clear you are - but it's a general annoyance that you can't get across what you know to help them, rather than hitting out at someone. There's a minimal space for banter and back and forth comments and being able to rely on regular commenters to answer - and an inward feeling of "Oh, no, not THAT topic again from THAT person" when some things come out again for the Nth time - but no space for belittling people for what they do/do not know or contribute. There's a space for knowing when to stop, when to not hit reply, when to step away from the keyboard or give it up as a lost cause. That step might be to be silent - or it might be a chance to put in a reflective comment on "The following topics come up regularly" and a collected summation of several discussions. There is space for a better "How to ask meaningful questions" topic either here or in wiki.debian.org. There's certainly space for a better "These come up regularly FAQ" but there isn't really space from experienced folk for some of the behaviour that happens here. For the avoidance of doubt: I can be (and have been) guilty of a lot of the less good behaviour outlined above on various lists and forums over the years - this is not a holier than thou post - but it's sometimes good to look back dispassionately and see what impression some threads leave. I'm also more than willing to try and draft better FAQs. With every good wish, as ever, Andy Cater [As a member of the Community Team]
Re: Why ``color_prompt`` is only set for ``xterm``? (colored prompt examples)
On Sun, Sep 05, 2021 at 09:38:34AM -0400, Cindy Sue Causey wrote:
> There's also that thing about how terminals will interpret the
> different types of quotes (dumb/typewriter/ASCII versus
> typographic/curly/smart) very literally. I experienced THAT fail
> firsthand and now try to remember to plug anything I copy into a plain
> text editor before then recopying over to a terminal.
Quotes matter to the shell, not to the terminal. And the only kinds
of quotes that the shell cares about are ASCII single and double
quotes (' and " respectively).
Unicode curly quotes are not considered quotes by any shell. They're
just random un-special characters, like q or 5 or ñ.
> Am wondering if, am more like hoping that this quotation marks part of
> it would have stood out when I research how to further customize my
> own setup. This will be a priceless personalization if I can
> eventually coerce it to say "(debootstrap) chroot" in place of some of
> the characters there. Shh, don't tell me how. Lead a fish to water,
> yada-yada. :)
Look for "chroot" in your ~/.bashrc and see how it's currently being
done. Make sure you understand each command, and what it does.
Re: Why ``color_prompt`` is only set for ``xterm``? (colored prompt examples)
On 9/5/21, Greg Wooledge wrote:
> On Sun, Sep 05, 2021 at 12:28:37AM -0500, Intense Red wrote:
>> > In /root/.bashrc I use this to give a red prompt including host and
>> > full path followed by a new line.
>>
>>I take this idea a bit further, setting a longer prompt and setting
>> workstation hosts for specific colors for user logins, and then doing a
>> red
>> prompt for servers.
>>
>>Part of my ~/.bashrc:
>>
>> # Set the hostname to a specific color
>> HostName=`hostname -s`
>> if [ $HostName = "capncrunch" ]; then
>>HostColor="\[\033[1;36m\]" # Bright Cyan
>> elif [ $HostName = "piglet" ]; then
>>HostColor="\[\033[1;35m\]" # Bright Purple
>> elif [ $HostName = "wiseguy" ]; then
>
> Not a big fan of case statements?
>
> HostName=${HOSTNAME%%.*}
> case $HostName in
> capncrunch) HostColor="\[\033[1;36m\]";; # Bright Cyan
> piglet) HostColor="\[\033[1;35m\]";; # Bright Purple
> ...
> esac
>
> Also, for the record, your quotes are in the wrong place in your [
> commands. You need quotes around "$HostName" to avoid globbing and
> word splitting. You don't need quotes around simple strings like
> capncrunch and piglet, unless one of them contains whitespace or
> punctuation that's significant to the shell.
>
> The missing quotes around "$HostName" have never mattered because so
> far all of your hostnames have been safe. Maybe that's even a
> guarantee -- I'm not sure what characters are actually allowed in a
> Linux hostname. But quoting correctly still a good habit to get into.
Thank you for the tweak, Greg. That's a biggie because along the way,
we've seen how things change and fail for that very reason. One of my
partitions' terminal root now calls itself by a package name. NO CLUE
how it happened because I wasn't root while the associated package was
being extracted. Confirming placement of quotation marks, if used,
would be one checkpoint for a massive fail like that.
There's also that thing about how terminals will interpret the
different types of quotes (dumb/typewriter/ASCII versus
typographic/curly/smart) very literally. I experienced THAT fail
firsthand and now try to remember to plug anything I copy into a plain
text editor before then recopying over to a terminal.
Am wondering if, am more like hoping that this quotation marks part of
it would have stood out when I research how to further customize my
own setup. This will be a priceless personalization if I can
eventually coerce it to say "(debootstrap) chroot" in place of some of
the characters there. Shh, don't tell me how. Lead a fish to water,
yada-yada. :)
Cindy :)
--
Cindy-Sue Causey
Talking Rock, Pickens County, Georgia, USA
* runs with birdseed *
Re: Lock screen with Gnome 3 login greeter
On Sun, Sep 05, 2021 at 12:50:43PM +0200, Richard Forst wrote: > > I switched to use i3. So now after the screen goes blank, it won't display > > greeter asking to login again. I read somewhere else on the internet that > > gdm3 (I use Debian 11 with kernel 5.10.0-8-amd64) no longer uses > > screensaver to lock the session. Instead it uses dbus to activate login > > asking user to enter id, password for login. I'm not sure if this is helpful, because our setups are clearly different, but: What I use is startx + fvwm, and I have installed the i3lock package as a replacement for xlockmore, which was removed a very long time ago. When I want to lock my screen, I run "i3lock" as a command (either in a terminal, or activated from a WM menu). This gives a plain white screen. To unlock it, I simply type my password. There's a green circle in the middle of the screen that appears while typing, to let me know that my keystrokes are actually being received. In my setup, there is no automatic locking at all. It's always done manually, by running i3lock. I don't know what your setup is (you mention gdm3, but you didn't actually make it clear that you're *running* gdm3 -- it might simply have been something that you found during your Internet searches). You mentioned i3, so I thought it might be possible that you're also using i3lock, perhaps in some sort of automatic mode. I don't know how the rest of i3 works, outside of the i3lock program. When your screen locks, what do you see? Is it a plain white screen? A plain black screen? A still image? An animation? Which display manager are you using, if any? If you use "startx", then the answer is "none", and your setup is similar to mine. Otherwise, you might be using lightdm, or gdm3, or xdm, or any of several other display mangers, most of which have the substring "dm" somewhere in their name. I wonder if switching from gdm3 to lightdm would be enough to solve your problem, if you are in fact using gdm3. I don't know a lot about display managers, since I don't use them myself. Maybe worth looking into...?
Re: Why ``color_prompt`` is only set for ``xterm``?
On Sat, Sep 04, 2021 at 11:23:48PM -0500, David Wright wrote: > $ cat /var/local/lib/myhosts/colours/axis > 5 magenta bbarbutton=white,magenta,none:bbarhotkey=magenta,white,none > $ > > 5 is for ANSI colours, > magenta is for, eg emacs, mutt, etc, > bbarbutton is for mc. > > The lines from /root/.bashrc are: > > if [[ $- = *i* ]]; then > > export PROMPT_COMMAND='Myprompt="$? " && [ "$Myprompt" = "0 " ] && > Myprompt=""' > > # read value from file, and sanitise it > Hue=$(head -c 1 /var/local/lib/myhosts/colours/"$HOSTNAME") > if grep -q ·"$Hue"· <<<"·0·1·2·3·4·5·6·7·8·9·"; then It's strange that you used shell builtins for the $- check but not for this check. read -r Hue _ < /var/local/lib/myhosts/colours/"$HOSTNAME" if [[ $Hue = [0-9] ]]; then Or if you happen to use the other fields of that file later in the script, then change the read command to store all 3 fields in shell variables.
Re: Why ``color_prompt`` is only set for ``xterm``? (colored prompt examples)
On Sun, Sep 05, 2021 at 12:28:37AM -0500, Intense Red wrote:
> > In /root/.bashrc I use this to give a red prompt including host and
> > full path followed by a new line.
>
>I take this idea a bit further, setting a longer prompt and setting
> workstation hosts for specific colors for user logins, and then doing a red
> prompt for servers.
>
>Part of my ~/.bashrc:
>
> # Set the hostname to a specific color
> HostName=`hostname -s`
> if [ $HostName = "capncrunch" ]; then
>HostColor="\[\033[1;36m\]" # Bright Cyan
> elif [ $HostName = "piglet" ]; then
>HostColor="\[\033[1;35m\]" # Bright Purple
> elif [ $HostName = "wiseguy" ]; then
Not a big fan of case statements?
HostName=${HOSTNAME%%.*}
case $HostName in
capncrunch) HostColor="\[\033[1;36m\]";; # Bright Cyan
piglet) HostColor="\[\033[1;35m\]";; # Bright Purple
...
esac
Also, for the record, your quotes are in the wrong place in your [
commands. You need quotes around "$HostName" to avoid globbing and
word splitting. You don't need quotes around simple strings like
capncrunch and piglet, unless one of them contains whitespace or
punctuation that's significant to the shell.
The missing quotes around "$HostName" have never mattered because so
far all of your hostnames have been safe. Maybe that's even a
guarantee -- I'm not sure what characters are actually allowed in a
Linux hostname. But quoting correctly still a good habit to get into.
Re: Installing old/deprecated packages
riveravaldez writes:
I have this `phwmon.py`[1] which I use with fluxbox to have a couple
of system monitors at hand. It depends on some python2 packages, so
stopped working some time ago.
Any specific reason for preferring `phwmon.py` over a tool like `conky`?
I've just made it work, installing manually (# apt-get install packages.deb)
this packages that I've downloaded from Debian OldStable official archives:
python-psutil
python-is-python2 (this is in fact in Testing)
python-numpy
python-pkg-resources
python-cairo
libffi6
python-gobject-2
python-gtk2
Therefore, my questions:
How safe is this?
IMHO it's pretty OK because that is quite similar to having upgraded from an
old system with the legacy packages installed to a new release where they
are no longer part of.
Is it better to install them as I did, or adding the corresponding line in
sources.list and pull them from there? Is there any difference?
There are differences: Whenever you install packages, you may not notice
that they are only avaliable in old releases because the output of
`apt-cache search` and similar tools will include old packages. Also,
running a release with stable+oldstable in sources.list is less common than
the other case: stable in sources.list and some oldstable packages leftover
from upgrades. In case bugs are fixed in the oldstable package, you will get
them automatically if you have them in sources.list.
My personal choice would be to install the packages without adding the
oldstable repositories as to be reminded that they are obsolete and are
likely to stop working in the future.
Be aware that libraries like `python-psutil` may not work with newer
kernels. Here (on oldstable with a backported kernel 5.10) the script would
not run due to excess fields reported by the kernel for disk statistics:
| $ ./phwmon.py
| Traceback (most recent call last):
| File "./phwmon.py", line 341, in
| HardwareMonitor()
| File "./phwmon.py", line 128, in __init__
| self.initDiskIo()
| File "./phwmon.py", line 274, in initDiskIo
| v = psutil.disk_io_counters(perdisk=False)
| File "/usr/lib/python2.7/dist-packages/psutil/__init__.py", line 2131, in
disk_io_counters
| rawdict = _psplatform.disk_io_counters(**kwargs)
| File "/usr/lib/python2.7/dist-packages/psutil/_pslinux.py", line 1121, in
disk_io_counters
| for entry in gen:
| File "/usr/lib/python2.7/dist-packages/psutil/_pslinux.py", line 1094, in
read_procfs
| raise ValueError("not sure how to interpret line %r" % line)
| ValueError: not sure how to interpret line ' 259 0 nvme0n1 42428 17299
3905792 8439 49354 7425 3352623 15456 0 48512 26929 43429 11 476835656 3033 0
0\n'
See also: https://forums.bunsenlabs.org/viewtopic.php?id=967
[...]
[1] https://gitlab.com/o9000/phwmon
Btw. it looks as if `python-is-python2` is not needed for this to run?
`phwmon.py` states `python2` explicitly.
HTH
Linux-Fan
öö
PS: If you are interested in my thoughts on status bars, see here:
https://masysma.lima-city.de/32/i3bar.xhtml
pgpGp_ovIoHsX.pgp
Description: PGP signature
Re: Installation Done after login it Hangs
On Sun, Sep 05, 2021 at 04:52:27PM +0530, Tushar Dasondi wrote: > Dear Debian team, > > After intallation of Debian 11 it hangs at login page after entering the > password.please support > > Regards > Tushar Dasondi > > Tus Can you please give more details. What machine? What desktop environment? if it hangs, does it recover after a minute? With these details, we can do nothing but make guesses: the more information you can give, the better we can help. With every good wish, as ever, Andy Cater
Re: Why ``color_prompt`` is only set for ``xterm``? (colored prompt examples)
On 9/5/21, Intense Red wrote: >> In /root/.bashrc I use this to give a red prompt including host and >> full path followed by a new line. > >I take this idea a bit further, setting a longer prompt and setting > workstation hosts for specific colors for user logins, and then doing a red > > prompt for servers. > >Part of my ~/.bashrc: > > # Set the hostname to a specific color > HostName=`hostname -s` > if [ $HostName = "capncrunch" ]; then >HostColor="\[\033[1;36m\]" # Bright Cyan > < lots of snip for brevity > > > # Then we need to export them so the shell picks things up. > export PS1 LS_COLORS > >That's worth playing with to change some of the default settings if one > wants to play with the colors of a terminal (I use KDE's "konsole" rather > than > xterm, but it's the same idea. That's cool. I did play with it last night as my normal user. I'm going to leave it for now as a reminder to personalize it more. I've seen very similar in some operating system somewhere but never stopped to pursue how it was accomplished. Maybe it was even for a different terminal (emulator) from Debian's repositories. Its first tweak needs to address that it renders the tab "Title" as "Untitled window". I've tried a few times over the years to alter that via "Edit > Preferences > General (tab)" in xfce4-terminal, but it never stuck. It would always reverse back each time a terminal window was fully closed. I could see putting something under (dot)bashrc along with this other as being more geeky fun. PS I just plugged it in under root after going back above and reading from the original snippets. I'm really liking that blinking reminder as to where one is. That might be just what I've always wanted to cognitively distinguish between chroot and root terminal tabs. That's been a worry quite often over time. Thank you! Cindy :) -- Cindy-Sue Causey Talking Rock, Pickens County, Georgia, USA * runs with birdseed *
Installation Done after login it Hangs
Dear Debian team, After intallation of Debian 11 it hangs at login page after entering the password.please support Regards Tushar Dasondi Tus
Re: new to debian
> See here for detailed explanation: > > https://wiki.debian.org/SourcesList I added five lines, thanks Karthik! Sincerely, Byung-Hee
Re: new to debian
On Sun, Sep 5, 2021, 4:15 PM 황병희 wrote: > Hellow~ > > Actually i am new to Debian. Especially i did install Debian 11 Bullseye > udner Chrome OS (ARM64 MT8173 Chromebook). > > Before i used Ubuntu 18.04 LTS. > > My question is: > Is this canonical way? > > #+BEGIN_SRC: sh > soyeomul@penguin:/etc/apt$ cat sources.list > deb https://deb.debian.org/debian bullseye main > soyeomul@penguin:/etc/apt$ > #+END_SRC > > That is only one line. > > Are you asking that one line is enough?. If you're then, yeah that one > line is enough for main repository binary packages > If you need updates(security and point releases) ,sources and non-free > packages you need an extra line > See here for detailed explanation: https://wiki.debian.org/SourcesList
Re: Lock screen with Gnome 3 login greeter
Ok. The link I found previously is [1]. But executing the command provided in that link doesn't work either. dbus-send --type=method_call \ --dest=org.gnome.ScreenSaver \ /org/gnome/ScreenSaver \ org.gnome.ScreenSaver.Lock I appreciate any suggestions or comments. Many thanks. [1]. https://unix.stackexchange.com/questions/86221/how-can-i-lock-my-screen-in-gnome-3-without-gdm/86275#86275 Sep 5, 2021, 17:32 by sterbl...@tutanota.com: > I switched to use i3. So now after the screen goes blank, it won't display > greeter asking to login again. I read somewhere else on the internet that > gdm3 (I use Debian 11 with kernel 5.10.0-8-amd64) no longer uses screensaver > to lock the session. Instead it uses dbus to activate login asking user to > enter id, password for login. However I don't find any related solutions that > work. The closest one is [1] (sorry it's Ubuntu env not Debian). But > executing that command (in [1]), > > > dbus-send --type=method_call \ > --dest=org.gnome.ScreenSaver \ > /org/gnome/ScreenSaver \ > org.gnome.ScreenSaver.Lock > > login screen doesn't come back. So what command should I use so that when the > system becomes idle, the login screen will be active? > > I tried configuring gnome-control-center's Privacy > Screen Lock > Automatic > Screen Lock set to on (round icon moved to right). And Automatic Screen Lock > Delay is set to 30 seconds. However, it's still active (when becoming idle, > screen won't be lock with login display). > > What steps should I perform in order to lock the screen with login being > displayed? I know there is xscreensaver. But this time I want to use login > instead of xscreensaver, which is nice of course. > > Thanks > > [1]. > https://askubuntu.com/a/983451 >
new to debian
Hellow~ Actually i am new to Debian. Especially i did install Debian 11 Bullseye udner Chrome OS (ARM64 MT8173 Chromebook). Before i used Ubuntu 18.04 LTS. My question is: Is this canonical way? #+BEGIN_SRC: sh soyeomul@penguin:/etc/apt$ cat sources.list deb https://deb.debian.org/debian bullseye main soyeomul@penguin:/etc/apt$ #+END_SRC That is only one line. Thanks for any comments! Sincerely, Byung-Hee
Lock screen with Gnome 3 login greeter
I switched to use i3. So now after the screen goes blank, it won't display greeter asking to login again. I read somewhere else on the internet that gdm3 (I use Debian 11 with kernel 5.10.0-8-amd64) no longer uses screensaver to lock the session. Instead it uses dbus to activate login asking user to enter id, password for login. However I don't find any related solutions that work. The closest one is [1] (sorry it's Ubuntu env not Debian). But executing that command (in [1]), dbus-send --type=method_call \ --dest=org.gnome.ScreenSaver \ /org/gnome/ScreenSaver \ org.gnome.ScreenSaver.Lock login screen doesn't come back. So what command should I use so that when the system becomes idle, the login screen will be active? I tried configuring gnome-control-center's Privacy > Screen Lock > Automatic Screen Lock set to on (round icon moved to right). And Automatic Screen Lock Delay is set to 30 seconds. However, it's still active (when becoming idle, screen won't be lock with login display). What steps should I perform in order to lock the screen with login being displayed? I know there is xscreensaver. But this time I want to use login instead of xscreensaver, which is nice of course. Thanks [1]. https://askubuntu.com/a/983451
