[arch-general] archlinux.org/pacman/pacman.conf.5.html from Aug-2019
The time stamp at the bottom of https://archlinux.org/pacman/pacman.conf.5.html is Aug-2019. The way I understand pacman v6 conf, there are some changes. Such as retiring TotalDownload. I haven't checked if there are differences with other files at the site. Is updating pacman.conf.5.html already scheduled? -- u34
My guess is that they'll update that once pacman 6 hits stable.
Yash Karandikar
d5c4 1cbc 071b 41b3 f235 da13 f01e 92cd 2390 f369
Jun 1, 2021 6:12:53 PM u34--- via arch-general
The time stamp at the bottom of https://archlinux.org/pacman/pacman.conf.5.html is Aug-2019. The way I understand pacman v6 conf, there are some changes. Such as retiring TotalDownload. I haven't checked if there are differences with other files at the site. Is updating pacman.conf.5.html already scheduled?
-- u34
On Tue, 2021-06-01 at 23:15 +0000, Yash Karandikar via arch-general wrote:
My guess is that they'll update that once pacman 6 hits stable. Yash Karandikar d5c4 1cbc 071b 41b3 f235 da13 f01e 92cd 2390 f369
Jun 1, 2021 6:12:53 PM u34--- via arch-general < arch-general@lists.archlinux.org>:
The time stamp at the bottom of https://archlinux.org/pacman/pacman.conf.5.html is Aug-2019. The way I understand pacman v6 conf, there are some changes. Such as retiring TotalDownload. I haven't checked if there are differences with other files at the site. Is updating pacman.conf.5.html already scheduled?
-- u34
I don't know. But perhaps that page should be deprecated in favor of [1]. [1] https://man.archlinux.org/man/pacman.conf.5 Cheers, Filipe Laíns
On 01/06/2021 23:15, Yash Karandikar via arch-general wrote:
My guess is that they'll update that once pacman 6 hits stable.
pacman=6.0.0-2 is already in stable. ;)
On 6/1/21 6:57 PM, Jonathon Fernyhough via arch-general wrote:
On 01/06/2021 23:15, Yash Karandikar via arch-general wrote:
My guess is that they'll update that once pacman 6 hits stable.
pacman=6.0.0-2 is already in stable. ;)
and it has terminal output issues bigtime. Whatever is causing it the download meter and filename jump back and for by 4-8 characters hundreds of time, sometime injecting a newline. The following is the output from tonight's update: :: Proceed with installation? [Y/n] :: Retrieving packages... ca-certificates-utils-20210529-1-any Total (0/6) 0.0 B 0.00 B/s 00:00 [--------------------------------------] 0% ca-certificates-20210529-1-any-1-any 7.9 KiB 149 KiB/s 00:00 [######################################] 100% Total (1/6) 7.9 KiB 0.00 B/s --:-- [--------------------------------------] 0% js78-78.11.0-1-x86_64529-1-any 2009.0 B 36.3 KiB/s 00:00 [######################################] 100% Total (2/6) 7.9 KiB 0.00 B/s --:-- [--------------------------------------] 0% poppler-21.06.0-1-x86_64 10.5 MiB 5.92 MiB/s 00:02 [######################################] 100% Total (3/6) 9.7 MiB 5.83 MiB/s 00:23 [##------------------------------------] 6% libreoffice-still-7.0.6-2-x86_64 1333.0 KiB 4.88 MiB/s 00:00 [######################################] 100% Total (4/6) 11.7 MiB 5.35 MiB/s 00:25 [###-----------------------------------] 8% poppler-glib-21.06.0-1-x86_64_64 134.0 MiB 5.94 MiB/s 00:23 [######################################] 100% Total (5/6) 144.9 MiB 6.09 MiB/s 00:00 [#####################################-] 99% Total (6/6) 146.1 MiB 5.82 MiB/s 00:25 [######################################] 100% That's just bizarre. Haven't seen this behavior before in the last decade+. -- David C. Rankin, J.D.,P.E.
Just to make sure, you're not using archinstall when that "jumping around" happens right? Because if that's the case, then it's my fault and not pacman :) On Thu, 3 Jun 2021, 10:34 David C. Rankin via arch-general, < arch-general@lists.archlinux.org> wrote:
On 6/1/21 6:57 PM, Jonathon Fernyhough via arch-general wrote:
On 01/06/2021 23:15, Yash Karandikar via arch-general wrote:
My guess is that they'll update that once pacman 6 hits stable.
pacman=6.0.0-2 is already in stable. ;)
and it has terminal output issues bigtime. Whatever is causing it the download meter and filename jump back and for by 4-8 characters hundreds of time, sometime injecting a newline. The following is the output from tonight's update:
:: Proceed with installation? [Y/n] :: Retrieving packages... ca-certificates-utils-20210529-1-any Total (0/6) 0.0 B 0.00 B/s 00:00 [--------------------------------------] 0% ca-certificates-20210529-1-any-1-any 7.9 KiB 149 KiB/s 00:00 [######################################] 100% Total (1/6) 7.9 KiB 0.00 B/s --:-- [--------------------------------------] 0% js78-78.11.0-1-x86_64529-1-any 2009.0 B 36.3 KiB/s 00:00 [######################################] 100% Total (2/6) 7.9 KiB 0.00 B/s --:-- [--------------------------------------] 0% poppler-21.06.0-1-x86_64 10.5 MiB 5.92 MiB/s 00:02 [######################################] 100% Total (3/6) 9.7 MiB 5.83 MiB/s 00:23 [##------------------------------------] 6% libreoffice-still-7.0.6-2-x86_64 1333.0 KiB 4.88 MiB/s 00:00 [######################################] 100% Total (4/6) 11.7 MiB 5.35 MiB/s 00:25 [###-----------------------------------] 8% poppler-glib-21.06.0-1-x86_64_64 134.0 MiB 5.94 MiB/s 00:23 [######################################] 100% Total (5/6) 144.9 MiB 6.09 MiB/s 00:00 [#####################################-] 99% Total (6/6) 146.1 MiB 5.82 MiB/s 00:25 [######################################] 100%
That's just bizarre. Haven't seen this behavior before in the last decade+.
-- David C. Rankin, J.D.,P.E.
On 6/3/21 4:05 AM, Anton Hvornum via arch-general wrote:
Just to make sure, you're not using archinstall when that "jumping around" happens right? Because if that's the case, then it's my fault and not pacman :)
Right, this is plain old 'pacman -Syu' on 3 boxes that have been running since 2013 or so. There was obviously a interface change that was messing things up. It was almost like a stray '\r' was injected or a bad ANSI escape was being hit. -- David C. Rankin, J.D.,P.E.
On 6/1/21 6:57 PM, Jonathon Fernyhough via arch-general wrote:
On 01/06/2021 23:15, Yash Karandikar via arch-general wrote:
My guess is that they'll update that once pacman 6 hits stable.
pacman=6.0.0-2 is already in stable. ;)
Follow up on my earlier report of jittery output on pacman update. I updated 3 machines and 2 out of 3 showed the character jumps and newline, the third appears normal. Mirror issue? -- David C. Rankin, J.D.,P.E.
On 03-06-2021 10:37, David C. Rankin via arch-general wrote:
On 6/1/21 6:57 PM, Jonathon Fernyhough via arch-general wrote:
My guess is that they'll update that once pacman 6 hits stable.
On 01/06/2021 23:15, Yash Karandikar via arch-general wrote: pacman=6.0.0-2 is already in stable. ;)
Follow up on my earlier report of jittery output on pacman update. I updated 3 machines and 2 out of 3 showed the character jumps and newline, the third appears normal. Mirror issue?
The new progressbar requires specific escape codes (part of VT100 terminal definition I think) to be supported. Atleast one terminal emulator (qterminal ) has already been found not to support those , see https://bbs.archlinux.org/viewtopic.php?id=266831 . Are those 3 machines using different terminal emulators ? Lone_Wolf
On 6/3/21 4:17 AM, Lone_Wolf via arch-general wrote:
The new progressbar requires specific escape codes (part of VT100 terminal definition I think) to be supported.
Atleast one terminal emulator (qterminal ) has already been found not to support those , see https://bbs.archlinux.org/viewtopic.php?id=266831 .
Are those 3 machines using different terminal emulators ?
Lone_Wolf
I did all 3 updates from different tabs in the same konsole (KDE), so the terminal is VT100 compatible. It really looked like there was a bad ANSI escape somewhere. It was night/day different than the last time I updated and I've probably update each of these boxes of 1000 times during the past 10 years or so. -- David C. Rankin, J.D.,P.E.
On 6/3/21 1:09 PM, David C. Rankin via arch-general wrote:
I did all 3 updates from different tabs in the same konsole (KDE), so the terminal is VT100 compatible. It really looked like there was a bad ANSI escape somewhere. It was night/day different than the last time I updated and I've probably update each of these boxes of 1000 times during the past 10 years or so.
My guess would be that the old (KDE Trinity) version of konsole that you're using doesn't support some modern feature of terminals. DR
On 2021-06-03 at 14:00:08 -0400,
David Rosenstrauch via arch-general
On 6/3/21 1:09 PM, David C. Rankin via arch-general wrote:
I did all 3 updates from different tabs in the same konsole (KDE), so the terminal is VT100 compatible. It really looked like there was a bad ANSI escape somewhere. It was night/day different than the last time I updated and I've probably update each of these boxes of 1000 times during the past 10 years or so.
My guess would be that the old (KDE Trinity) version of konsole that you're using doesn't support some modern feature of terminals.
FWIW, I use xterm (currently at version 367), and I also noticed a difference in the output after the recent pacman upgrade. I haven't seen the level of disruption reported in this thread, but I'm pretty sure that I didn't see the "total" bandwidth/time output until the downloads were complete before, either. There's an extra blank line after that total line, too, instead of whatever is being downloaded showing progress on the very last line of the xterm window. (I figured it was a small improvement to see that line during the downloads, and let it go at that. No, I have no physical evidence of the change, only my recollection.)
On 6/3/21 1:00 PM, David Rosenstrauch via arch-general wrote:
On 6/3/21 1:09 PM, David C. Rankin via arch-general wrote:
I did all 3 updates from different tabs in the same konsole (KDE), so the terminal is VT100 compatible. It really looked like there was a bad ANSI escape somewhere. It was night/day different than the last time I updated and I've probably update each of these boxes of 1000 times during the past 10 years or so.
My guess would be that the old (KDE Trinity) version of konsole that you're using doesn't support some modern feature of terminals.
DR
No, that's not it either, I have used ALL ANSI escapes (and tput, etc..) in the terminals without issue. The key here seems to relate to the "Total" line triggering a new line part way through the package download. The faster the host, the less noticeable the jitter. This is from the newer box that I though was NOT affected: linux-docs-5.12.9.arch1-1-x86_64 Total (0/4) 0.0 B 0.00 B/s 00:00 [--------------------------------------] 0% linux-lts-5.10.42-1-x86_64x86_64 22.0 MiB 5.75 MiB/s 00:04 [######################################] 100% Total (1/4) 20.9 MiB 6.08 MiB/s 00:17 [######--------------------------------] 16% linux-lts-headers-5.10.42-1-x86_64 74.3 MiB 5.85 MiB/s 00:13 [######################################] 100% Total (2/4) 95.5 MiB 6.08 MiB/s 00:05 [############################----------] 74% nodejs-16.3.0-1-x86_64.42-1-x86_64 22.7 MiB 5.57 MiB/s 00:04 [######################################] 100% Total (3/4) 117.8 MiB 6.06 MiB/s 00:01 [##################################----] 92% Total (4/4) 127.8 MiB 5.73 MiB/s 00:22 [######################################] 100% It was affected, it just wasn't as noticeable when pacman was running because the screen refresh was fast. Also note, I do all updates over SSH (and Wifi) to headless boxes, so there many be some network delay contributing as well. But the point being, for 12 years I have never seen anything like this issue, and now, it is quite notable. -- David C. Rankin, J.D.,P.E.
On 6/3/21 4:59 PM, David C. Rankin via arch-general wrote:
On 6/3/21 1:00 PM, David Rosenstrauch via arch-general wrote:
On 6/3/21 1:09 PM, David C. Rankin via arch-general wrote:
I did all 3 updates from different tabs in the same konsole (KDE), so the terminal is VT100 compatible. It really looked like there was a bad ANSI escape somewhere. It was night/day different than the last time I updated and I've probably update each of these boxes of 1000 times during the past 10 years or so.
My guess would be that the old (KDE Trinity) version of konsole that you're using doesn't support some modern feature of terminals.
DR
No, that's not it either, I have used ALL ANSI escapes (and tput, etc..) in the terminals without issue.
Actually, I think that might be the issue. (Or at least part of it.) I don't have these issues using xfce4-terminal. DR
On 6/3/21 4:07 PM, David Rosenstrauch via arch-general wrote:
Actually, I think that might be the issue. (Or at least part of it.) I don't have these issues using xfce4-terminal.
g-darnit, now I have to patch that too! Use xterm and there was still a bit of wonkiness like it was trying to download and display in parallel, but no duplicate totals. Sheeze.... -- David C. Rankin, J.D.,P.E.
Does adding ILoveCandy to /etc/pacman.conf make a difference? I don't notice a progress bar issue. My setup: $ grep IL /etc/pacman.conf ILoveCandy $ pacman -Qi roxterm | grep On Depends On : dbus-glib vte3 hicolor-icon-theme $ pacman -Q roxterm vte3 roxterm 3.11.1-1 vte3 0.64.1-1
On 6/3/21 4:21 PM, Ralf Mardorf via arch-general wrote:
Does adding
ILoveCandy
Sounds like a slogan for a lollipop store rather than something that should be part of a Linux package manager. I'll try anything, but this seems like a largely self-imposed wound. Why the informative text output from a package manager should use anything other than basic terminal capabilities seems to fly in the face of the Arch KISS philosophy. (damn kids with crayons...) -- David C. Rankin, J.D.,P.E.
Hi David,
Why the informative text output from a package manager should use anything other than basic terminal capabilities seems to fly in the face of the Arch KISS philosophy.
As long as it's querying terminfo(5) to check for the capability and using the escape sequence presented there if it exists then that would seem fine. If it just blindly writes bytes and assumes the world can understand them... And Thomas Dickey's xterm(1) is superb in its correctness so that's the main thing to test it with. -- Cheers, Ralph.
On 6/6/21 2:56 AM, Ralph Corderoy via arch-general wrote:
Hi David,
Why the informative text output from a package manager should use anything other than basic terminal capabilities seems to fly in the face of the Arch KISS philosophy.
As long as it's querying terminfo(5) to check for the capability and using the escape sequence presented there if it exists then that would seem fine. If it just blindly writes bytes and assumes the world can understand them... And Thomas Dickey's xterm(1) is superb in its correctness so that's the main thing to test it with.
Is there any simple way to turn of the parallel download of the packages files that seems to be at the root of this "new technology". I would be quite happy to just have the files download sequentially again? Also, when I update my server at home, this seems aimed ab better saturating the bandwidth to speed things up. Only problem is if my kids are playing games -- the howls of bad ping times and cries of "Dad what are you doing?" see to come without much delay :) -- David C. Rankin, J.D.,P.E.
On 07-06-2021 10:46, David C. Rankin via arch-general wrote:
On 6/6/21 2:56 AM, Ralph Corderoy via arch-general wrote:
Hi David,
Why the informative text output from a package manager should use anything other than basic terminal capabilities seems to fly in the face of the Arch KISS philosophy. As long as it's querying terminfo(5) to check for the capability and using the escape sequence presented there if it exists then that would seem fine. If it just blindly writes bytes and assumes the world can understand them... And Thomas Dickey's xterm(1) is superb in its correctness so that's the main thing to test it with.
Is there any simple way to turn of the parallel download of the packages files that seems to be at the root of this "new technology". I would be quite happy to just have the files download sequentially again?
Also, when I update my server at home, this seems aimed ab better saturating the bandwidth to speed things up. Only problem is if my kids are playing games -- the howls of bad ping times and cries of "Dad what are you doing?" see to come without much delay :)
Maybe you could test with a console terminal like getty / agetty instead of a terminal emulator ? Try setting ParallelDownloads = 0 in pacman.conf . LW
On 6/7/21 3:56 AM, Lone_Wolf via arch-general wrote:
Maybe you could test with a console terminal like getty / agetty instead of a terminal emulator ?
Try setting ParallelDownloads = 0 in pacman.conf .
LW
Good thought, but ... error: config file /etc/pacman.conf, line 38: value for 'ParallelDownloads' has to be positive : '0' So it appears we have a new "feature" that is rendered a "bug" for some because it cannot be turned off :( -- David C. Rankin, J.D.,P.E.
On 07/06/2021 11.43, David C. Rankin via arch-general wrote:
On 6/7/21 3:56 AM, Lone_Wolf via arch-general wrote:
Maybe you could test with a console terminal like getty / agetty instead of a terminal emulator ?
Try setting ParallelDownloads = 0 in pacman.conf .
LW Good thought, but ...
error: config file /etc/pacman.conf, line 38: value for 'ParallelDownloads' has to be positive : '0'
So it appears we have a new "feature" that is rendered a "bug" for some because it cannot be turned off :(
From pacman.conf(5):
ParallelDownloads Specifies number of concurrent download streams. The value needs to be a positive integer. If this config option is not set then only one download stream is used (i.e. downloads happen sequentially).
On 6/7/21 4:46 AM, Xiretza via arch-general wrote:
From pacman.conf(5):
ParallelDownloads Specifies number of concurrent download streams. The value needs to be a positive integer. If this config option is not set then only one download stream is used (i.e. downloads happen sequentially).
Found that after the error, but setting ParallelDownloads = 1 Makes no difference whatsoever. The same discombobulated output is seen, e.g. :: Proceed with installation? [Y/n] :: Retrieving packages... inetutils-2.0-1-x86_64 Total (0/5) 0.0 B 0.00 B/s 00:00 [--------------------------------------] 0% libgusb-0.3.7-1-x86_64 345.3 KiB 2.06 MiB/s 00:00 [######################################] 100% Total (1/5) 15.8 KiB 0.00 B/s --:-- [--------------------------------------] 0% liblouis-3.18.0-1-x86_64 65.1 KiB 857 KiB/s 00:00 [######################################] 100% Total (2/5) 361.1 KiB 523 KiB/s 00:11 [#-------------------------------------] 5% mpg123-1.28.0-1-x86_6464 2.0 MiB 4.67 MiB/s 00:00 [######################################] 100% Total (3/5) 1898.4 KiB 2.48 MiB/s 00:01 [###########---------------------------] 30% openmpi-4.1.1-1-x86_64 432.7 KiB 2.35 MiB/s 00:00 [######################################] 100% Total (4/5) 2.5 MiB 2.58 MiB/s 00:01 [##############------------------------] 39% Total (5/5) 6.2 MiB 3.13 MiB/s 00:02 [######################################] 100% I'll just live with it until I have time to dig into the pacman source and find out what this supposed "new feature" is that screws up old terminals. -- David C. Rankin, J.D.,P.E.
On 6/7/21 3:49 PM, David C. Rankin via arch-general wrote:
I'll just live with it until I have time to dig into the pacman source and find out what this supposed "new feature" is that screws up old terminals.
Or you could just switch to a modern terminal app! :-) DR
On Mon, Jun 7, 2021 at 10:13 PM David Rosenstrauch via arch-general < arch-general@lists.archlinux.org> wrote:
On 6/7/21 3:49 PM, David C. Rankin via arch-general wrote:
I'll just live with it until I have time to dig into the pacman source and find out what this supposed "new feature" is that screws up old terminals.
Or you could just switch to a modern terminal app! :-)
DR
It works fine in konsole - but not in qterninal. I have put in an 'issue' report upstream for qterminal but it has not been acted upon yet. So I am happy using konsole. I guess that in time the necessary escapes will be fixed in those terminal emulators where it fails. The easiest 'solution' as has been mentioned is to use a terminal application where it does work. -- mike c
On 6/7/21 4:50 PM, Mike Cloaked via arch-general wrote:
It works fine in konsole
Well... this is KDE3 konsole :) -- David C. Rankin, J.D.,P.E.
On Mon, 7 Jun 2021 22:50:21 +0100
Mike Cloaked via arch-general
On Mon, Jun 7, 2021 at 10:13 PM David Rosenstrauch via arch-general < arch-general@lists.archlinux.org> wrote:
On 6/7/21 3:49 PM, David C. Rankin via arch-general wrote:
I'll just live with it until I have time to dig into the pacman source and find out what this supposed "new feature" is that screws up old terminals.
Or you could just switch to a modern terminal app! :-)
DR
It works fine in konsole - but not in qterninal. I have put in an 'issue' report upstream for qterminal but it has not been acted upon yet. So I am happy using konsole. I guess that in time the necessary escapes will be fixed in those terminal emulators where it fails. The easiest 'solution' as has been mentioned is to use a terminal application where it does work.
Works just fine in Xterm as well or is that too older tech these days Pete
On 6/7/21 4:13 PM, David Rosenstrauch via arch-general wrote:
Or you could just switch to a modern terminal app! :-)
DR
Blasphemy! When they pry my cold dead fingers off my konsole :) -- David C. Rankin, J.D.,P.E.
On 2021-06-08 10:19, David C. Rankin via arch-general wrote:
Blasphemy! When they pry my cold dead fingers off my konsole :)
Modern Konsole has tiling and all kinds of niceties, great wayland support and just works :) Maybe it's time to go with the times... Cheers
On 06/07/21 at 02:49pm, David C. Rankin via arch-general wrote:
On 6/7/21 4:46 AM, Xiretza via arch-general wrote:
From pacman.conf(5):
ParallelDownloads Specifies number of concurrent download streams. The value needs to be a positive integer. If this config option is not set then only one download stream is used (i.e. downloads happen sequentially).
Found that after the error, but setting
ParallelDownloads = 1
Makes no difference whatsoever. The same discombobulated output is seen, e.g.
:: Proceed with installation? [Y/n] :: Retrieving packages... inetutils-2.0-1-x86_64 Total (0/5) 0.0 B 0.00 B/s 00:00 [--------------------------------------] 0% libgusb-0.3.7-1-x86_64 345.3 KiB 2.06 MiB/s 00:00 [######################################] 100% Total (1/5) 15.8 KiB 0.00 B/s --:-- [--------------------------------------] 0% liblouis-3.18.0-1-x86_64 65.1 KiB 857 KiB/s 00:00 [######################################] 100% Total (2/5) 361.1 KiB 523 KiB/s 00:11 [#-------------------------------------] 5% mpg123-1.28.0-1-x86_6464 2.0 MiB 4.67 MiB/s 00:00 [######################################] 100% Total (3/5) 1898.4 KiB 2.48 MiB/s 00:01 [###########---------------------------] 30% openmpi-4.1.1-1-x86_64 432.7 KiB 2.35 MiB/s 00:00 [######################################] 100% Total (4/5) 2.5 MiB 2.58 MiB/s 00:01 [##############------------------------] 39% Total (5/5) 6.2 MiB 3.13 MiB/s 00:02 [######################################] 100%
I'll just live with it until I have time to dig into the pacman source and find out what this supposed "new feature" is that screws up old terminals.
We are now 26 messages deep in a thread that has nothing to do with the original post and with no real progress being made. The obvious suspect is that your terminal does not support the ANSI escape codes used for the new multi-line progress bars. You have two options: use a terminal that does or disable the progress bars. If you find something to indicate that this is not the source of your problem, file a bug with pacman. Otherwise, can we please let this thread die? apg
On 6/7/21 5:43 AM, David C. Rankin via arch-general wrote:
On 6/7/21 3:56 AM, Lone_Wolf via arch-general wrote:
Maybe you could test with a console terminal like getty / agetty instead of a terminal emulator ?
Try setting ParallelDownloads = 0 in pacman.conf .
LW
Good thought, but ...
error: config file /etc/pacman.conf, line 38: value for 'ParallelDownloads' has to be positive : '0'
So it appears we have a new "feature" that is rendered a "bug" for some because it cannot be turned off :(
Zero parallel downloads, i.e. downloading zero things at a time, seems like it would be a serious regression from pacman 5, given you could not ever download anything. Seriously, this thread feels weird. If you don't like ParallelDownloads, the natural first step is to disable it after reading the manpage to see how it works. Instead we're getting all kinds of wild assertions about easter eggs being "damn kids with crayons", even though it's been in your pacman since 2006. (I do not care if someone else suggested it might help. You could have figured out it would not help, and then just not complained about it.) It's hard to take your problem seriously when you come across as... weirdly grumpy... about it. It sort of feels like you don't truly want the solution, and actually trying to help you out with a solution would entail fighting you to force you to accept it. On 6/7/21 5:45 AM, David C. Rankin via arch-general wrote:
ParallelDownloads = 1
works (and hopefully that means 1 at-a-time and not 1 + 1 in parallel...
... Ok, so if it works then that means your output is happy now, right? Does it even *matter* if your output is happy, but you're retrieving 2 things in parallel? Your objective here was to get uncorrupted output... *right*? -- Eli Schwartz Bug Wrangler and Trusted User
On 6/7/21 6:48 AM, Eli Schwartz via arch-general wrote:
... Ok, so if it works then that means your output is happy now, right? Does it even *matter* if your output is happy, but you're retrieving 2 things in parallel? Your objective here was to get uncorrupted output... *right*?
No, No need worrying further about it. Since no one is divulging what changed, I'll just set this issue aside until I have time to wade into the changes and figure it out. Nothing is "broken", the terminal output changed and no longer displays without showing jagged progress bars and extra (Total X/Y) between each package download on TDE/KDE3 konsole. It's more a visual thing. pacman downloads and updates just fine. I'll look into in in the next week or so. -- David C. Rankin, J.D.,P.E.
On 6/7/21 3:56 AM, Lone_Wolf via arch-general wrote:
Maybe you could test with a console terminal like getty / agetty instead of a terminal emulator ?
Try setting ParallelDownloads = 0 in pacman.conf .
LW
ParallelDownloads = 1 works (and hopefully that means 1 at-a-time and not 1 + 1 in parallel... -- David C. Rankin, J.D.,P.E.
Hello David,
ILoveCandy Sounds like a slogan for a lollipop store rather than something that should be part of a Linux package manager.
It's an _easter egg_, not an official feature - but besides that, I don't actually believe it could have the potential to fix your issue.
I'll try anything [...] -- David C. Rankin, J.D.,P.E.
I haven't tried that myself on any affected terminal emulator, but it might be worth a shot to see how it behaves when you apply `--noprogessbar` to your pacman commands. Best of luck ! Kriss -- PGP: 7A19 4E3F 7A8F 867B EA8A 5339 023F 0788 62AC FE50 https://keys.openpgp.org/vks/v1/by-keyid/023F078862ACFE50
Kr1ss via arch-general
Hello David,
ILoveCandy Sounds like a slogan for a lollipop store rather than something that should be part of a Linux package manager.
It's an _easter egg_, not an official feature - but besides that, I don't actually believe it could have the potential to fix your issue.
I'll try anything [...] -- David C. Rankin, J.D.,P.E.
I haven't tried that myself on any affected terminal emulator, but it might be worth a shot to see how it behaves when you apply `--noprogessbar` to your pacman commands.
Is David using ParallelDownloads? If yes, perhaps that worse turning off either. -- u34
Best of luck !
Kriss -- PGP: 7A19 4E3F 7A8F 867B EA8A 5339 023F 0788 62AC FE50 https://keys.openpgp.org/vks/v1/by-keyid/023F078862ACFE50
On 2021-06-03 at 12:09:59 -0500,
"David C. Rankin via arch-general"
On 6/3/21 4:17 AM, Lone_Wolf via arch-general wrote:
The new progressbar requires specific escape codes (part of VT100 terminal definition I think) to be supported.
Atleast one terminal emulator (qterminal ) has already been found not to support those , see https://bbs.archlinux.org/viewtopic.php?id=266831 .
Are those 3 machines using different terminal emulators ?
Lone_Wolf
I did all 3 updates from different tabs in the same konsole (KDE), so the terminal is VT100 compatible. It really looked like there was a bad ANSI escape somewhere. It was night/day different than the last time I updated and I've probably update each of these boxes of 1000 times during the past 10 years or so.
I have a stupid question: Is it possible that one or more of those tabs are out of sync with respect to the actual size of the window and the remote systems? Years (if not decades) ago, there used to be problems along those lines if you (generic you) resized the terminal on the local machine because the remote machine wouldn't know about it. Since then, things like SIGWINCH work way better than they used to. To test that hypothesis, open a new local window and log in to one of the remote systems giving you trouble. (Now that I re-read that, it came out condescending, but that's not at all my intent. Someone downthread mentioned an ancient version of konsole and my brain went to that old problem. I'm only trying to explain your symptoms and a possible solution.)
participants (18)
-
2QdxY4RzWzUUiLuE＠potatochowder.com
-
Andrew Gregory
-
Anton Hvornum
-
Bennett Piater
-
David C. Rankin
-
David Rosenstrauch
-
Eli Schwartz
-
Filipe Laíns
-
Jonathon Fernyhough
-
Kr1ss
-
Lone_Wolf
-
Mike Cloaked
-
pete
-
Ralf Mardorf
-
Ralph Corderoy
-
u34＠net9.ga
-
Xiretza
-
Yash Karandikar