You are not logged in.
Hi, Devuman. Two quick things that may or may not be obvious:
1. For my double-ssh exercise to work, make sure R2 is not forwarding port 22.
2. The < and > symbols in my posts (and everything between them) are just placeholders. This will never work:
$ ssh -t <bob>@<192.168.18.8> "ssh <bob>@<192.168.1.101>"The proper syntax would be something like this:
$ ssh -t bob@192.168.18.8 "ssh bob@192.168.1.101"P.S. Actually, before trying do do fancy stuff (two-hop ssh), it's best to establish that your setup is allowing the easy stuff. Can you ssh from LT2 to R2? How about from LT1 to PC1? Those should be trivial. If not, fix this first.
Here's how to kill openvpn in one line in terminal:
sudo pkill -f openvpnDepending on how you run openvpn, you may not need the -f.
If the above "double ssh" command works, then we've established that our assumptions about ports (that the default port 22 is being used everywhere) and firewalls (that R2 and PC1 are allowing incoming on port 22) are correct.
To achieve SFTP via port forwarding, you have to tell R1 to take incoming traffic on port 22 and forward it to R2. Then you have to tell R2 to take incoming traffic on port 22 and forward it to PC1. I think that should do it.
I can SFTP from LT1 into PC1.
That's not surprising because they're both on the same network (LAN2).
Not from LT2 into PC1.
That's because LT2 and PC1 are on different networks, so it's not trivial to establish a connection. We have to go through R2. Let's see if we can SSH from LT2 to PC1 (if you can do SSH, then you can do SFTP).
At the risk of stating the obvious, remember that a router by definition stands between two networks, so it should have at least two IP addresses. R2 in your example has an IP address on LAN1 and a different IP address on LAN2.
Try this:
1. Make sure R2 and PC1 are both running SSHD (I'm assuming those two devices are not running a firewall, or at least have the SSH port open)
2. Run this command on LT2:
ssh -t <R2user>@<R2address-on-LAN1> "ssh <PC1user>@<PC1address-on-LAN2>"Here is an example with made-up passwords and ip addresses:
ssh -t devuman@192.168.1.150 "ssh devuman@192.168.2.135"You will be prompted for a password twice (first your password on R2, then your password on PC1).
P.S. Nice diagram!
EDIT: If you can establish an SSH connection from LT2 to PC1 using this long method, then we can be optimistic that with the right port forwarding rules we can achieve an SFTP connection without too much fuss. See my subsequent post below.
@contrast - The best place to share this AppImage with a larger community would be at https://www.appimagehub.com but unfortunately the requirements for submitting an AppImage there are too strict (the application needs to be compiled on a machine with very ancient glibc, for instance). It's just too much of a hassle.
Hopefully people who want an AppImage of gimp 2.8 will find this thread. I'll leave my tweaked, final version of the AppImage at this link indefinitely:
http://files.dantas.airpost.net/public/ … 4.AppImage
Cheers!
3 x Tiny Core Linux (my daily driver laptop plus my two wireless routers)
2 x Devuan ASCII (wife's laptops)
1 x Arch Linux ("media player" laptop connected to television--has not been updated since 2015)
I figured it out. Gimp cannot find the plugins inside the AppImage. I came up with a quick fix, which involves two changes:
1. At root of AppDir there is gimp.desktop. Change its Exec line from this:
Exec=gimp-2.8 %UTo this:
Exec=gimp-2.8-wrapper %U2. Create (and mark as executable) AppDir/usr/bin/gimp-2.8-wrapper with this in it:
#!/bin/sh
dir=/tmp/gimp2.8-plugins
if [ ! -d "$dir" ]; then
mkdir -p "$dir"
HERE="$(dirname "$(readlink -f "${0}")")"
cp -r "$HERE"/../lib/gimp/2.0/plug-ins "$dir"
fi
export GIMP2_PLUGINDIR="$dir"
exec gimp-2.8 "$@"Create AppImage from the tweaked AppDir. Now the AppImage works perfectly. Thanks again, Tatwi! ![]()
EDIT: I cleaned up the wrapper script a bit.
Hi, Tatwi. Thanks for sharing! The AppImage loads quickly and everything looks good. However, I can't open any files.
When I select File -> Open, no files show up in the file selection window unless I choose "All files". Then, whatever image file I select (even a common format such as jpeg) I get an error window saying "Unknown file type".
Alas, I don't know enough about how GIMP/GTK file selection works to fix this.
Hi, Tatwi. I use an AppImage of GIMP 2.10.8 but, like you, I'm unhappy with the recent changes.
Can you please share your AppImage of GIMP 2.8.x for 64-bit?
AppImage is also available. Sometimes issues in .deb version do not affect AppImage version:
https://musescore.org/en/download
If proficient UNIX/Linux usage is a craft, then the shell is the craftsman's workbench. I think having a solid grasp of bash--and knowing how it's different from other shells--is key. You can't go wrong by starting with the shell because it's essential no matter what your specific needs are. Where you go from there depends on your specific needs and interests.
- best bash tutorial anywhere: http://tldp.org/LDP/abs/html/
- compilation of bash pitfalls: https://mywiki.wooledge.org/BashPitfalls
- learn to recognize bashisms: https://mywiki.wooledge.org/Bashism
Write as many shell scripts as you can. After a script is working, make it correct. After it's correct, make it beautiful. Here is a tool to help you improve your existing shell scripts: https://github.com/koalaman/shellcheck
I hope that helps ![]()
Until I discovered fpm, creating my own packages was always a hassle. With fpm it becomes quite trivial.
In case you'd like to check it out: https://github.com/jordansissel/fpm
The official Tiny Core Linux wiki is down. Developers are aware and working on it. In the meantime, wiki can be accessed via this mirror:
https://www.linuxsecrets.com/tinycoreli … start.html
Their forum is alive and well:
http://forum.tinycorelinux.net
Main reason to strip binaries is to make them smaller (I've sometimes seen them shrink to 50-75% of original size). TCL packages contain stripped binaries in the interest of keeping everything as tiny as possible.
Lopsided, like I said. They expect blind trust but are themselves unwilling to trust.
Trust should go both ways, especially in times of crisis.
We can disagree on this, but I don't think it's ever acceptable to ask someone to run closed source software. Principles don't just fly out the window when times are tough. I think it should be the opposite--good principles provide guidance, should be like light beams to help us find our way when the path is dark.
These folks need to come up with an algorithm that disallows data manipulation/cheating, then open up the code. Keeping the code closed for the highlighted reason is just a copout for bad software design.
Asking people to pitch in computing power but they refuse to show the source code? That's lopsided and creepy.
I'll be happy to contribute and spread the word when they make the source code available.
I stand corrected, and edited my walk-through to reflect this. Thank you, HoaS.
You are not careful! I repeat!
Must be:hwclock --utc
2020-02-28 20:15:50.889268+00:00
That's not correct. The output of the hwclock command is your local time, and what comes after the final + or - is your timezone's offset from UTC. +00:00 would only be correct for users who live in Greenwich ![]()
Please go over the steps above in my longer post. If you still have questions after that, let me know and I'll try to help.
I agree with Eaglet that the way hwclock works is a bit confusing. So I decided to get to the bottom of things on my machine. Here's what I did. I hope it helps...
First, we have to realize that the hardware clock knows nothing about timezone. It holds a date and a time, that's it.
It's up to the user to tell the operating system the correct timezone and whether the hardware clock is set to UTC (recommended for Unix-like OS) or localtime (recommended if you dual boot with Windows).
Before I show you the output of some commands, here are some facts:
- It's currently 15:35 UTC
- I live in Eastern USA, where timezone is EST (UTC-5)
- My local time is 10:35 a.m.
Okay, now for the commands:
$ sudo hwclock --get --debug | grep "Time read" # this command will show exactly what the hardware clock is saying
Time read from Hardware Clock: 2020/03/05 15:35:32
# Hmmm, hardware clock says it's 15:35, so it's obviously set to UTC. Let's make sure my OS knows that...
$ cat /etc/adjtime
0.000000 1583421556 0.000000
1583421556
UTC
# Notice the third line in the output above. That line tells the OS how to interpret the hardware clock. Mine should say UTC, which it does.
$ cat /etc/timezone
America/New_York
# So hardware clock is set to the correct UTC time, OS knows to interpret hardware clock as UTC, and my timezone is correct. Therefore, when "date" or "hwclock" command puts it all together, it should spit out my correct local time:
$ date
Thu Mar 5 10:35:06 EST 2020 # yep, that's the correct local time
$ sudo hwclock # or sudo hwclock --utc
2020-03-05 10:35:09.546002-0500 # yep, that's the correct local timeSo as you can see, your OS showing you the correct local time is a complex operation. It requires 3 things: hardware clock being set correctly (to either UTC or local time), your OS knowing whether hardware clock is set to UTC or local time, and your OS knowing your time zone.
Hope that helps.
P.S. The --utc and --localtime flags to hwclock are to tell hwclock which timescale the hardware clock is set to (hwclock assumes UTC, so the --utc flag is redundant). Eaglet, If you your hardware clock is set to UTC you should never use the --localtime flag because it will give you incorrect (and potentially very confusing) results.
No resolution. Also, I don't know if this is peculiar to T400 or all laptops.
Note that the ExpressCard came with a power cable, but I don't use it because I don't want to carry around an extra cable.
So here is what I've discovered:
- ExpressCard never works on battery power.
- ExpressCard works if laptop was on AC when it booted and has been on AC continuously up to and including when I use the ExpressCard.
- If laptop was on battery power when it booted and later I switch to AC power, I need to eject and re-insert the ExpressCard in order for it to work.
Hope this helps.
If that's true, switching your email client to UTF-16 encoding would properly display the wide characters in your friend's emails--but would cause problems displaying wide characters in emails to you from people who use UTF-8 (the de facto standard encoding these days).
Ron,
I don't think the fact that your friend's email client is proprietary has anything to do with it. Also, uni2ascii is unlikely to help. It sounds like an encoding issue.
The emoji in question is the "slightly smiling face" (formal Unicode notation: U+1F642, details about it here: https://www.iemoji.com/view/emoji/1112/ … ling-face). Here are the relevant details:
UTF-8 encoding: 0xf09f9982
UTF-16 encoding: 0xd83dde42 (notice this matches the "D83D "DE42" in your original post)
How are you viewing your friend's emails? In a web browser? If so, take a look at that emoji here:
https://unicode.org/emoji/charts/full-e … html#1f602
If you can see the "slightly smiling face" emoji there under the "Browser" column, we'll know it's not a browser or font issue on your end.
My best guess is that your friend's email client is using UTF-16 encoding, while your email client is expecting UTF-8 encoding. Different encodings agree on how to represent 7-bit ASCII characters but differ on how they represent wider characters such as emojis.
Have your friend check their outgoing mail encoding and change it to UTF-8 if it's currently something else.
I bumped into this the other day and found it to be the best and most level-headed criticism of systemd I've yet seen:
http://landley.net/notes-2014.html#23-04-2014
It's worth reading.
Bottomline:
Linux-based OSes have always been modular--built of interchangeable components (with the exception of the kernel, which gets a pass because it defines these OSes). This allows for a lot of flexibility and is the reason why tinkerers find Linux attractive.
BusyBox and systemd are, in a way, good twin and evil twin: Both are monolithic userspaces, but while BusyBox goes out of its way to respect modularity (user can enable all, none, or any subset of its functionality; it can coexist with tools that provide similar functionality), systemd does the opposite.
Forking the OpenBSD kernel in order to avoid systemd and/or other linuxland woes makes no sense. By virtue of not being Linux, OpenBSD (as-is) does not suffer from linuxland woes.
@nobodyuknow - Try adding blacklist pcspkr to /etc/modprobe.d/blacklist (create /etc/modprobe.d/blacklist if it doesn't already exist).
With extensions/packages. I don't think it would be appropriate to take up space on Devuan's servers to go into great detail here. If you decide to try it and need help, start a thread on forum.tinycorelinux.net or send me a personal message there (my user name is GNUser there, too).
Everything you need to know to get started is in the gratis e-book "Into the Core":
http://www.tinycorelinux.net/corebook.pdf