I have this error trying to download the new version:
Is that also happening to you?
Same problem for me also. curl and wget not working. However, downloading through browser works.
I want to use julia in a headless server to which I have only ssh access.
Please tell how to download that latest version to server?
Ok, I solved that problem (DNS_PROBE_FINISHED_NXDOMAIN) to access the new version and the Julia docs following this: networking - Ubuntu DNS Error Chrome: DNS_PROBE_FINISHED_NXDOMAIN, Firefox similar - Ask Ubuntu
Not working for me 
It’s updated but awaiting moderation. You can install it ahead of time by adding the flag --version 1.6.0 (it worked fine for me just now), but this may leave you with a bugged install if the Julia team decides to change the chocolatey package before it gets approved.
I could see that the 1.5.4 announcement and the 1.5.4 update on chocolatey were 1 day apart, so I think I will wait a day or two before doing something like that. But thanks for the tip on how to force it!
Are you behind a MITM proxy firewall?
First time to plot: 2.45 seconds. Bravo!
@time using PyPlot; plot(1:10)
2.451022 seconds (5.49 M allocations: 328.448 MiB, 7.24% gc time, 0.11% compilation time)
1-element Vector{PyCall.PyObject}:
PyObject <matplotlib.lines.Line2D object at 0x7f032c8981f0>
3 Likes
You must have a fast machine. Which CPU do you use?
Note that this is possibly a bit misleading, since you are only measuring the time using PyPlot takes. To measure the time it takes to actually see the first plot, you can do something like:
julia> @time (using PyPlot; display(plot(1:10)))
1-element Vector{PyCall.PyObject}:
PyObject <matplotlib.lines.Line2D object at 0x7f0923cf2390>
7.021832 seconds (11.55 M allocations: 674.571 MiB, 5.33% gc time)
4 Likes
I tested it on two computers that have an m.2 SSD and reached an FTTP of 3.5 seconds.
julia> @time (using PyPlot; display(plot(1:10)))
1-element Vector{PyCall.PyObject}:
PyObject <matplotlib.lines.Line2D object at 0x7fd80cdcfe20>
3.505962 seconds (10.30 M allocations: 602.439 MiB, 3.54% gc time, 0.17% compilation time)
Update:
On my Ubuntu 20.04 laptop the first time to plot was 7s. Then I changed the cpu governor from powersave to performance, and now it is down to 3.6 s. 
See: Ubuntu – How to set “Performance” instead of “Powersave” as default – Unix Server Solutions
6 Likes
Help can someone please remind me how to force IJulia to use Julia-1.6
I looked at my cheat notes and I did not put in instructions on how to get it to use Julia-1.6
using IJulia
notebook(dir="/Users/ssiew/juliascript/Notebooks",detached=true)
http://localhost:8888/tree/Notebooks
to stop Jupyter
/Users/ssiew/.julia/conda/3/bin/jupyter notebook stop 8888
to check the kernels of julia
/Users/ssiew/.julia/conda/3/bin/jupyter kernelspec list
Available kernels:
julia-1.4 /Users/ssiew/Library/Jupyter/kernels/julia-1.4
julia-1.5 /Users/ssiew/Library/Jupyter/kernels/julia-1.5
python3 /Users/ssiew/.julia/conda/3/share/jupyter/kernels/python3
usually you just need to ]build IJulia again.
2 Likes
I am using my university server and to the best of my knowledge it is not.
wget worked when I wanted to setup v1.5, but not working now.
I also tried to see what is the ip address of server so that I can use it directly in the URL.
I ran ping in my local machine and the output goes like
$ ping julialang-s3.julialang.org
PING julialang-s3.julialang.org(2a04:4e42:2::561 (2a04:4e42:2::561)) 56 data bytes
64 bytes from 2a04:4e42:2::561 (2a04:4e42:2::561): icmp_seq=1 ttl=48 time=1012 ms
64 bytes from 2a04:4e42:2::561 (2a04:4e42:2::561): icmp_seq=2 ttl=48 time=1005 ms
64 bytes from 2a04:4e42:2::561 (2a04:4e42:2::561): icmp_seq=3 ttl=48 time=1111 ms
64 bytes from 2a04:4e42:2::561 (2a04:4e42:2::561): icmp_seq=4 ttl=48 time=352 ms
64 bytes from 2a04:4e42:2::561 (2a04:4e42:2::561): icmp_seq=5 ttl=48 time=949 ms
...
But when I use the ip address directly in the wget, I get another error
$ wget https://2a04:4e42:2::561/bin/linux/x64/1.6/julia-1.6.0-linux-x86_64.tar.gz
https://2a04:4e42:2::561/bin/linux/x64/1.6/julia-1.6.0-linux-x86_64.tar.gz: Bad port number.
You can’t use an IPv6 in a URL in that way: IPv6 address - Wikipedia
I installed Julia late in octomber 2020 (Julia 1.5.2), coming from Python. Hence this is the first experience with installing a new version.
I searched for a tutorial addressed to beginners but no reference found. It’s a nightmare to navigate among hundreds
of SO and Julia discourse answers to questions similar with what I’d like to ask myself. Most of them refer to outdated versions
or are given in an ultra-specialised terms which a beginner is not yet familiar with.
Hence if you want to attract more users, please provide succint and clear instructions on what we have to do when a version is updated.
Finally after 2 hours I succeeded to make IJulia work. My older notebooks are running fine.
I installed Julia 1.6.0 on windows 10, and although I’ve set the path to it in Atom, 1.55.0, it still displays the error:
Julia cannot be started
We tried to launch Julia from:
c:\Users\myname\AppData\Local\Programs\Julia-1.6.0\bin\julia.exe
What is wrong or what I have to do to get it work?
Thanks!
1 Like
Can you start julia if you navigate to that folder on the command line and type
julia
?
1 Like