macports install obsolete port - pyqt

When I execute sudo port install py33-pyqt5 in my terminal, I get this error message:
---> Configuring py33-pyqt5
Error: py33-pyqt5 has been made obsolete by the port py34-pyqt5. Please install py34-pyqt5 instead.
Error: org.macports.configure for port py33-pyqt5 returned: obsolete port
Please see the log file for port py33-pyqt5 for details:
/opt/local/var/macports/logs/_opt_local_var_macports_sources_rsync.macports.org_release_tarballs_ports_python_py-graveyard/py33-pyqt5/main.log
To report a bug, follow the instructions in the guide:
http://guide.macports.org/#project.tickets
Error: Processing of port py33-pyqt5 failed
But I need to install py33-pyqt5 and not py34-pyqt5. Is there a possible way to install the old version?

You can use the same approach that is used to install an older port: http://trac.macports.org/wiki/howto/InstallingOlderPort. For a Python port that uses the subport feature, that probably means finding an older version of the Portfile that now provides py34-pyqt5 that did still provide py33-pyqt5 and running sudo port install -- subport=py33-pyqt5 in a directory that contains this older Portfile. Note that the Portfile and the directory must be readable by the macports user, so your home directory will probably not work very well. I recommend a place in /tmp.
There has been a decision to phase out support for python 3.3 in MacPorts, though -- you will run into more and more problems when trying to use Python 3.3, so you should really consider switching to a newer version.

Related

How to run application that installed using Snapcraft

I just installed Snapcraft and then using it to install Ngrok, but unfortunately I still unable to run Ngrox from my terminal.
My Ngrok Installation
╭─yoesoff#yoesoff-ThinkPad-X220 ~/Documents/labs/hello-world ‹master›
╰─➤ sudo snap install ngrok 1 ↵
ngrok 2.3.27 from Khiem Doan (khiemdoan) installed
Ngrk Is Listed Now
╭─yoesoff#yoesoff-ThinkPad-X220 ~/Documents/labs/hello-world ‹master›
╰─➤ snap list 127 ↵
Name Version Rev Tracking Publisher Notes
core 16-2.39 6964 stable canonical✓ core
ngrok 2.3.27 13 stable khiemdoan -
But It Does not Work
╭─yoesoff#yoesoff-ThinkPad-X220 ~
╰─➤ ngrok
zsh: command not found: ngrok
My question is how to run an application that installed using snapcraft ?
Thanks you
The problem is caused by Snapcraft poor documentation I think. the only way to run the installed application by Snap should be run by Snap either. Here is how to run an application that installed using Snap:
snap run ngrok --version
ngrok version 2.3.27
This way will not work
ngrok
zsh: command not found: ngrok
A little bit weird but I think it's better to unzip Ngrok manually instead of using this poor tool.
You can run the application using snap run application (see snap run --help).
If you want the application to display in the desktop applications list you should find the application.desktop file in /var/lib/snapd/desktop/applications/, then you can symlink it to .local/share/applications/.
Or better you can add /var/lib/snapd/desktop/applications/ to XDG_DATA_DIRS by putting in ~/.profile (or ~/.gnomerc if you want to apply it only to gnome) something like this:
export XDG_DATA_DIRS="${XDG_DATA_DIRS}:/var/lib/snapd/desktop"
or if /etc/profile.d/snapd.sh exists
source /etc/profile.d/snapd.sh

PCRE issue when setting up WSGI application

I am working with Ubuntu 16.04.2 LTS.
I have been following a guide
How To Set Up uWSGI and Nginx to Serve Python Apps on Ubuntu 14.04. Once I have set up the virtualenv I follow the instructions:
pip install uwsgi
You can verify that it is now available by typing:
uwsgi --version
If it returns a version number, the uWSGI server is available for use.
However when I do this I get:
uwsgi: error while loading shared libraries: libpcre.so.1: cannot open
shared object file: No such file or directory
If I push on and work further through the guide things fall over when I try use uwsgi.
My research tells me that PCRE is Perl Compatible Regular Expressions and several people have asked questions online with libpcre.so.1 issues with other applications.
For example a response to a similar issue relating to nginx:
The message means what it says. The nginx executable was compiled to
expect the PCRE (Perl-compatible Regular Expression) shared library to
be available somewhere on LD_LIBRARY_PATH or specified in
/etc/ld.so.conf or whatever equivalent library-locating mechanisms
apply to your operating system, and it cannot find the library.
You will need to install PCRE - or configure your environment so that
nginx will look for the PCRE library where it is installed.
But I can't find much relevant to installing PCRE or configuring it. Most install instructions use: apt-get install libpcre3 libpcre3-dev and then reinstalling uwsgi pip install uwsgi -I. As in this example. Where I have tried everything posted and got nowhere.
I think my principle issue is that I don't understand the problem very well or how to do the things mentioned in the nginx example above.
Any insight or guidance would be much appreciated.
Even though my context may be different, the following steps should help you as well.
I did pip install uwsgi into my environment created by conda create -yn <env_name> python. Note, that one wouldn't even need to install PCRE into the environment, because it is included with Anaconda. We can see this issue in the environment, after source activate <env_name>:
# uwsgi --version
uwsgi: error while loading shared libraries: libpcre.so.1: cannot open...
With root/sudo access you can find where libpcre.so.1 is/will be:
# find / -name libpcre.so.1
/opt/anaconda3/lib/libpcre.so.1
Now let Linux know how to access it:
# ldconfig /opt/anaconda3/lib/
That's all you need to make it work. You can see the change you are making:
# find / -name uwsgi
/opt/anaconda3/envs/<env_name>/bin/uwsgi
# ldd -d /opt/anaconda3/envs/<env_name>/bin/uwsgi
linux-vdso.so.1 => (0x00007fff2d1ba000)
...
/lib64/ld-linux-x86-64.so.2 (0x00007ff98dbc5000)
undefined symbol: pcre_free (/opt/anaconda3/envs/cts/bin/uwsgi)
PS Turned out ldconfig above populates global cache /etc/ld.so.cache, which, in my case, clashed with system library (/lib/x86_64-linux-gnu/libdbus-1.so.3). So I had to revert the change by running ldconfig without parameters and resort to runtime linking = starting uwsgi as
# LD_LIBRARY_PATH=/opt/anaconda3/lib uwsgi --version

installing logstash plugins on windows 10

Wondering if anyone has had any luck running logstash on Windows 10. I can get the simple stdin input and stdout outputs to work, but I can't install any additional plugins. For example, running .\logstash-plugin.bat install logstash-input-heroku yields this output:
Validating logstash-input-heroku Installing logstash-input-heroku
Error Bundler::InstallError, retrying 1/10 An error occurred while
installing net-ssh (4.0.1), and Bundler cannot continue. Make sure
that gem install net-ssh -v '4.0.1' succeeds before bundling.
WARNING: SSLSocket#session= is not supported Error
Bundler::InstallError, retrying 2/10 An error occurred while
installing net-ssh (4.0.1), and Bundler cannot continue. Make sure
that gem install net-ssh -v '4.0.1' succeeds before bundling.
WARNING: SSLSocket#session= is not supported
(etc.)
I don't see windows 10 in the logstash support matrix so I'm guessing it's not intended to work. Perhaps on the linux subsystem for windows? Has anyone tried that? Doesn't look like java8 runs on it.
Well installation seems to have now succeeded. Thanks #Darth_Vader for the suggestion. Here's the source github issue that I referred to. The issue was resolved after updating the Gemfile (in logstash root folder) and changing the net-ssh and rest-client library versions:
# This is a Logstash generated Gemfile.
# If you modify this file manually all comments and formatting will be lost.
source "https://rubygems.org"
gem "net-ssh", "2.9.0"
gem "rest-client", "1.7.2"
gem "logstash-core", :path => "./logstash-core"
(etc...)

I want to install lucee on my VPS but cannot as I am getting error during post installation

I want to install lucee on my VPS but cannot as I am getting error during post installation.
Error running /opt/lucee/sys/install_mod_cfml.sh -m install -f
/usr/local/apache/conf/httpd.conf -d /usr/local/apache/modules -c
/usr/local/apache/bin/apachectl -k
a9681d10cecf61fea7aa9e4bd203e096e7e5a0c94fa1534d32cf6c0cd82f7d2c: httpd: Syntax
error on line 570 of /usr/local/apache/conf/httpd.conf: Cannot load
modules/mod_cfml.so into server: /lib64/libc.so.6: version `GLIBC_2.14' not
found (required by /usr/local/apache/modules/mod_cfml.so)
Can anybody point me in a right direction, how to solve the issue?
This means that the auto-detection of your Apache version or bit-type is failing. This means you'll need to install mod_cfml manually, or live without it.
When you run the Lucee installer, simply un-check the option to install mod_cfml. From there, you can install mod_cfml manually by following the instructions found on this page:
http://www.modcfml.org/index.cfm/install/web-server-components/
You can also add hosts to your tomcat config manually by editing your Tomcat server.xml file directly, and restarting Tomcat/Lucee every time you add a new host.

Check if a port is installed on FreeBSD

How can one check if a port is already installed on FreeBSD?
This is to be used in a shell script for provisioning. i.e. if port is already installed skip install, else call make command
Under FreeBSD 10+, you can use "pkg". The good reference is: https://wiki.freebsd.org/PkgPrimer
If your port is installed, it will result in:
# pkg info apache-jmeter
apache-jmeter-3.0.r2
Name : apache-jmeter
Version : 3.0.r2
Installed on : Thu May 5 18:58:45 2016 MSK
Origin : www/jmeter
Architecture : freebsd:10:x86:64
...
Otherwise:
# pkg info jmeter
pkg: No package(s) matching jmeter
You should never use "pkg delete"/"pkg remove" on software that you have installed via
the ports, but rather "make deinstall" from inside the ports tree. Doing "pkg install"
downloads a precompiled package to your system and installs it.
Doing "pkg info " will *not tell you if it was installed from ports.
I'm afraid the quoted statements are incorrect.
A 'port' is simply a 'package' that hasn't been compiled and installed yet.
Installed ports are in fact packages and can be treated like any other package regardless if it was downloaded pre-compiled and installed via pkg(8) or compiled locally from the ports tree and then installed. The terms 'port' and 'package' are synonymous once the software is 'installed'.
pkg delete/remove <portname> can and should be used and may be preferable as it will prompt to also uninstall no longer needed dependencies.
make deinstall in a specific port's directory will only uninstall that port and nothing else.
pkg info pkg info <portname> is the correct answer.

Resources