I'm a new hire that the company has recently given me a job, and I've got a
work notebook Mac, I need to code Linux server using C++ on Mac. I wonder if
is there any tips to do that?
I used to run Fedora on Mac using Parallels Desktop 11, and I heard that Docker
support Container carrying Linux on Mac, and I know that I could use Web Services
to run C++ code under Linux, as Amazon Elastic Compute Cloud(EC2: Cloud Server)
do, also, I could buy an external Raspberry Pi to run the need C++ Code at a
lower price and better solution.
So, that's clear.
Kill Linux C++ Back-end on Mac
1.Physical Linux Notebook(dual boot Linux and MacOS)
Dual boot building harmful.
2.Virtual Machine(running Linux. Connect via SSH and S FTP)
e.g: Vagrant+ VirtualBox.
This is the simplest route, but is effectively just building in Linux.
3.Docker(Linux based containers running on MacOS, via SSH and S FTP)
The docker image, this could end up being difficult to run/debug your
application and have the MacOS IDE understand error messages. Also use vagrant
instead of docker.
4.Cloud Host(remote Linux. Connect via SSH and S FTP)
Well idea.
5.[Raspberry Pi](https://www.raspberrypi.com/products/raspberry-pi-4-model-b/)(a proper Linux computer. Could as remote Linux. Connect via SSH and S FTP)
Nice!
6.Cross GCC on MacOS
Do Not recommend!
I'm ready to checkout the Journal, Blog, Stack Overflow, or Others.
Reading notes:
1.Setting up a Simple Docker Dev Environment
<about using 3.Docker>
Docker has two modules. The one is the static way, dependencies and testing,
or just as App release version; the another is dynamic way, application
under active development using "rsync", just as App debug version.
2.A productive development environment with Docker on OS X
<about using 3.Docker>
Edit Code at Mac, compile and test at LXC via Docker, sync the mac directory
to the virtual machine Docker. Mount my folder into my Docker via "rsync".
3.Running Docker on AWS from the ground up
<about using 4.Cloud Host>
Compute, ship and post your App using using Elastic Compute Cloud(EC2) via
Amazon Web Services. Multiple environment using Docker in EC2 Container
Services(ECS).
4.Infrastructure as code: running microservices on AWS using Docker, Terraform, and ECS
<about using 4.Cloud Host>
Auto build, ship and post platform via Docker, ECS and Terraform. That's
fancy and powerful. I could use that auto-idea on my Mac Docker for build,
ship or post.
5.Can you use a Mac OS to develop a C/C++ application that runs on a Linux server?
<about using 5.Raspberry Pi>
Using Raspberry Pi as external Linux computer, I could compile and test on
it. Hook it into my network so I could just remote logo into it - and arrange
for it to share it's filing system with my Mac. At that time, I could sit
in front of my Mac, use my Notebook Mac's editors and others stuff - and just
have a shell window opened with an "ssh"session onto that Raspberry Pi.
If I install "apache" on that Raspberry Pi - so it'll not only like a Linux
server - it'll actually BE a Linux server. Also, I could probably even do
full-up testing on it.
As Steve Baker said, for the sake of a few bucks and a day to get it set
up - I'll save myself a ton of grief.
6.Why can't I compile a C++ program on my Mac and run it on a Ubuntu server?
<about 2.Virtual Machine(running Linux)>
MacOS(Xcode/Command Line Tools) use clang-llvm with support C++14+.
<about 6.Cross GCC on MacOS>
Compiling for Linux on MacOS X, This is a cross compilation of GCC:
https://gcc.gnu.org/news.html. You can find here the releases of cross
compilers for Linux 32 and Linux 64.
8.Development environment setup for Mac and CentOS using Docker
<about using 3.Docker>
It's a really good idea to use Docker to standardise the development
configuration (softwares, custom packages, env variables, ...).
9.How to quit Docker for Mac and build a Docker development environment with Ubuntu + Vagrant
<about 2.Virtual Machine(running Linux)>
On Linux, the VFS is shared between the host and the container, so the
reflection without overhead is guaranteed.
However, on macOS (and other non-Linux platforms), there is significant
overhead for full consistency ...
10.Shipping C++ Programs in Docker: A Journey Begins
<about using 3.Docker>
Deploying C++ applications using Docker and Kubernetes.
The idea with multi-stage builds is that I can build the application in one
Docker container, and then copy only the production files to a second
Docker container.
The second Docker container can then be deployed and released to customers
without giving customers access to your source code.
Customers can run and connect to the container using Bash, but can only see
the configuration files and application binaries.
11.How to cross compile from Mac OS X to Linux x86?
<about using 3.Docker>
"backup your container as base build image docker commit [id] centos:build"
Because if you turn down Docker, the container will miss, until you use last
time with manually rebuild. Also, that could automate building images.
12.How to cross compile from Mac OS X to Linux x86?
<about 6.Cross GCC on MacOS>
Your simplest solution is to just run CentOS 5.3 in a VM (e.g. Sun VirtualBox).
This requires minimal setup, has quite reasonable overhead (assuming an Intel
Mac), and you'll be able to actually test and debug what you are building.
If you really insist on cross-compiling, you must build a cross-compiler.
Instructions are here and here, but beware: it will likely take you several
days to get it right, and then you'll still need a VM to test the result, so
I don't see any point in doing it that way.
13.Docker Explained: Using Dockerfiles to Automate Building of Images
<about using 3.Docker>
In this DigitalOcean article, we will see about automating this process as
much as possible, as well as demonstrate the best practices and methods to
make most of Docker and containers via Dockerfiles: scripts to build
containers, step-by-step, layer-by-layer, automatically from a base image.
And,
Editing Dockerfile just like write embeded language, matlab, or python.
I wonder if is it strictly speaking a programming language? I mean that
embeded, Matlab, or python, or even R.
Automate building of Images means that package the currently created
basic image with other installed app stack into the new image via Dockerfile.
Dockerfile initial image to automate building image, and easiest migration.
14.Notes about separately installing Linux via Docker on Mac(Chinese)
<about using 3.Docker>
Create a image, then start variety of containers. There may be Nginx, Redis,
MySQL, MongoDB, or even Memcached.
15.Config LNMP environment using Docker(Chinese)
<about using 3.Docker>
(DNMP)Docker+Nginx+MySQL+PHP+Redis.
(LNMP)Linux+Nginx+MySQL+PHP.
(LAMP)Linux+Apache+MySQL+PHP.
Directory architecture:
.
|_data
| |_elasticsearchdata
| |_mongodb
| |_mysql
| |_mysql8
|_services
| |_elasticsearch
| |_mysql
| |_mysql8
| |_nginx
| |_redis
|_log
|_docker-compose.yml
|_env
|_www
16.Develop C++ via Docker under CLion and VSCode(Chinese)
<about using 3.Docker>
Using VSCode debug remote virtual machine containers via Microsoft's'"Remote
Development", current workspace folder ".devcontainer" with "devcontainer.json"
and CMake extensions optional "Remote-Containers: Open Folder in Container".
17.The idea to config the development environment of LNMP via Docker on Mac(Chinese)
<about using 3.Docker>
The usage of Docker is that a container do only provide one service.
The designed principle of Docker is not to run services at Backend, actually
to run services at Frontend.
From downloading image command "docker pull linux_release", then, creating
container command "docker run ...", to entering container command "docker
exec ...".
I wonder if are you creating all containers under one image or one image
with one container?
Now, I know that the auth create container one by one image, all the image
are separated, also, the container under image are separated.
And, my guess was right. There are two ways to create multiple services,
the one is that one kind image create one kind container; the another
is that all services are created below one base image with starting
one container one service.
If create all containers under only one image, that just like a physical
Linux;
If create container one by one image, that all called separated system,
distributed system, computer cluster, or service farm.
18.Setup C/C++ development environment via Docker on Mac(Chinese)
<about using 3.Docker>
Automate Docker via four files: Dockerfile, entrypoint.sh, rsync.conf,
docker-compose.yml.
That's a good idea.
19.Create the development environment of DNMP via Docker on Mac(Chinese)
<about using 3.Docker>
Great DNMP directory:
.
|_conf
| |_conf.d
| | |_certs
| | | |_site2
| | |_site1.conf
| | |_site2.conf
| |__my.cnf
| |__nginx.conf
| |_php-fpm.d
| | |_www.conf
| |_php.ini
|__docker-compose.yml
|__docker-compose54.yml
|__docker-compose56.yml
|_log
| |__mysql.slow.log
| |__nginx.error.log
| |__nginx.site1.error.log
| |__nginx.site2.error.log
|_mysql
|_php
|_www
|_site1
|_site2
This blog also has more description about MySQL, Redis.
20.Docker build LNMP via deployer posting(Chinese)
<about using 4.Cloud Host + using 3.Docker>
"docker pull service" is to create image;
"apt-get install service" it to install services below that base
image.
Like LNMP or DNMP, one image, one container, one service;
Base image also could install many services.
21.Using Docker build up Linux server on Mac(Chinese)
<about using 3.Docker>
The installed many applications of Docker containers should be sync as a
new sync_image.
22.Setup the development environment ofr Linux server(Chinese)
<about using 2.Virtual Machine>
The "environment installed package" of Linux release version should be
installed separating from the system directory in order to conveniently
delete without affecting the system directory.
"oh-my-zsh" is the functional assistant of shell zsh.
"vim7.4+" is to witness the liking lambda of C++11.
"GCC9.1" supports the C11 C14 C+=17, well error prompt.
"bison" is to solve the warming of "'bison' is missing on your system".
"textinfo" is to solve the warming of "'makeinfo' is missing on your system".
"automake1.15+" is requisite for GCC.
"autoconf2.69" is essential to "automake1.15+".
"GCC" is crucial that that is vital to more than 30 minutes to install.
"GDB8.3" is the requirement of "GCC" with installing after "GCC".
"CMake3.15" product "Makefile" with source files.
"Ragel6.10" is the limited state machine to edit the parser protocol.
Others like "psmic", "openssl-devel", "net-tools".
23.Build Linux C++ development environmnet with VS2017(Chinese)
<about using 2.Virtual Machine>
VS2017 log into virtual machine via SSH.
Linux release below virtual machine install like "openssh-server", "g++",
"gdb" and "gdbserver".
24.What's the development environment of C/C++ under Linux(Chinese)
<about using Development Tools>
editor: Jetbrains 's paid CLion, vim, MicroSoft's VSCode,
script: zsh, bash, shellcheck,
debugger:
(Online)google-breakpad;
(LIVE)gdb(functionally stack trace, that bugger home location),
printf(<syslog, one printf let IO console apply more resources than
once IPC call syslog. Could relocate syslog, file, udp socket
(receive log output), define marco to realize classify sign or
level),
syslog(call IPC, e.g:log4cxx(support UDP log), or log file), open
source chromium's logging module,
"UDP log" could record multiple devices's log at a time, easy to
checkout and relocate or track. e.g: when a bugger happens, you
could track that bugger cia UDP log, do not need to go to Test
Group to checkout their testing log,
printk (using in printf kernel, every changed debug should be delete
every time, every debug every recompile, not always function that
means not output print every time),
systemtap, perf, kdump,
valgrind (memory lock(dynamic mem statistics)),
SAFECode(LLVM subproject, to solve memory lock, faster valgrind),
memory lock: not free that applied mem, access illegal mem(Segmentation
fault), use not initialized mem, ...
log: web checkout,
optimization: gprof,
test framework: DejiGnu, CppUnit,
single unit test: googletest,
others's code review: grep, Source Insight,
follow code: cscope, ctags, gtags,
grammar parser: global,
checkout static code: splint,
autocomplete: YouCompleteMe,
lint: pypep8, pylint, jslint,
compiler:
(Makefile)GNUMake,
(CMake)
compiler platform: remote Jenkins(hudson), chroot
code sync: virtual machine settings share folder with VS editing code,
"Ctrl+shift+s" syncing code, recompiling before debugging.
project: autoconf,
code style: cpplint.py, google-styleguide.google.com...,
style format: astyle, astyle.sourceforge.net/...,
document generation/organization: doxygen,
"apt-get" is being called to update the package repository;
"build-essential" package which contains the C and C++ compilers and standard
libraries for Linux;
"autoconf", "libtool", and "pkg-config" packages are also used for building
the C++ programs;
Also, "automake", "make".
"CMake"("cmake") is a generator tool that takes a description of the build process and
can generate a Makefile for the target platform; On the Linux, it will use GCC
and G++ to build the program.
"openssh-server"
"g++"
"gdb"
"gdbserver"
"net-tools" includes "ifconfig"
"tar"
"rsync"
"python3"
"python3-pip"
"libboost-dev"
"SSH" is to connect the remote Host, the virtual Machine, that Docker contain
-ers with the local Host Mac.
"Nginx" is a architecture of server. That need: GCC environment(gcc-g++),
lib PCRE(parser Lambda: pcre, pcre-devel), zlib(compress and decompress:
zlib, zlib-devel), OpenSSL protocol(openssl).