Why can't I double-click to install software in Linux?-linux Culture

Abstract:
1: windows only supports x86 x64 hardware platforms, while linux software usually supports x86, arm, x64, and many other hardware platforms. Therefore, the./configure source code can be used across platforms. Jun does not see 32-bit and 64-bit, the general software provides two installation versions, at least the startup is different!

2: Pass. the/configure command can add reasonable parameters without compiling unnecessary modules. In windows, even if you choose not to follow a certain module, the main program code is compiled to support loading multiple modules, the module is not loaded when it is not found.

3: For public modules that depend on other programs, in Windows, they must be in C: Windows/System32 (or set the environment variable PATH)
In linux, The./configure parameter specifies the installation path of other modules on your computer (such as libiconv ).

4: If you do not need to span so many hardware platforms, you can double-click the software to install ubuntu and Fedora systems!

5: If the tar.gz source code version is released on QQ, the installation command I want is:
Cd qq2012-beta2 // configure -- without-ad -- without-qqshow -- without-popnews -- without-qzone -- with-ipaddress -- with-showip make; make install

Body:
According to the recent cnbeta comments, one of my friends repeatedly said, "linux is the worst system in the world. You cannot install software by double-clicking." many comments have been made on this comment, but there are very few true ones. In summary, the majority of cber, especially those who like to comment, have little understanding of Linux, and many of them are compared with Windows (of course, I admit that there are many experts here ), so I want to explain here why Linux cannot install software by double-clicking.

First of all, many Linux distributions, such as Ubuntu, CentOS, and Fedora, have long been able to double-click to install their own installation packages, such as deb and rpm.
Why? Now, the major program software is still installing a tar.gz and./configure/make install tedious command?

In fact, the biggest difference between Linux and Windows is that most of the software in Linux is open-source, which only provides source code download and does not provide compiled binary downloads. Therefore, the above three commands are actually to compile a software, and then copy it to the corresponding directory of the system, not to be accurate.

But why do we need these three commands instead of automatically executing the preceding three commands after double-clicking?

It mainly involves cross-platform and custom considerations. We know that there are many Linux releases and many Linux platforms, from x86 to arm, solaris, and so on. the/configure script supports many compilation parameters so that the software supports almost all hardware platforms.

In Windows, because almost all software is compiled and released in binary packaging, the software is either a 32-bit version or a 64-bit version. Fortunately, Windows currently only supports x86 and x64 hardware platforms. If Windows supports other platforms in the future, there will be more differences between our daily software versions. In Linux, you only need to publish a source code package, which hardware platform you are using has been compiled by yourself.

Even those who have learned software engineering a little bit know that software development is modular. the/configure command can add reasonable parameters without compiling unnecessary modules, such as the most commonly used php, with dozens of modules. Which one should be compiled normally.

Some people may say that Windows installation software can also be installed selectively. I do not need to install modules that I don't need.

In fact, for Windows, even if you do not install the module you choose, it is only that this module is not installed. The Module Access of the main program is compiled, but it can automatically determine that this module does not exist and will not be loaded.

In addition, many Linux modules are common. This is the dependency. For example, to use the iconv function of php, you need to install the libiconv library in the system. The installation paths of the libiconv library on different computers are different, so you need to pass. the/configure parameter specifies the libiconv installation path on your computer so that php can be compiled successfully. Windows must be in C: WindowsSystem32. customization is too bad.

Let's talk about the make command. The make command is to compile the source code into a binary file based on the configuration parameters of./configure, and put it in the current directory instead of copying it to the system directory. Why can't I put make and make install together?

Sometimes, we only need a dynamic link library in a software package, rather than the entire program. For example, we often need spawn-fcgi in lighttpd, so we don't have to install lighttpd, but compile it directly to get the spawn-fcgi in it.

Finally, make install copies the compiled binary files, link libraries, configuration files, and other files to various directories of the system. The installation is complete.

Speaking of this, some netizens may not have read the text above. They just need to know how to use it if they don't need to know so much about it. Yes. If you don't need Windows, you can try Ubuntu or Fedora. You can double-click the installation program for both releases. However, the disadvantage is that deep customization is not possible.

If the tar.gz source code version is released on QQ, the installation command I want is:

Cd qq2012-beta2 // configure -- without-ad -- without-qqshow -- without-popnews -- without-qzone -- with-ipaddress -- with-showip make; make install