Software builds are a core part of the software development lifecycle. Builds are where source files are converted into executable software products.
The classic tool for automating software builds is the Unix "make" program, which was invented around 1974 to reduce software build times for C programs on the new Unix operating system. Make programs decide whether source files need rebuilding based by using text Makefiles that contain dependency trees, and by comparing timestamps on source files and object files.
Beginning in the 1980s, integrated development environments (IDEs) began to do software builds of IDE "project files" themselves, reducing the need for people to write traditional makefiles for software builds of IDE projects.
However, makefiles are still very useful for software build operations that go beyond individual IDE projects, and so many modern IDEs generate external makefiles for IDE projects.
Around the turn of the millennium (2000), a new generation of build manager tools began to emerge, as people tried to solve the still-present problem of software builds using interactive GUI interface technologies developed in the 1990s. The trend of developing software on personal computers under Windows played a big part in motivating the new generation of GUI-interface tools, since the Windows world was a GUI world, unlike previous decades of Unix systems.
This new generation of GUI-oriented tools helped people to interactively script and manage build processes using a GUI interface, rather than using traditional makefiles. Instead of coding lines in a troublesome makefile syntax, people could "code" the steps of their build process using more friendly GUI dialog boxes, and could click to execute various steps in the build process. Especially for individual programmers, these tools proved quite useful.
Some of the essential goals of a build manager tool are to provide a nicer GUI interface for people, to provide a more friendly way of specifying build process steps, and to provide additional functionality and interactive build management convenience through the interface.
But build manager tools do not typically address the deeper problems of larger-scale builds that involve multiple languages, larger numbers of files and platforms and components, pervasive parallelism and distribution over remote machines, or other complex aspects of scalable builds. For this kind of problem, build management _systems_ are required.
Build management systems try to go beyond simple build scripting tools, by providing more general software build models that involve some level of abstraction away from the details of software builds. For example, a build system might provide a layer of abstraction above the traditional makefile layer, leading to platform-independent makefiles (or equivalents). Or a build management system might abstract builds even further, by calculating and executing multiplatform build sequences for itself, in a totally automated way, thereby reducing the need for people to be involved in scripting their builds at all.
This category is for research papers, tutorials, and other documentation related to software builds. Products for performing software builds should be submitted to one of the other software build categories.
Build management systems improve productivity for software builds by providing more functionality than build manager tools. Some of the main characteristics of larger build management systems are that they provide broader and deeper support for abstraction, scope, and automation than do the smaller, interactive build manager tools.
Build manager tools improve productivity for software builds by providing more functionality than make programs provide. Build manager tools usually include a GUI or web interface, a more convenient build process scripting language, and means for executing all or part of scripted build processes, and for logging and reporting build results. Some of the main characteristics of build manager tools are a GUI interface, increased interaction with builds, a build scripting language or interface, and built-in capabilities for build execution, logging, and reporting.
Makefile generators reduce the human costs of software builds by automating some of the tedious and error-prone work of creating makefiles. The makefile generation problem is very difficult to solve, so most makefile generators are simple. They depend upon human labor for the difficult parts of makefile construction.
Simple generators produce dependency files by scanning source files for include file statements. Intermediate generators produce whole makefiles by expanding manually-constructed input files. Advanced generators produce whole makefiles from scratch, with no manually-constructed input files or human labor required.
Make programs help to manage the complexity of building
software programs from many source file and library
components. Make programs read Makefiles, which are
description files that specify the parts and processes that
should be used to build a software product.
Make programs have been the dominant programs for doing
software builds for more than 25 years. The first make
program was created in the mid-1970s. In the 1980s, make
was extended with more advanced features. In the 1990s,
several new make replacement programs were created to offer
new description languages for makefiles.
This category lists research papers and websites that provide useful information related to make programs and the overall software build problem.
This category is for research papers, tutorials, and other documentation related to software builds. Products for performing software builds should be submitted to one of the other software build categories.