_
_ _ ___ _____ ___| |_ ___
| | | | | .'| '_| -_|
|___|_|_|_|_|_|__,|_,_|___|
unmake
is a makefile linter emphasizing portability. We target the POSIX make standard.
With unmake
, you can develop practical, portable makefile
scripts, such as for provisioning dev environments.
No need for more heavyweight tools like Ansible, Docker, Lua, Python, or WSL! Just good ol' make.
$ cd fixtures/parse-valid
$ unmake .
warning: ./Makefile: MAKEFILE_PRECEDENCE: lowercase Makefile to makefile for launch speed
warning: ./boilerplate-ats.mk:4: SIMPLIFY_AT: replace individual at (@) signs with .SILENT target declaration(s)
warning: ./missing-posix.mk:1: STRICT_POSIX: lead makefiles with the ".POSIX:" compliance marker, or else rename include files like *.include.mk
...
See unmake -h
for more options.
See makefile for a live example of a portable dev environment provisioning script for this Rust project.
unmake
applies a stricter reading of POSIX syntax than make -n
. Whereas make -n
may skip inactive sections depending on control flow, unmake
scans each line. For example, make -n
may only check instructions specific to building the default task.
In fact, the two checks complement each other. make -n
checks for dry-run runtime issues. unmake
checks for syntactic portability issues.
unmake
automatically recurses over directories.
When recursing over directories, unmake
skips symlinks.
unmake
's linter rules skip many implementation-specific files such as GNUmakefile
.
unmake
skips many machine-generated makefiles. For example, makefiles produced by autotools; Perl; and cmake when using the Unix Makefile generator (both in-source builds and out-of-source builds).
unmake
skips any third party makefiles housed in subdirectories like .git
, node_modules
, or vendor
.
To investigate makefiles in more detail, see the --debug
or --inspect
command line options for unmake
.
unmake
can identify low level makefile quirks, such as invalid syntax.
See SYNTAX.md for more information.
-n
/ --dry-run
performs passthrough dry run validation with external make implementation tools, e.g. bmake -nf
, gmake -nf
, make -nf
, etc.
Per POSIX, rule commands prefixed with plus (+
) may continue to execute in dry run mode.
A few benefits of unmake dry run checks:
find
\ xargs
snippetsThe unmake dry run option aggressively assumes that most makefiles are buildable, top level project configurations, as opposed to make include files named like include.mk
, *.include.mk
, etc.
-l
/ --list
emits paths of any matching makefiles that unmake finds in the given file paths.
This is useful for feeding large make projects into external linters. Unfortunately, many linters are poorly designed, lacking directory recursion and automatic file type identification. As a stopgap, unmake can perform these duties, exporting a subset of makefiles within a large project, through xargs
, to an external linter.
Like dry run, the list option automatically skips common machine generated makefiles.
When piping unmake makefile lists through xargs, we recommend adding a --print0
flag to unmake, and adding a -0
flag to xargs. This informs both programs to transfer data in null delimited form, as a precaution against errors related to any spaces in file paths.
unmake
can identify higher level portability recommendations for makefiles.
See WARNINGS.md for more information.
https://crates.io/crates/unmake
https://docs.rs/unmake/latest/unmake/
https://github.com/mcandre/unmake/releases
$ cargo install --force --path .
(None)
For more details on developing unmake itself, see DEVELOPMENT.md.
BSD-2-Clause
Some projects using unmake
to safeguard their makefiles:
Note that the GNU and BSD implementations of make
support -j [4]
options for accelerating workflows using parallelism. Strict POSIX compliance helps to ensure that makefiles are portable, able to run with GNU, BSD, and other make implementations.