namcap

Namcap is a tool to check binary packages and source PKGBUILDs for common packaging mistakes, which can also be automatically enabled. It was created by Jason Chu.

Changes: The NEWS file in the git repository contains the changes from the previous version concisely.

Development Branch: https://gitlab.archlinux.org/pacman/namcap

To report a bug or a feature request for namcap, file a bug in the Packages:Extra section of the Arch Linux bugtracker and the set the importance accordingly. If you are reporting a bug, please give concrete examples of packages where the problem occurs and remember to insert the version number.

If you have a patch (fixing a bug or a new namcap module), then you can send it to the arch-projects mailing list. Namcap development is managed with git, so git-formatted patches are preferred.

Installation

Install the namcap package.

Usage

To run namcap on a file, where filename is PKGBUILD or the name of a binary pkg.tar.zst:

$ namcap filename

If you want to see extra informational messages, then invoke namcap with the -i flag:

$ namcap -i filename

For more info on usage, see the man page namcap(1).

Namcap uses a system of tags to classify the output. Currently, tags are of three types, errors (denoted by E), warnings (denoted by W) and informational (denoted by I). An error is important and should be fixed immediately; mostly they relate to insufficient security, missing licenses or permission problems.

Normally namcap prints a human-readable explanation (sometimes with suggestions on how to fix the problem). If you want output which can be easily parsed by a program, then pass the -m (machine-readable) flag to namcap (this feature is currently in the development branch).

The tag file /usr/share/namcap/namcap-tags[1] consists of lines specifying the human-readable form of the hyphenated tags used in the namcap code. A line beginning with a # is treated as a comment:

# The comment
machine-parsable-tag %s :: The human-readable description for the tag %s

Note:

:: (double colon) separates the hyphenated machine-readable tag from the human-readable description.
%s is a format specifier, see #Making a namcap module for more information.

Making a namcap module

This section documents the innards of namcap and specifies how to create a new namcap module.

The main namcap program namcap.py takes as parameters the filename of a package or a PKGBUILD and makes a pkginfo object, which it passes to a list of rules defined in __tarball__ and __pkgbuild__.

__tarball__ defines the rules which process binary packages.
__pkgbuild__ defines the rules which process PKGBUILDs

Once your module is finalized, remember to add it to the appropriate array (__tarball__ or __pkgbuild__) defined in Namcap/__init__.py

A sample namcap module is like this:

namcap/url.py

  import pacman

  class package:
  	def short_name(self):
		return "url"
	def long_name(self):
		return "Verifies url is included in a PKGBUILD"
	def prereq(self):
		return ""
	def analyze(self, pkginfo, tar):
		ret = [[],[],[]]
		if not hasattr(pkginfo, 'url'):
			ret[0].append(("missing-url", ()))
		return ret
	def type(self):
		return "pkgbuild"

Mostly, the code is self-explanatory. The following are the list of the methods that each namcap module must have:

short_name(self) Returns a string containing a short name of the module. Usually, this is the same as the basename of the module file.
long_name(self) Returns a string containing a concise description of the module. This description is used when listing all the rules using namcap -r list.
prereq(self) Return a string containing the prerequisites needed for the module to operate properly. Usually "" for modules processing PKGBUILDs and "tar" for modules processing package files. "extract" should be specified if the package contents should be extracted to a temporary directory before further processing.
analyze(self, pkginfo, tar) Should return a list comprising in turn of three lists: of error tags, warning tags and information tags respectively. Each member of these tag lists should be a tuple consisting of two components: the short, hyphenated form of the tag with the appropriate format specifiers (%s, etc.) The first word of this string must be the tag name. The human readable form of the tag should be put in the tag file. The format of the tag file is described above; and the parameters which should replace the format specifier tokens in the final output.
type(self) "pkgbuild" for a module processing PKGBUILDs, "tarball" for a module processing a binary package file.