We build and release software by massively consuming and producing software packages such as NPMs, RPMs, Rubygems, etc.
Each package manager, platform, type or ecosystem has its own conventions and protocols to identify, locate and provision software packages.
When tools, APIs and databases process or store multiple package types, it is difficult to reference the same software package across tools in a uniform way.
For example, these tools, specifications and API use relatively similar approaches to identify and locate software packages, each with subtle differences in syntax, naming and conventions:
- Grafeas uses a scheme, namespace, name and version in a URL-like string. https://github.com/Grafeas/Grafeas
- Here.com OSRK uses a package manager, name and version field and a colon- separated URL-like string https://github.com/heremaps/oss-review-toolkit
- JFrog XRay uses a scheme, namespace, name and version in a URL-like string https://www.jfrog.com/confluence/display/XRAY/Xray+REST+API#XrayRESTAPI-ComponentIdentifiers
- Libraries.io uses a platform, name and version https://libraries.io/
- OpenShift fabric8 analytics uses ecosystem, name and version https://github.com/fabric8-analytics/
- ScanCode and AboutCode.org use a type, name and version https://github.com/nexB/scancode-toolkit
- SPDX has an appendix for external repository references and uses a type and a locator with a type-specific syntax for component separators in a URL-like string https://github.com/spdx/spdx-spec/blob/master/chapters/3-package-information.md
- versioneye uses a type, name and version https://github.com/versioneye/
- Sonatype Lifecycle uses a format id followed by format specific coordinates. https://help.sonatype.com/display/NXIQ/Component+Details+API+-+v2
A purl or package URL is an attempt to standardize existing approaches to reliably identify and locate software packages.
A purl is a URL string used to identify and locate a software package in a mostly universal and uniform way across programing languages, package managers, packaging conventions, tools, APIs and databases.
Such a package URL is useful to reliably reference the same software package using a simple and expressive syntax and conventions based on familiar URLs.
purl stands for package URL.
A purl is a URL composed of six components:
type:namespace/name@version?qualifiers#subpath
Components are separated by a specific character for unambiguous parsing.
The defintion for each components is:
- type: the package "type" or package "protocol" such as maven, npm, nuget, gem, pypi, etc. Required.
- namespace: some name prefix such as a Maven groupid, a Docker image owner, a GitHub user or organization. Optional and type-specific.
- name: the name of the package. Required.
- version: the version of the package. Optional.
- qualifiers: extra qualifying data for a package such as an OS, architecture, a distro, etc. Optional and type-specific.
- subpath: extra subpath within a package, relative to the package root. Optional.
Components are designed such that they for a hierarchy from the most significant on the left to the least significant components on the right.
A purl must NOT contain a URL Authority i.e. there is no support for username, password, host and port components. A namespace segment may sometimes look like a host but its interpretation is specific to a type.
bitbucket:birkenfeld/pygments-main@244fd47e07d1014f0aed9c deb:debian/curl@7.50.3-1?arch=i386&distro=jessie docker:cassandra@sha256:244fd47e07d1004f0aed9c docker:gcr.io/customer/dockerimage@sha256:244fd47e07d1004f0aed9c gem:jruby-launcher@1.1.2?platform=java gem:ruby-advisory-db-check@0.12.4 github:package-url/purl-spec@244fd47e07d1004f0aed9c golang:google.golang.org/genproto#googleapis/api/annotations maven:org.apache.xmlgraphics/batik-anim@1.9.1?packaging=sources maven:org.apache.xmlgraphics/batik-anim@1.9.1?repository_url=repo.spring.io/release npm:%40angular/animation@12.3.1 npm:foobar@12.3.1 nuget:EnterpriseLibrary.Common@6.0.1304 pypi:django@1.11.1 rpm:fedora/curl@7.50.3-1.fc25?arch=i386&distro=fedora-25 rpm:opensuse/curl@7.56.1-1.1.?arch=i386&distro=opensuse-tumbleweed
(NB: some checksums are truncated for brevity)
- A purl is a valid URL and URI that conforms to the URL definitions or specifications at:
- The purl components are mapped to these URL components:
- purl type: this is a URL scheme
- purl namespace, name and version components: these are collectively mapped to a URL path
- purl qualifiers: this maps to a URL query
- purl subpath: this is a URL fragment
- In a purl there is no support for a URL Authority (e.g. NO username, password, host and port components).
- Special URL schemes as defined in https://url.spec.whatwg.org/ such as file://, https://, http:// and ftp:// are NOT valid purl types. They may be used to reference URLs in separate attributes outside of a purl or in a purl qualifier.
- Version control system (VCS) URLs such git://, svn://, hg:// or as defined in Python pip or SPDX download locations are NOT valid purl types. They are a closely related, compact and uniform way to reference vcs URLs. They may be used as references in separate attributes outside of a purl or in a purl qualifier.
A purl string is an ASCII URL string composed of six components.
Some components are allowed to use other characters beyond ASCII: these components must then be UTF-8-encoded strings and percent-encoded as defined in the "Character encoding" section.
The rules for each component are:
type:
The package type is composed only of ASCII letters and numbers, '.', '+' and '-' (period, plus, and dash)
The type cannot start with a number
The type cannot contains spaces
The type must NOT be percent-encoded
The type is case insensitive. The canonical form is lowercase
Since a purl never contains a URL Authority, its type must not be suffixed with double slash as in 'docker://' and should use instead 'docker:'. Otherwise this would be an invalid URI per rfc3986 at https://tools.ietf.org/html/rfc3986#section-3.3:
If a URI does not contain an authority component, then the path cannot begin with two slash characters ("//").
While it is acceptable to use such '://' suffix, its is not significant and not needed for unambiguous parsing even if it makes a purl look like a familiar web URL. In its canonical form, a purl must NOT use such '://' type suffix.
purl parsers must accept URLs with such '://' and must ignore the '//'.
purl builders must not create invalid URLs with such double slash '//'.
The type is followed by a ':' separator
For example these two purls are strictly equivalent and the first is in canonical form. The second purl with a '//' is an acceptable purl but is an invalid URI/URL per rfc3986:
gem:ruby-advisory-db-check@0.12.4 gem://ruby-advisory-db-check@0.12.4
namespace:
- The optional namespace contains zero or more segments, separated by slash '/'
- Leading and trailing slashes '/' are not significant and should be stripped in the canonical form. They are not part of the namespace
- Each namespace segment must be a percent-encoded string
- When percent-decoded, a segment:
- must not contain a '/'
- must not be empty
- A URL host or Authority must NOT be used as a namespace. Use instead a repository_url qualifier. Note however that for some types, the namespace may look like a host.
name:
- The name is prefixed by a '/' separator when the namespace is not empty
- This '/' is not part of the name
- A name must be a percent-encoded string
version:
- The version is prefixed by a '@' separator when not empty
- This '@' is not part of the version
- A version must be a percent-encoded string
- A version is a plain and opaque string. Some package types use versioning conventions such as semver for NPMs or nevra conventions for RPMS. A type may define a procedure to compare and sort versions, but there is no reliable and uniform way to do such comparison consistently.
qualifiers:
- The qualifiers string is prefixed by a '?' separator when not empty
- This '?' is not part of the qualifiers
- This is a query string composed of zero or more key=value pairs each separated by a '&' ampersand. A key and value are separated by the equal '=' character
- These '&' are not part of the key=value pairs.
- key must be unique within the keys of the qualifiers string
- value cannot be an empty string: a key=value pair with an empty value is the same as no key/value at all for this key
- For each pair of key = value:
- The key must be composed only of ASCII letters and numbers, '.', '-' and '_' (period, dash and underscore)
- A key cannot start with a number
- A key must NOT be percent-encoded
- A key is case insensitive. The canonical form is lowercase
- A key cannot contains spaces
- A value must be must be a percent-encoded string
- The '=' separator is neither part of the key nor of the value
subpath:
- The subpath string is prefixed by a '#' separator when not empty
- This '#' is not part of the subpath
- The subpath contains zero or more segments, separated by slash '/'
- Leading and trailing slashes '/' are not significant and should be stripped in the canonical form
- Each subpath segment must be a percent-encoded string
- When percent-decoded, a segment:
- must not contain a '/'
- must not be any of '..' or '.'
- must not be empty
- The subpath must be interpreted as relative to the root of the package
For clarity and simplicity a purl is always an ASCII string. To ensure that there is no ambiguity when parsing a purl, separator characters and non-ASCII characters must be UTF-encoded and then percent-encoded as defined at:
https://en.wikipedia.org/wiki/Percent-encoding
Use these rules for percent-encoding and decoding purl components:
- the type must NOT be encoded and must NOT contain separators
- the '#', '?', '@' and ':' characters must NOT be encoded when used as separators. They may need to be encoded elsewhere
- the ':' type separator does not need to and must NOT be encoded. It is unambiguous unencoded everywhere
- the '/' used as namespace/name and subpath segments separator does not need to and must NOT be percent-encoded. It is unambiguous unencoded everywhere
- the '@' version separator must be encoded as %40 elsewhere
- the '?' qualifiers separator must be encoded as %3F elsewhere
- the '=' qualifiers key/value separator must NOT be encoded
- the '#' subpath separator must be encoded as %23 elsewhere
- All non-ASCII characters must be encoded as UTF-8 and then percent-encoded
It is OK to percent-encode purl components otherwise except for the type. Parsers and builders must always percent-decode and percent-encode purl components and component segments as explained in the "How to parse" and "How to build" sections.
Building a purl ASCII string works from left to right, from type to subpath.
Note: some extra type-specific normalizations are required. See the "Known types section" for details.
To build a purl string from its components:
- Start a purl string with the type as a lowercase ASCII string
- Append ':' to the purl
- If the namespace is not empty:
- Strip the namespace from leading and trailing '/'
- Split on '/' as segments
- Apply type-specific normalization to each segment if needed
- UTF-8-encode each segment if needed in your programming language
- Percent-encode each segment
- Join the segments with '/'
- Append this to the purl
- Append '/' to the purl
- Strip the name from leading and trailing '/'
- Apply type-specific normalization to the name if needed
- UTF-8-encode the name if needed in your programming language
- Append the percent-encoded name to the purl
- If the namespace is empty:
- Apply type-specific normalization to the name if needed
- UTF-8-encode the name if needed in your programming language
- Append the percent-encoded name to the purl
- If the version is not empty:
- Append '@' to the purl
- UTF-8-encode the version if needed in your programming language
- Append the percent-encoded version to the purl
- If the qualifiers are not empty and not composed only of key/value pairs
where the value is empty:
- Append '?' to the purl
- Build a list from all key/value pair:
- discard any pair where the value is empty.
- UTF-8-encode each value if needed in your programming language
- If the key is checksums and this is a list of checksums join this list with a ',' to create this qualifier value
- create a string by joining the lowercased key, the equal '=' sign and the percent-encoded value to create a qualifier
- sort this list of qualifier strings lexicographically
- join this list of qualifier strings with a '&' ampersand
- Append this string to the purl
- If the subpath is not empty and not composed only of empty, '.' and '..'
segments:
- Append '#' to the purl
- Strip the subpath from leading and trailing '/'
- Split this on '/' as segments
- Discard empty, '.' and '..' segments
- Percent-encode each segment
- UTF-8-encode each segment if needed in your programming language
- Join the segments with '/'
- Append this to the purl
Parsing a purl ASCII string into its components works from right to left, from subpath to type.
Note: some extra type-specific normalizations are required. See the "Known types section" for details.
To parse a purl string in its components:
- Split the purl string once from right on '#'
- The left side is the remainder
- Strip the right side from leading and trailing '/'
- Split this on '/'
- Discard any empty string segment from that split
- Discard any '.' or '..' segment from that split
- Percent-decode each segment
- UTF-8-decode each segment if needed in your programming language
- Join segments back with a '/'
- This is the subpath
- Split the remainder once from right on '?'
- The left side is the remainder
- The right side is the qualifiers string
- Split the qualifiers on '&'. Each part is a key=value pair
- For each pair, split the key=value once from left on '=':
- The key is the lowercase left side
- The value is the percent-decoded right side
- UTF-8-decode the value if needed in your programming language
- Discard any key/value pairs where the value is empty
- If the key is checksums, split the value on ',' to create a list of checksums
- This list of key/value is the qualifiers object
- Split the remainder once from left on ':'
- The left side lowercased is the type
- The right side is the remainder
- Strip the remainder from leading and trailing '/'
- Split this once from right on '/'
- The left side is the remainder
- Percent-decode the right side
- UTF-8-decode the name if needed in your programming language
- Apply type-specific normalization to the name if needed
- This is the name
- Split the remainder on '/'
- Discard any empty segment from that split
- Percent-decode each segment
- UTF-8-decode the each segment if needed in your programming language
- Apply type-specific normalization to each segment if needed
- Join segments back with a '/'
- This is the namespace
These are known purl package type definitions. More should be added. See candidate list further down.
bitbucket for Bitbucket-based packages:
The default repository is bitbucket.org
The namespace is the user or organization. It is not case sensitive and must be lowercased.
The name is the repository name. It is not case sensitive and must be lowercased.
The version is a commit or tag
Examples:
bitbucket:birkenfeld/pygments-main@244fd47e07d1014f0aed9c
composer for Composer PHP packages:
The default repository is packagist.org
The namespace is the vendor.
Note: private, local packages may have no name. In this casse you cannot create a purl for these.
Examples:
composer:laravel/laravel@5.5.0
deb for Debian, Debian derivatives and Ubuntu packages:
There is no default package repository: this should be implied either from the distro qualifiers key or using a base url as a repository_url qualifiers key
The namespace is the "vendor" name such as "debian" or "ubuntu". It is not case sensitive and must be lowercased.
The name is not case sensitive and must be lowercased.
The version is is the package version.
arch is the qualifiers key for a package architecture
Examples:
deb:debian/curl@7.50.3-1?arch=i386&distro=jessie deb:debian/dpkg@1.19.0.4?arch=amd64&distro=stretch deb:ubuntu/dpkg@1.19.0.4?arch=amd64
docker for Docker images
The default repository is hub.docker.com
The namespace is the registry/user/organization if present
The version should be the image id sha256 or a tag. Since tags can be moved, a sha256 image id is preferred.
Examples:
docker:cassandra@latest docker:smartentry/debian@dc437cc87d10 docker:gcr.io/customer/dockerimage@sha256:244fd47e07d10
gem for Rubygems:
The default repository is rubygems.org
The platform qualifiers key is used to specify an alternative platform such as java for JRuby. The implied default is ruby for Ruby MRI.
Examples:
gem:ruby-advisory-db-check@0.12.4 gem:jruby-launcher@1.1.2?platform=java
generic for plain, generic packages that do not fit anywhere else such as for "upstream -from-distro" packages. In particular this is handy for a plain version control repository such as a bare git repo.
Their is no default repository. A download_url and checksum may be provided in qualifiers or as separate attributes outside of a purl for proper identification and location.
When possible another or a new purl type should be used instead of using the generic type and eventually contributed back to this specification
as for other type, the name component is mandatory. In the worst case it can be a file or directory name.
Examples (truncated for brevity):
generic:openssl@1.1.10g generic:openssl@1.1.10g?download_url=https://openssl.org/source/openssl-1.1.0g.tar.gz&checksum=sha256:de4d501267da3931090 generic:bitwarderl?vcs_url=https://git.fsfe.org/dxtr/bitwarderl@cc55108da32042a0e385bd8e
github for Github-based packages:
The default repository is github.com
The namespace is the user or organization. It is not case sensitive and must be lowercased.
The name is the repository name. It is not case sensitive and must be lowercased.
The version is a commit or tag
Examples:
github:package-url/purl-spec@244fd47e07d1004f0aed9c github:package-url/purl-spec@244fd47e07d1004f0aed9c#everybody/loves/dogs
golang for Go packages
There is no default package repository: this is implied in the namespace using the go get command conventions
The namespace and name must be lowercased.
The subpath is used to point to a subpath inside a package
The version is often empty when a commit is not specified and should be the commit in most cases when available.
Examples:
golang:github.com/gorilla/context@234fd47e07d1004f0aed9c golang:google.golang.org/genproto#googleapis/api/annotations golang:github.com/gorilla/context@234fd47e07d1004f0aed9c#api
maven for Maven JARs and related artifacts
The default repository is repo1.maven.org
The group id is the namespace and the artifact id is the name
Known qualifiers keys are: classifier and type as defined in the POM documentation. Note that Maven uses a concept / coordinate called packaging which does not map directly 1:1 to a file extension. In this use case, we need to construct a link to one of many possible artifacts. Maven itself uses type in a dependency declaration when needed to disambiguate between them.
Examples:
maven:org.apache.xmlgraphics/batik-anim@1.9.1 maven:org.apache.xmlgraphics/batik-anim@1.9.1?type=pom maven:org.apache.xmlgraphics/batik-anim@1.9.1?classifier=sources maven:org.apache.xmlgraphics/batik-anim@1.9.1?type=zip&classifier=dist maven:net.sf.jacob-projec/jacob@1.14.3?classifier=x86&type=dll maven:net.sf.jacob-projec/jacob@1.14.3?classifier=x64&type=dll
npm for Node NPM packages:
The default repository is registry.npmjs.org
The namespace is used for the scope of a scoped NPM package.
Per the package.json spec, new package "must not have uppercase letters in the name", therefore the must be lowercased.
Examples:
npm:foobar@12.3.1 npm:%40angular/animation@12.3.1 npm:mypackage@12.4.5?vcs_url=git://host.com/path/to/repo.git@4345abcd34343
nuget for NuGet .NET packages:
The default repository is nuget.org
There is no namespace per se even if the common convention is to use dot-separated package names where the first segment is namespace-like. TBD: should we split the first segment as a namespace?
Examples:
nuget:EnterpriseLibrary.Common@6.0.1304
pypi for Python packages:
The default repository is pypi.python.org
PyPi treats '-' and '_' as the same character and is not case sensitive. Therefore a Pypi package name must be lowercased and underscore '_' replaced with a dash '-'
TBD: we could specify a format qualifiers key to specify a package format with values of egg, wheel , sdist, exe or may be a file extension?
TBD: we could specify a markers qualifiers key to specify PEP 508 environment markers but this is extra complexity. See https://www.python.org/dev/peps/pep-0508/
Examples:
pypi:django@1.11.1 pypi:django-allauth@12.23
rpm for RPMs:
There is no default package repository: this should be implied either from the distro qualifiers key or using a repository base url as a repository_url qualifiers key
the namespace is the vendor such as fedora or opensuse It is not case sensitive and must be lowercased.
the name is the RPM name and is case sensitive.
the version is the combined epoch (if not 0), version and release of an RPM.
arch is the qualifiers key for a package architecture
Examples:
rpm:fedora/curl@7.50.3-1.fc25?arch=i386&distro=fedora-25 rpm:opensuse/curl@7.56.1-1.1.?arch=i386&distro=opensuse-tumbleweed
- alpine for Alpine Linux apk packages:
- apache for Apache projects packages:
- android for Android apk packages:
- arch for Arch Linux packages:
- atom for Atom packages:
- bower for Bower JavaScript packages:
- brew for Homebrew packages:
- buildroot for Buildroot packages
- cargo for Rust packages:
- carthage for Cocoapods Cocoa packages:
- chef for Chef packages:
- clojars for Clojure packages:
- cocoapods for Cocoapods iOS packages:
- conan for Conan C/C++ packages:
- coreos for CoreOS packages:
- cpan for CPAN Perl packages:
- cran for CRAN R packages:
- ctan for CTAN TeX packages:
- crystal for Crystal Shards packages:
- drupal for Drupal packages:
- dtype for DefinitelyTyped TypeScript type definitions:
- dub for D packages:
- elm for Elm packages:
- eclipse for Eclipse projects packages:
- gitea for Gitea-based packages:
- gitlab for Gitlab-based packages:
- guix for Guix packages:
- hackage for Haskell packages:
- haxe for Haxe packages:
- hex for Erlang and Elixir packages
- julia for Julia packages:
- lua for LuaRocks packages:
- melpa for Emacs packages
- meteor for Meteor JavaScript packages:
- nim for Nim packages:
- nix for Nixos packages:
- opam for OCaml packages:
- openwrt for OpenWRT packages:
- osgi for OSGi bundle packages:
- p2 for Eclipse p2 packages:
- pear for Pear PHP packages:
- pecl for PECL PHP packages:
- perl6 for Perl 6 module packages:
- platformio for PlatformIO packages:
- ebuild for Gentoo Linux portage packages:
- pub for Dart packages:
- puppet for Puppet Forge packages:
- sourceforge for Sourceforge-based packages:
- sublime for Sublime packages:
- swift for Swift packages:
- vim for Vim scripts packages:
- wordpress for Wordpress packages:
- yocto for Yocto recipe packages
Note: Do not abuse qualifiers: it can be tempting to use many qualifier keys but their usage should be limited to the bare minimum for proper package identification to ensure that a purl stays compact and readable in most cases.
Additional, separate external attributes stored outside of a purl are the preferred mechanism to convey extra long and optional information such as a download URL, vcs URL or checksums in an API, database or web form.
With this warning, the known key and value defined here are valid for use in all package types:
repository_url is an extra URL for an alternative, non-default package repository or registry. When a package does not come from the default public package repository for its type a purl may be qualified with this extra URL. The default repository or registry of a type is documented in the "Known purl types" section.
download_url is an extra URL for a direct package web download URL to optionally qualify a purl.
vcs_url is an extra URL for a package version control system URL to optionally qualify a purl. The syntax for this URL should be as defined in Python pip or the SPDX specification. See https://github.com/spdx/spdx- spec/blob/cfa1b9d08903/chapters/3-package-information.md#37-package-download- location- TODO: incorporate the details from SPDX here.
file_name is an extra file name of a package archive.
checksum is a qualifier for one or more checksums stored as a comma-separated list. Each item in the value is in form of lowercase_algorithm:hex_encoded_lowercase_value such as sha1:ad9503c3e994a4f611a4892f2e67ac82df727086. For example (with checksums truncated for brevity)
`checksum=sha1:ad9503c3e994a4f,sha256:41bf9088b3a1e6c1ef1d`
This list is TBD!
- in JavaScript:
- in Golang: https://github.com/package-url/packageurl-go
- for .NET:
- for the JVM:
- in Perl:
- in Python: https://github.com/package-url/packageurl-python
- in Ruby:
This list is TBD!
- https://github.com/nexB/scancode-toolkit will report purl from parsed package manifests using https://github.com/package-url/packageurl-python The code lives in the 275 branch for now.
- [OWASP Dependency-Track](https://www.owasp.org/index.php/OWASP_Dependency_Track_Project): Software Composition Analysis (SCA) platform
- [CycloneDX](https://github.com/CycloneDX): A lightweight software bill-of-material (BOM) specification
To support the language-neutral testing of purl implementations, a test suite is provided as JSON document named test-suite-data.json. This JSON document contains an array of objects. Each object represents a test with these key/value pairs some of which may not be normalized:
- purl: a purl string.
- canonical: the same purl string in canonical, normalized form
- type: the type corresponding to this purl.
- namespace: the namespace corresponding to this purl.
- name: the name corresponding to this purl.
- version: the version corresponding to this purl.
- qualifiers: the qualifiers corresponding to this purl as an object of {key: value} qualifier pairs.
- subpath: the subpath corresponding to this purl.
- is_invalid: a boolean flag set to true if the test should report an error
To test purl parsing and building, a tool can use this test suite and for every listed test object, run these tests:
- parsing the test canonical purl then re-building a purl from these parsed components should return the test canonical purl
- parsing the test purl should return the components parsed from the test canonical purl
- parsing the test purl then re-building a purl from these parsed components should return the test canonical purl
- building a purl from the test components should return the test canonical purl
This document is dedicated to the public domain