Attributes¶
Package reference¶
Recipe attributes that can define the main pkg/version@user/channel
package reference.
name¶
The name of the package. A valid name is all lowercase and has:
A minimum of 2 and a maximum of 101 characters (though shorter names are recommended).
- Matches the following regex
^[a-z0-9_][a-z0-9_+.-]{1,100}$
: so starts with alphanumeric or_
, then from 1 to 100 characters between alphanumeric,
_
,+
,.
or-
.
- Matches the following regex
- The name is only necessary for
export
-ing the recipe into the local cache (export
,export-pkg
and
create
commands), if they are not defined in the command line with--name=<pkgname>
.
version¶
The version of the package. A valid version follows the same rules than the name
attribute.
In case the version follows semantic versioning in the form X.Y.Z-pre1+build2
, that value might be used for requiring
this package through version ranges instead of exact versions.
The version is only strictly necessary for export
-ing the recipe into the local cache (export
, export-pkg
and create
commands), if they are not defined in the command line with --version=<pkgversion>
The version
can be dynamically defined in the command line, and also programmatically in the recipe with the
set_version() method.
user¶
A valid string for the user
field follows the same rules than the name
attribute.
This is an optional attribute. It can be used to identify your own packages with pkg/version@user/channel
,
where user
could be the name of your team, org or company. ConanCenter recipes don’t have user/channel
,
so they are in the form of pkg/version
only. You can also name your packages without user and channel, or using
only the user as pkg/version@user
.
The user can be specified in the command line with --user=<myuser>
channel¶
A valid string for the channel
field follows the same rules than the name
attribute.
This is an optional attribute. It is sometimes used to identify a maturity of the package (“stable”, “testing”…),
but in general this is not necessary, and the maturity of packages is better managed by putting them in different
server repositories.
The user can be specified in the command line with --channel=<mychannel>
Metadata¶
Optional metadata, like license, description, author, etc. Not necessary for most cases, but can be useful to have.
description¶
This is an optional, but recommended text field, containing the description of the package, and any information that might be useful for the consumers. The first line might be used as a short description of the package.
class HelloConan(ConanFile):
name = "hello"
version = "0.1"
description = """This is a Hello World library.
A fully featured, portable, C++ library to say Hello World in the stdout,
with incredible iostreams performance"""
license¶
License of the target source code and binaries, i.e. the code
that is being packaged, not the conanfile.py
itself.
Can contain several, comma separated licenses. It is a text string, so it can
contain any text, but it is strongly recommended that recipes of Open Source projects use
SPDX identifiers from the SPDX license list
This will help people wanting to automate license compatibility checks, like consumers of your package, or you if your package has Open-Source dependencies.
class Pkg(ConanFile):
license = "MIT"
topics¶
Tags to group related packages together and describe what the code is about. Used as a search filter in ConanCenter. Optional attribute. It should be a tuple of strings.
class ProtocInstallerConan(ConanFile):
name = "protoc_installer"
version = "0.1"
topics = ("protocol-buffers", "protocol-compiler", "serialization", "rpc")
homepage¶
The home web page of the library being packaged.
Used to link the recipe to further explanations of the library itself like an overview of its features, documentation, FAQ as well as other related information.
class EigenConan(ConanFile):
name = "eigen"
version = "3.3.4"
homepage = "http://eigen.tuxfamily.org"
url¶
URL of the package repository, i.e. not necessarily of the original source code. Recommended, but not mandatory attribute.
class HelloConan(ConanFile):
name = "hello"
version = "0.1"
url = "https://github.com/conan-io/hello.git"
Requirements¶
Attribute form of the dependencies simple declarations, like requires
, tool_requires
.
For more advanced way to define requirements, use the requirements()
, build_requirements()
methods instead.
requires¶
List or tuple of strings for regular dependencies in the host context, like a library.
class MyLibConan(ConanFile):
requires = "hello/1.0", "otherlib/2.1@otheruser/testing"
You can specify version ranges, the syntax is using brackets:
class HelloConan(ConanFile):
requires = "pkg/[>1.0 <1.8]"
Accepted expressions would be:
Expression |
Versions in range |
Versions outside of range |
---|---|---|
[>=1.0 <2] |
1.0.0, 1.0.1, 1.1, 1.2.3 |
0.2, 2.0, 2.1, 3.0 |
[<3.2.1] |
0.1, 1.2, 2.4, 3.1.1 |
3.2.2 |
[>2.0] |
2.1, 2.2, 3.1, 14.2 |
1.1, 1.2, 2.0 |
If pre-releases are activated, like defining configuration core.version_ranges:resolve_prereleases=True
:
Expression |
Versions in range |
Versions outside of range |
---|---|---|
[>=1.0 <2] |
1.0.0-pre.1, 1.0.0, 1.0.1, 1.1, 1.2.3 |
0.2, 2.0-pre.1, 2.0, 2.1, 3.0 |
[<3.2.1] |
0.1, 1.2, 1.8-beta.1, 2.0-alpha.2, 2.4, 3.1.1 |
3.2.1-pre.1, 3.2.1, 3.2.2, 3.3 |
[>2.0] |
2.1-pre.1, 2.1, 2.2, 3.1, 14.2 |
1.1, 1.2, 2.0-pre.1, 2.0 |
See also
Check Range expressions version_ranges tutorial section
tool_requires¶
List or tuple of strings for dependencies. Represents a build tool like “cmake”. If there is an existing pre-compiled binary for the current package, the binaries for the tool_require won’t be retrieved. They cannot conflict.
class MyPkg(ConanFile):
tool_requires = "tool_a/0.2", "tool_b/0.2@user/testing"
This is the declarative way to add tool_requires
. Check the tool_requires()
conanfile.py method to learn a more flexible way to add them.
build_requires¶
build_requires are used in Conan 2 to provide compatibility with the Conan 1.X syntax, but their use is discouraged in Conan 2 and will be deprecated in future 2.X releases. Please use tool_requires instead of build_requires in your Conan 2 recipes.
test_requires¶
List or tuple of strings for dependencies in the host context only. Represents a test tool like “gtest”. Used when the current package is built from sources. They don’t propagate information to the downstream consumers. If there is an existing pre-compiled binary for the current package, the binaries for the test_require won’t be retrieved. They cannot conflict.
class MyPkg(ConanFile):
test_requires = "gtest/1.11.0", "other_test_tool/0.2@user/testing"
This is the declarative way to add test_requires
.
Check the test_requires() method
to learn a more flexible way to add them.
python_requires¶
This class attribute allows to define a dependency to another Conan recipe and reuse its code. Its basic syntax is:
from conan import ConanFile
class Pkg(ConanFile):
python_requires = "pyreq/0.1@user/channel" # recipe to reuse code from
def build(self):
self.python_requires["pyreq"].module # access to the whole conanfile.py module
self.python_requires["pyreq"].module.myvar # access to a variable
self.python_requires["pyreq"].module.myfunct() # access to a global function
self.python_requires["pyreq"].path # access to the folder where the reused file is
Read more about this attribute in Python requires
python_requires_extend¶
This class attribute defines one or more classes that will be injected in runtime as base classes of
the recipe class. Syntax for each of these classes should be a string like pyreq.MyConanfileBase
where the pyreq
is the name of a python_requires
and MyConanfileBase
is the name of the class
to use.
from conan import ConanFile
class Pkg(ConanFile):
python_requires = "pyreq/0.1@user/channel", "utils/0.1@user/channel"
python_requires_extend = "pyreq.MyConanfileBase", "utils.UtilsBase" # class/es to inject
Sources¶
exports¶
List or tuple of strings with file names or fnmatch patterns that should be exported and stored side by side with the conanfile.py file to make the recipe work: other python files that the recipe will import, some text file with data to read,…
For example, if we have some python code that we want the recipe to use in a helpers.py
file,
and have some text file info.txt we want to read and display during the recipe evaluation
we would do something like:
exports = "helpers.py", "info.txt"
Exclude patterns are also possible, with the !
prefix:
exports = "*.py", "!*tmp.py"
See also
exports_sources¶
List or tuple of strings with file names or
fnmatch patterns that should be exported
and will be available to generate the package. Unlike the exports
attribute, these files
shouldn’t be used by the conanfile.py
Python code, but to compile the library or generate
the final package. And, due to its purpose, these files will only be retrieved if requested
binaries are not available or the user forces Conan to compile from sources.
This is an alternative to getting the sources with the source()
method. Used when we are not packaging a third party
library and we have together the recipe and the C/C++ project:
exports_sources = "include*", "src*"
Exclude patterns are also possible, with the !
prefix:
exports_sources = "include*", "src*", "!src/build/*"
Note, if the recipe defines the layout()
method and specifies a self.folders.source = "src"
it won’t affect
where the files (from the exports_sources
) are copied. They will be copied to the base source folder. So, if you
want to replace some file that got into the source()
method, you need to explicitly copy it from the parent folder
or even better, from self.export_sources_folder
.
import os, shutil
from conan import ConanFile
from conan.tools.files import save, load
class Pkg(ConanFile):
...
exports_sources = "CMakeLists.txt"
def layout(self):
self.folders.source = "src"
self.folders.build = "build"
def source(self):
# emulate a download from web site
save(self, "CMakeLists.txt", "MISTAKE: Very old CMakeLists to be replaced")
# Now I fix it with one of the exported files
shutil.copy("../CMakeLists.txt", ".")
shutil.copy(os.path.join(self.export_sources_folder, "CMakeLists.txt", "."))
conan_data¶
Read only attribute with a dictionary with the keys and values provided in a conandata.yml file format placed next to the conanfile.py. This YAML file is automatically exported with the recipe and automatically loaded with it too.
You can declare information in the conandata.yml file and then access it inside any of the methods of the recipe. For example, a conandata.yml with information about sources that looks like this:
sources:
"1.1.0":
url: "https://www.url.org/source/mylib-1.0.0.tar.gz"
sha256: "8c48baf3babe0d505d16cfc0cf272589c66d3624264098213db0fb00034728e9"
"1.1.1":
url: "https://www.url.org/source/mylib-1.0.1.tar.gz"
sha256: "15b6393c20030aab02c8e2fe0243cb1d1d18062f6c095d67bca91871dc7f324a"
def source(self):
get(self, **self.conan_data["sources"][self.version])
source_buildenv¶
Boolean attribute to opt-in injecting the VirtualBuildEnv generated environment while running the source() method.
Setting this attribute to True (default value False) will inject the VirtualBuildEnv
generated environment from tool requires
when executing the source() method.
class MyConan:
name = "mylib"
version = "1.0.0"
source_buildenv = True
tool_requires = "7zip/1.2.0"
def source(self):
get(self, **self.conan_data["sources"][self.version])
self.run("7z x *.zip -o*") ## Can run 7z in the source method
Binary model¶
Important attributes that define the package binaries model, which settings, options, package type, etc. affect the final packaged binaries.
package_type¶
Optional. Declaring the package_type
will help Conan:
To choose better the default
package_id_mode
for each dependency, that is, how a change in a dependency should affect thepackage_id
to the current package.Which information from the dependencies should be propagated to the consumers, like headers, libraries, runtime information. See here to see what traits are propagated based on the
package_type
information.
The valid values are:
application: The package is an application.
library: The package is a generic library. It will try to determine the type of library (from
shared-library
,static-library
,header-library
) reading theself.options.shared
(if declared) and theself.options.header_only
shared-library: The package is a shared library.
static-library: The package is a static library.
header-library: The package is a header only library.
build-scripts: The package only contains build scripts.
python-require: The package is a python require.
unknown: The type of the package is unknown.
settings¶
List of strings with the first level settings (from settings.yml) that the recipe
needs, because:
- They are read for building (e.g: if self.settings.compiler == “gcc”)
- They affect the package_id
. If a value of the declared setting changes, the package_id
has to be different.
The most common is to declare:
settings = "os", "compiler", "build_type", "arch"
Once the recipe is loaded by Conan, the settings
are processed and they can be read in the recipe, also
the sub-settings:
settings = "os", "arch"
def build(self):
if self.settings.compiler == "gcc":
if self.settings.compiler.cppstd == "gnu20":
# do some special build commands
If you try to access some setting that doesn’t exist, like self.settings.compiler.libcxx
for the msvc
setting, Conan will fail telling that libcxx
does not exist for that compiler.
If you want to do a safe check of settings values, you could use the get_safe()
method:
def build(self):
# Will be None if doesn't exist (not declared)
arch = self.settings.get_safe("arch")
# Will be None if doesn't exist (doesn't exist for the current compiler)
compiler_version = self.settings.get_safe("compiler.version")
# Will be the default version if the return is None
build_type = self.settings.get_safe("build_type", default="Release")
The get_safe()
method returns None
if that setting or sub-setting doesn’t
exist and there is no default value assigned.
It’s also feasible to check the possible values defined in settings.yml using the
possible_values()
method:
def generate(self):
# Print if Android exists as OS in the whole settings.yml
is_android = "Android" in self.settings.possible_values()["os"]
self.output.info(f"Android in settings.yml: {is_android}")
# Print the available versions for the compiler used by the HOST profile
compiler_versions = self.settings.compiler.version.possible_values()
self.output.info(f"[HOST] Versions for {str(self.settings.compiler)}: {', '.join(compiler_versions)}")
# Print the available versions for the compiler used by the BUILD profile
compiler_versions = self.settings_build.compiler.version.possible_values()
self.output.info(f"[BUILD] Versions for {str(self.settings.compiler)}: {', '.join(compiler_versions)}")
As you can see above, doing self.settings.possible_values()
returns the
whole settings.yml as a Python dict-like object, and doing
self.settings.compiler.version.possible_values()
for instance returns the available versions for the compiler
used by the consumer.
If you want to do a safe deletion of settings, you could use the rm_safe()
method.
For example, in the configure()
method a typical pattern for a C library would be:
def configure(self):
self.settings.rm_safe("compiler.libcxx")
self.settings.rm_safe("compiler.cppstd")
options¶
Dictionary with traits that affects only the current recipe, where the key is the option
name and the value is a list of different values that the option can take. By default any
value change in an option, changes the package_id
. Check the default_options
and
default_build_options
fields to define default values for the options.
Values for each option can be typed or plain strings ("value"
, True
, 42
,…).
There are two special values:
None
: Allow the option to have aNone
value (not specified) without erroring."ANY"
: For options that can take any value, not restricted to a set.
class MyPkg(ConanFile):
...
options = {
"shared": [True, False],
"option1": ["value1", "value2"],
"option2": ["ANY"],
"option3": [None, "value1", "value2"],
"option4": [True, False, "value"],
}
Once the recipe is loaded by Conan, the options
are processed and they can be read in the recipe. You can also
use the method .get_safe()
(see settings attribute) to avoid Conan raising an Exception if the option
doesn’t exist:
class MyPkg(ConanFile):
options = {"shared": [True, False]}
def build(self):
if self.options.shared:
# build the shared library
if self.options.get_safe("foo", True):
pass
In boolean expressions, like if self.options.shared
:
equals
True
for the valuesTrue
,"True"
and"true"
, and any other value that would be evaluated the same way in Python code.equals
False
for the valuesFalse
,"False"
and"false"
, also for the empty string and for0
and"0"
as expected.
Notice that a comparison using is
is always False
because the types would be different as it is encapsulated
inside a Python class.
If you want to do a safe deletion of options, you could use the rm_safe()
method.
For example, in the config_options()
method a typical pattern for Windows library
would be:
def config_options(self):
if self.settings.os == "Windows":
self.options.rm_safe("fPIC")
See also
Read the Getting started, creating packages to know how to declare and how to define a value to an option.
Removing options in the
package_id()
method. <MISSING PAGE>About the package_type and how it plays when a
shared
option is declared. <MISSING PAGE>
default_options¶
The attribute default_options
defines the default values for the options, both for the
current recipe and for any requirement.
This attribute should be defined as a python dictionary.
class MyPkg(ConanFile):
...
requires = "zlib/1.2.8", "zwave/2.0"
options = {"build_tests": [True, False],
"option2": "ANY"}
default_options = {"build_tests": True,
"option1": 42,
"z*:shared": True}
You can also assign default values for options of your requirements using “<reference_pattern>: option_name”, being
a valid reference_pattern
a name/version
or any pattern with *
like the example above.
You can also set the options conditionally to a final value with configure()
instead of using default_options
:
class OtherPkg(ConanFile):
settings = "os", "arch", "compiler", "build_type"
options = {"some_option": [True, False]}
# Do NOT declare 'default_options', use 'config_options()'
def configure(self):
if self.options.some_option == None:
if self.settings.os == 'Android':
self.options.some_option = True
else:
self.options.some_option = False
Take into account that if a value is assigned in the configure()
method it cannot be overridden.
See also
There are 2 different ways that a recipe can try to define options values for its dependencies.
Using default_options = {"mypkg/*:myoption", 123}
the current recipe can define the 123
value
to the dependency mypkg
myoption
. This way of defining options for dependencies has
some limitations:
Any other downstream user of the current recipe that defines the same option for
mypkg
will have precedence, overwriting the current recipe123
value. Also any definition in the profile or command line will also have precedence. The recipedefault_options
have the least precedence. If a recipe will not work at all with some dependencies options, then recipes can check and raiseConanInvalidConfiguration
errors accordingly.Any sibling package that depends on
mypkg
will also define its options and it will be the only one being taken into account. In other words, the first timemypkg
is required by any other package will “freeze” its currently assigned options values. Any other package that depends later onmypkg
, closing the diamond structures in the dependency graph will not have any influence on themypkg
options. Only the first one requiring it will.
The second way to define the options values is defining them as important!
.
Warning
The important!
syntax is experimental and can be changed or removed at any time.
A recipe can define its dependencies options as important!
with the syntax
default_options = {"mypkg/*:myoption!", 123}
. That means that the mypkg
myoption
will not be overriden by other downstream packages, profile or command line doing regular
definition of options (like -o *:myoption=234
).
But there are 2 cases in which this will still not define the final value of the dependency:
If any downstream recipe, command line or profile also uses the
myoption!
syntax, that will also have precedence and override the value upstreamIf there is any other package that requires first
mypkg
, the values defined at that moment will still have precedence.
In general the recommendation for defining options values is to do it in profile
files,
not in recipes, as in-recipe definition can be more complicated specially for complex
dependency graphs.
default_build_options¶
The attribute default_build_options
defines the default values for the options in the
build context and is typically used for defining options for tool_requires
.
from conan import ConanFile
class Consumer(ConanFile):
default_options = {"protobuf/*:shared": True}
default_build_options = {"protobuf/*:shared": False}
def requirements(self):
self.requires("protobuf/1.0")
def build_requirements(self):
self.tool_requires("protobuf/1.0")
options_description¶
The options_description
attribute is an optional attribute that can be defined in the
form of a dictionary where the key is the option name and the value is a description of
the option in text format. This attribute is useful for providing additional information
about the functionality and purpose of each option, particularly when the option is not
self-explanatory or has complex or special behavior.
The format for each dictionary entry should be:
Key: Option name. Must be a string and must match one of the keys in the
options
dictionary.Value: Description of the option. Must be a string and can be as long as necessary.
For example:
class MyPkg(ConanFile):
...
options = {"option1": [True, False],
"option2": "ANY"}
options_description = {
"option1": "Describe the purpose and functionality of 'option1'. ",
"option2": "Describe the purpose and functionality of 'option2'. ",
}
info¶
Object used exclusively in package_id()
method:
The :ref:package_id method<reference_conanfile_methods_package_id> to control the unique ID for a package:
def package_id(self): self.info.clear()
The self.info.clear()
method removes all the settings, options, requirements (requires
, tool_requires
, python_requires
)
and configuration (conf
) from the package_id
computation, so the package_id
will always result in the same binary, irrespective
of all those things. This would be the typical case of a header-only library, in which the packaged artifacts (files) are always identical.
package_id_{embed,non_embed,python,unknown}_mode¶
The package_id_embed_mode, package_id_non_embed_mode, package_id_python_mode, package_id_unknown_mode
are class attributes that can be defined in recipes to define the effect they have on their consumers package_id
. Can be declared as:
from conan import ConanFile
class Pkg(ConanFile):
name = "pkg"
version = "1.0.0"
# They are not mandatory, and it is not necessary to define all
package_id_embed_mode = "full_mode"
package_id_non_embed_mode = "patch_mode"
package_id_unknown_mode = "minor_mode"
package_id_python_mode = "major_mode"
In general, the Conan defaults are good ones, and allow providing users good control over when the consumers need to be re-built from source or not. Also, the Conan defaults can be changed globally in the global.conf
file (they should be changed globally for all users, CI, etc.) via the core.package_id:xxxx
configurations. The in-recipe attribute definition is useful to define behavior that deviates from the defaults.
Possible values are (following the semver definition of MAJOR.MINOR.PATCH):
patch_mode
: New patches, minors, and major releases of the package will require a new binary (newpackage_id
) of the consumers. New recipe revisions will not require new binaries of the consumers. For example if we create a newpkg/1.0.1
version and some consumer hasrequires = "pkg/[>=1.0 <2.0]"
, such a consumer will build a new binary against this specific new1.0.1
version. But if we just change the recipe, producing a newrecipe_revision
, the consumers will not require building a new binary.minor_mode
: New minor and major releases of this package will require a new binary of the consumers. New patches and new revisions will not require new binaries of the consumers. This is the default for the “non-embed-mode”, as it allows fine control by the users to decide when to rebuild things or not.major_mode
: Only new major releases will require new binaries. Any other modifications and new versions will not require new binaries from the consumers.full_mode
: The full identifier of this package, includingpkgname/version@user/channel#recipe_revision:package_id
will be used in the consumerspackage_id
, then requiring to build a new binary of the consumer for every change of this package (as any change either in source or configuration will produce a differentrecipe_revision
orpackage_id
respectively). This is the default for the “embed-mode”.unrelated_mode
: No change in this package will ever produce a new binary in the consumer.
The 4 different attributes are:
package_id_embed_mode
: Define the mode for “embedding” cases, that is, a shared library linking a static library, an application linking a static library, an application or a library linking a header-only library. The default for this mode isfull_mode
package_id_non_embed_mode
. Define the mode for “non-embedding” cases, that is, a shared library linking another shared library, a static library linking another static library, an application executable linking a shared library. The default for this mode isminor_mode
.package_id_unknown_mode
: Define the mode when the relationship between packages is unknown. If it is not possible to deduce the package type, because there are noshared
orheader_only
options defined, or becausepackage_type
is not defined, then, this mode will be used. The default for this mode issemver_mode
(similar to Conan 1.X behavior).package_id_python_mode
: Define the mode for consumers ofpython_requires
. By default it will beminor_mode
, and it is strongly recommended to use this default, and not define thepackage_id_python_mode
. This attribute is provided for completeness and exceptional cases like temporary migrations.
See also
Read the binary model reference for a full view of the Conan binary model.
Build¶
generators¶
List or tuple of strings with names of generators.
class MyLibConan(ConanFile):
generators = "CMakeDeps", "CMakeToolchain"
The generators can also be instantiated explicitly in the generate() method.
from conan.tools.cmake import CMakeToolchain
class MyLibConan(ConanFile):
...
def generate(self):
tc = CMakeToolchain(self)
tc.generate()
build_policy¶
Controls when the current package is built during a conan install
.
The allowed values are:
"missing"
: Conan builds it from source if there is no binary available."never"
: This package cannot be built from sources, it is always created withconan export-pkg
None
(default value): This package won’t be built unless the policy is specified in the command line (e.g--build=foo*
)class PocoTimerConan(ConanFile): build_policy = "missing"
win_bash¶
When True
it enables the new run in a subsystem bash in Windows mechanism.
from conan import ConanFile
class FooRecipe(ConanFile):
...
win_bash = True
It can also be declared as a property
based on any condition:
from conan import ConanFile
class FooRecipe(ConanFile):
...
@property
def win_bash(self):
return self.settings.arch == "armv8"
win_bash_run¶
When True
it enables running commands in the "run"
scope, to run them inside a bash shell.
from conan import ConanFile
class FooRecipe(ConanFile):
...
win_bash_run = True
def build(self):
self.run(cmd, scope="run") # will run <cmd> inside bash
Folders and layout¶
source_folder¶
The folder in which the source code lives. The path is built joining the base directory
(a cache directory when running in the cache or the output folder
when running locally)
with the value of folders.source
if declared in the layout()
method.
Note that the base directory for the source_folder
when running in the cache will point to the base folder of the
build unless no_copy_source is set to True
. But anyway it will
always point to the correct folder where the source code is.
export_sources_folder¶
The value depends on the method you access it:
At
source(self)
: Points to the base source folder (that means self.source_folder but without taking into account thefolders.source
declared in thelayout()
method). The declared exports_sources are copied to that base source folder always.At
exports_sources(self)
: Points to the folder in the cache where the export sources have to be copied.
build_folder¶
The folder used to build the source code. The path is built joining the base directory (a cache
directory when running in the cache or the output folder
when running locally) with
the value of folders.build
if declared in the layout()
method.
package_folder¶
The folder to copy the final artifacts for the binary package. In the local cache a package folder is created for every different package ID.
The most common usage of self.package_folder
is to copy
the files
at the package() method:
import os
from conan import ConanFile
from conan.tools.files import copy
class MyRecipe(ConanFile):
...
def package(self):
copy(self, "*.so", self.build_folder, os.path.join(self.package_folder, "lib"))
...
recipe_folder¶
The folder where the recipe conanfile.py is stored, either in the local folder or in
the cache. This is useful in order to access files that are exported along with the recipe,
or the origin folder when exporting files in export(self)
and export_sources(self)
methods.
The most common usage of self.recipe_folder
is in the export(self)
and export_sources(self)
methods,
as the folder from where we copy the files:
from conan import ConanFile
from conan.tools.files import copy
class MethodConan(ConanFile):
exports = "file.txt"
def export(self):
copy(self, "LICENSE.md", self.recipe_folder, self.export_folder)
recipe_metadata_folder¶
The self.recipe_metadata_folder
(experimental) can be used in the export()
and export_sources()
and source()
methods
to save or copy recipe metadata files. See metadata section for more information.
package_metadata_folder¶
The self.package_metadata_folder
(experimental) can be used in the generate()
, build()
and package()
methods
to save or copy package metadata files. See metadata section for more information.
no_copy_source¶
The attribute no_copy_source
tells the recipe that the source code will not be copied from
the source_folder
to the build_folder
. This is mostly an optimization for packages
with large source codebases or header-only, to avoid extra copies.
If you activate no_copy_source=True
, it is mandatory that the source code must not be modified at all by
the configure or build scripts, as the source code will be shared among all builds.
The recipes should always use self.source_folder
attribute, which will point to the build
folder when no_copy_source=False
and will point to the source
folder when no_copy_source=True
.
See also
Read header-only packages section for an
example using no_copy_source
attribute.
Layout¶
folders¶
The folders
attribute has to be set only in the layout()
method. Please check the
layout() method documentation to learn more about this
attribute.
cpp¶
Object storing all the information needed by the consumers
of a package: include directories, library names, library paths… Both for editable
and regular packages in the cache. It is only available at the layout()
method.
self.cpp.package
: For a regular package being used from the Conan cache. Same as declaringself.cpp_info
at thepackage_info()
method.self.cpp.source
: For “editable” packages, to describe the artifacts underself.source_folder
self.cpp.build
: For “editable” packages, to describe the artifacts underself.build_folder
.
The cpp
attribute has to be set only in the layout()
method. Please check the
layout() method documentation to learn more about this
attribute.
layouts¶
The layouts
attribute has to be set only in the layout()
method. Please check the
layout() method documentation to learn more about this
attribute.
The layouts
attribute contains information about environment variables and conf
that
would be path-dependent, and as a result it would contain a different value when the package
is in editable mode, or when the package is in the cache. The layouts
sub-attributes are:
self.layouts.build
: information related to the relativeself.folders.build
self.layouts.source
: information related to the relativeself.folders.source
self.layouts.package
: information related to the finalpackage_folder
Each one of those will contain:
buildenv_info
: environment variables build information for consumers (equivalent toself.buildenv_info
inpackage_info()
)runenv_info
: environment variables run information for consumers (equivalent toself.runenv_info
inpackage_info()
)conf_info
: configuration information for consumers (equivalent toself.conf_info
inpackage_info()
). Note this is only automatically propagated toself.conf
of consumers when this package is a directtool_require
.
For example, if we had an androidndk
recipe that contains the AndroidNDK, and we want to have that recipe in “editable” mode,
it is necessary where the androidndk will be locally, before being in the created package:
import os
from conan import ConanFile
from conan.tools.files import copy
class AndroidNDK(ConanFile):
def layout(self):
# When developing in user space it is in a "mybuild" folder (relative to current dir)
self.layouts.build.conf_info.define_path("tools.android:ndk_path", "mybuild")
# but when packaged it will be in a "mypkg" folder (inside the cache package folder)
self.layouts.package.conf_info.define_path("tools.android:ndk_path", "mypkg")
def package(self):
copy(self, "*", src=os.path.join(self.build_folder, "mybuild"),
dst=os.path.join(self.package_folder, "mypkg"))
Package information for consumers¶
cpp_info¶
Same as using self.cpp.package
in the layout()
method. Use it if you need to read
the package_folder
to locate the already located artifacts.
See also
CppInfo model.
Important
This attribute is only defined inside package_info()
method being None elsewhere.
buildenv_info¶
For the dependant recipes, the declared environment variables will be present during the
build process. Should be only filled in the package_info()
method.
Important
This attribute is only defined inside package_info()
method being None elsewhere.
def package_info(self):
self.buildenv_info.append_path("PATH", self.package_folder)
See also
Check the reference of the Environment object to know how to fill
the self.buildenv_info
.
runenv_info¶
For the dependant recipes, the declared environment variables will be present at runtime.
Should be only filled in the package_info()
method.
Important
This attribute is only defined inside package_info()
method being None elsewhere.
def package_info(self):
self.runenv_info.define_path("RUNTIME_VAR", "c:/path/to/exe")
See also
Check the reference of the Environment object to know how to fill
the self.runenv_info
.
conf_info¶
Configuration variables to be passed to the dependant recipes.
Should be only filled in the package_info()
method.
class Pkg(ConanFile):
name = "pkg"
def package_info(self):
self.conf_info.define("tools.build:verbosity", "debug")
self.conf_info.get("tools.build:verbosity") # == "debug"
self.conf_info.append("user.myconf.build:ldflags", "--flag3") # == ["--flag1", "--flag2", "--flag3"]
self.conf_info.update("tools.microsoft.msbuildtoolchain:compile_options", {"ExpandAttributedSource": "false"})
self.conf_info.unset("tools.microsoft.msbuildtoolchain:compile_options")
self.conf_info.remove("user.myconf.build:ldflags", "--flag1") # == ["--flag0", "--flag2", "--flag3"]
self.conf_info.pop("tools.system.package_manager:sudo")
See also
Read here the complete reference of self.conf_info.
deprecated¶
This attribute declares that the recipe is deprecated, causing a user-friendly warning message to be emitted whenever it is used
For example, the following code:
from conan import ConanFile
class Pkg(ConanFile):
name = "cpp-taskflow"
version = "1.0"
deprecated = True
may emit a warning like:
cpp-taskflow/1.0: WARN: Recipe 'cpp-taskflow/1.0' is deprecated. Please, consider changing your requirements.
Optionally, the attribute may specify the name of the suggested replacement:
from conan import ConanFile
class Pkg(ConanFile):
name = "cpp-taskflow"
version = "1.0"
deprecated = "taskflow"
This will emit a warning like:
cpp-taskflow/1.0: WARN: Recipe 'cpp-taskflow/1.0' is deprecated in favor of 'taskflow'. Please, consider changing your requirements.
If the value of the attribute evaluates to False
, no warning is printed.
provides¶
This attribute declares that the recipe provides the same functionality as other recipe(s). The attribute is usually needed if two or more libraries implement the same API to prevent link-time and run-time conflicts (ODR violations). One typical situation is forked libraries. Some examples are:
If Conan encounters two or more libraries providing the same functionality within a single graph, it raises an error:
At least two recipes provides the same functionality:
- 'libjpeg' provided by 'libjpeg/9d', 'libjpeg-turbo/2.0.5'
The attribute value should be a string with a recipe name or a tuple of such recipe names.
For example, to declare that libjpeg-turbo
recipe offers the same functionality as libjpeg
recipe, the following code could be used:
from conan import ConanFile
class LibJpegTurbo(ConanFile):
name = "libjpeg-turbo"
version = "1.0"
provides = "libjpeg"
To declare that a recipe provides the functionality of several different recipes at the same time, the following code could be used:
from conan import ConanFile
class OpenBLAS(ConanFile):
name = "openblas"
version = "1.0"
provides = "cblas", "lapack"
If the attribute is omitted, the value of the attribute is assumed to be equal to the current package name. Thus, it’s redundant for
libjpeg
recipe to declare that it provides libjpeg
, it’s already implicitly assumed by Conan.
Other¶
dependencies¶
Conan recipes provide access to their dependencies via the self.dependencies
attribute.
class Pkg(ConanFile):
requires = "openssl/0.1"
def generate(self):
openssl = self.dependencies["openssl"]
# access to members
openssl.ref.version
openssl.ref.revision # recipe revision
openssl.options
openssl.settings
See also
Read here the complete reference of self.dependencies.
conf¶
In the self.conf
attribute we can find all the conf entries declared in the [conf] section of the profiles.
in addition of the declared self.conf_info entries from the first level tool requirements.
The profile entries have priority.
from conan import ConanFile
class MyConsumer(ConanFile):
tool_requires = "my_android_ndk/1.0"
def generate(self):
# This is declared in the tool_requires
self.output.info("NDK host: %s" % self.conf.get("tools.android:ndk_path"))
# This is declared in the profile at [conf] section
self.output.info("Custom var1: %s" % self.conf.get("user.custom.var1"))
Note
The conf
attribute is a read-only attribute. It can only be defined in profiles and command lines, but it should never be set by recipes.
Recipes can only read its value via self.conf.get()
method.
Output¶
Output contents¶
Use the self.output
to print contents to the output.
self.output.success("This is good, should be green")
self.output.info("This is neutral, should be white")
self.output.warning("This is a warning, should be yellow")
self.output.error("Error, should be red")
Additional output methods are available and you can produce different outputs with different colors. See the output documentation for the list of available output methods.
revision_mode¶
This attribute allow each recipe to declare how the revision for the recipe itself should be computed. It can take three different values:
"hash"
(by default): Conan will use the checksum hash of the recipe manifest to compute the revision for the recipe."scm"
: if the project is inside a Git repository the commit ID will be used as the recipe revision. If there is no repository it will raise an error."scm_folder"
: This configuration applies when you have a mono-repository project, but still want to use scm revisions. In this scenario, the revision of the exported conanfile.py will correspond to the commit ID of the folder where it’s located. This approach allows multiple conanfile.py files to exist within the same Git repository, with each file exported under its distinct revision.
When scm
or scm_folder
is selected, the Git commit will be used, but by default
the repository must be clean, otherwise it would be very likely that there are uncommitted
changes and the build wouldn’t be reproducible. So if there are dirty files, Conan will raise
an error. If there are files that can be dirty in the repo, but do not belong at all to the
recipe or the package, then it is possible to exclude them from the check with the core.scm:excluded
configuration, which is a list of patterns (fnmatch) to exclude.
upload_policy¶
Controls when the current package built binaries are uploaded or not
"skip"
: The precompiled binaries are not uploaded. This is useful for “installer” packages that just download and unzip something heavy (e.g. android-ndk), and is useful together with thebuild_policy = "missing"
class Pkg(ConanFile): upload_policy = "skip"
required_conan_version¶
Recipes can define a module level required_conan_version
that defines a valid version range of
Conan versions that can load and understand the current conanfile.py
. The syntax is:
from conan import ConanFile
required_conan_version = ">=2.0"
class Pkg(ConanFile):
pass
Version ranges as in requires
are allowed.
Also there is a global.conf
file core:required_conan_version
configuration that can
define a global minimum, maximum or exact Conan version to run, which can be very convenient
to maintain teams of developers and CI machines to use the desired range of versions.
implements¶
A list is used to define a series of option configurations that Conan will handle automatically. This is especially handy for avoiding boilerplate code that tends to repeat in most of the recipes. The syntax is as follows:
from conan import ConanFile
class Pkg(ConanFile):
implements = ["auto_shared_fpic", "auto_header_only", ...]
Currently these are the automatic implementations provided by Conan:
"auto_shared_fpic"
: automatically managesfPIC
andshared
options. Adding this implementation will have both effect in the configure and config_options steps when those methods are not explicitly defined in the recipe."auto_header_only"
: automatically manages the package ID clearing settings. Adding this implementation will have effect in the package_id step when the method is not explicitly defined in the recipe.
Warning
This is a 2.0-only feature, and it will not work in 1.X
alias¶
Warning
While aliases can technically still be used in Conan 2, their usage is not recommended and they may be fully removed in future releases. Users are encouraged to adapt to the newer versioning features for a more standardized and efficient package management experience.
In Conan 2, the alias attribute remains a part of the recipe, allowing users to define
an alias for a package version. Normally, you would create one using the conan new
command with the alias
template and the exporting the recipe with conan export:
$ conan new alias -d name=mypkg -d version=latest -d target=1.0
$ conan export .
Note that when requiring the alias, you must place the version in parentheses ()
to
explicitly declare the use of an alias as a requirement:
class Consumer(ConanFile):
...
requires = "mypkg/(latest)"
...
extension_properties¶
The extensions_properties
attribute is a dictionary intended to define and pass information from the
recipes to the Conan extensions.
At the moment, the only defined property is compatibility_cppstd
, that allows disabling the behavior
of the default compatibility.py extension, that considers
binaries built with different compiler.cppstd
values ABI-compatible among them.
To disable this behavior for the current package, it is possible to do it with:
class Pkg(ConanFile):
extension_properties = {"compatibility_cppstd": False}
If it is necessary to do it conditionally, it is also possible to define its value inside recipe compatibility()
method:
class Pkg(ConanFile):
def compatibility(self):
self.extension_properties = {"compatibility_cppstd": False}