The BitBake task executor together with various types of configuration files form the OpenEmbedded Core. This section provides an overview of the task executor and the configuration files by describing their use and interaction. BitBake handles the parsing and execution of the data files. The data itself is of various types: Recipes: Details about particular pieces of software. Class Data: An abstraction of common build information (e.g. how to build a Linux kernel). Configuration Data: Machine-specific settings, policy decisions, and so forth. Configuration data acts as the glue to bind everything together. The remainder of this chapter provides examples of BitBake metadata. Any syntax not supported in any of the previously listed areas is documented as such.

This section provides some basic syntax examples.

The following example sets VARIABLE to "value". This assignment occurs immediately as the statement is parsed. It is a "hard" assignment. VARIABLE = "value"

BitBake supports variables referencing one another's contents using a syntax that is similar to shell scripting. Following is an example that results in A containing "aval" and B containing "preavalpost". A = "aval" B = "pre${A}post"

You can use the "?=" operator to achieve a "softer" assignment for a variable. This type of assignment allows you to define a variable if it is undefined when the statement is parsed, but to leave the value alone if the variable has a value. Here is an example: A ?= "aval" If A is set at the time this statement is parsed, the variable retains its value. However, if A is not set, the variable is set to "aval". This assignment is immediate. Consequently, if multiple "?=" assignments to a single variable exist, the first of those ends up getting used.

It is possible to use a "weaker" assignment that in the previous section by using the "??=" operator. This assignment behaves identical to "?=" except that the assignment is made at the end of the parsing process rather than immediately. Consequently, when multiple "??=" assignments exist, the last one is used. Also, any "=" or "?=" assignment will override the value set with "??=". Here is an example: A ??= "somevalue" A ??= "someothervalue" If A is set before the above statements are parsed, the variable retains its value. If A is not set, the variable is set to "someothervalue". Again, this assignment is a "lazy" or "weak" assignment because it does not occur until the end of the parsing process.

The ":=" operator results in a variable's contents being expanded immediately, rather than when the variable is actually used: T = "123" A := "${B} ${A} test ${T}" T = "456" B = "${T} bval" C = "cval" C := "${C}append" In this example, A contains "test 123" because ${B} and ${A} at the time of parsing are undefined, which leaves "test 123". And, the variable C contains "cvalappend" since ${C} immediately expands to "cval".

Appending and prepending values is common and can be accomplished using the "+=" and "=+" operators. These operators insert a space between the current value and prepended or appended value. Here are some examples: B = "bval" B += "additionaldata" C = "cval" C =+ "test" The variable B contains "bval additionaldata" and C contains "test cval".

If you want to append or prepend values without an inserted space, use the ".=" and "=." operators. Here are some examples: B = "bval" B .= "additionaldata" C = "cval" C =. "test" The variable B contains "bvaladditionaldata" and C contains "testcval".

You can also append and prepend a variable's value using an override style syntax. When you use this syntax, no spaces are inserted. Here are some examples: B = "bval" B_append = " additional data" C = "cval" C_prepend = "additional data " D = "dval" D_append = "additional data" The variable B becomes "bval additional data" and C becomes "additional data cval". The variable D becomes "dvaladditional data". You must control all spacing when you use the override syntax.

You can remove values from lists using the removal override style syntax. Specifying a value for removal causes all occurrences of that value to be removed from the variable. When you use this syntax, BitBake expects one or more strings. Surrounding spaces are removed as well. Here is an example: FOO = "123 456 789 123456 123 456 123 456" FOO_remove = "123" FOO_remove = "456" FOO2 = "abc def ghi abcdef abc def abc def" FOO2_remove = "abc def" The variable FOO becomes "789 123456" and FOO2 becomes "ghi abcdef".

Variable flags are BitBake's implementation of variable properties or attributes. It is a way of tagging extra information onto a variable. You can find more out about variable flags in general in the "Variable Flags" section. You can define, append, and prepend values to variable flags. All the standard syntax operations previously mentioned work for variable flags except for override style syntax (i.e. _prepend, _append, and _remove). Here are some examples showing how to set variable flags: FOO[a] = "abc" FOO[b] = "123" FOO[a] += "456" The variable FOO has two flags: a and b. The flags are immediately set to "abc" and "123", respectively. The a flag becomes "abc456".

You can use inline Python variable expansion to set variables. Here is an example: DATE = "${@time.strftime('%Y%m%d',time.gmtime())}" This example results in the DATE variable becoming the current date.

BitBake uses OVERRIDES to control what variables are overridden after BitBake parses recipes and configuration files. This section describes how you can use OVERRIDES as conditional metadata, talks about key expansion in relationship to OVERRIDES, and provides some examples to help with understanding.

You can use OVERRIDES to conditionally select a specific version of a variable and to conditionally append or prepend the value of a variable. Selecting a Variable: The OVERRIDES variable is a colon-character-separated list that contains items for which you want to satisfy conditions. Thus, if you have a variable that is conditional on “arm”, and “arm” is in OVERRIDES, then the “arm”-specific version of the variable is used rather than the non-conditional version. Here is an example: OVERRIDES = "architecture:os:machine" TEST = "default" TEST_os = "osspecific" TEST_nooverride = "othercondvalue" In this example, the OVERRIDES variable lists three overrides: "architecture", "os", and "machine". The variable TEST by itself has a default value of "default". You select the os-specific version of the TEST variable by appending the "os" override to the variable (i.e.TEST_os). Appending and Prepending: BitBake also supports append and prepend operations to variable values based on whether a specific item is listed in OVERRIDES. Here is an example: DEPENDS = "glibc ncurses" OVERRIDES = "machine:local" DEPENDS_append_machine = "libmad" In this example, DEPENDS becomes "glibc ncurses libmad".

Key expansion happens when the BitBake datastore is finalized just before BitBake expands overrides. To better understand this, consider the following example: A${B} = "X" B = "2" A2 = "Y" In this case, after all the parsing is complete, and before any overrides are handled, BitBake expands ${B} into "2". This expansion causes A2, which was set to "Y" before the expansion, to become "X".

Despite the previous explanations that show the different forms of variable definitions, it can be hard to work out exactly what happens when variable operators, conditional overrides, and unconditional overrides are combined. This section presents some common scenarios along with explanations for variable interactions that typically confuse users. There is often confusion concerning the order in which overrides and various "append" operators take effect. Recall that an append or prepend operation using "_append" and "_prepend" does not result in an immediate assignment as would "+=", ".=", "=+", or "=.". Consider the following example: OVERRIDES = "foo" A = "Z" A_foo_append = "X" For this case, A is unconditionally set to "Z" and "X" is unconditionally and immediately appended to the variable A_foo. Because overrides have not been applied yet, A_foo is set to "X" due to the append and A simply equals "Z". Applying overrides, however, changes things. Since "foo" is listed in OVERRIDES, the conditional variable A is replaced with the "foo" version, which is equal to "X". So effectively, A_foo replaces A. This next example changes the order of the override and the append: OVERRIDES = "foo" A = "Z" A_append_foo = "X" For this case, before overrides are handled, A is set to "Z" and A_append_foo is set to "X". Once the override for "foo" is applied, however, A gets appended with "X". Consequently, A becomes "ZX". Notice that spaces are not appended. This next example has the order of the appends and overrides reversed back as in the first example: OVERRIDES = "foo" A = "Y" A_foo_append = "Z" A_foo_append += "X" For this case, before any overrides are resolved, A is set to "Y" using an immediate assignment. After this immediate assignment, A_foo is set to "Z", and then further appended with "X" leaving the variable set to "Z X". Finally, applying the override for "foo" results in the conditional variable A becoming "Z X" (i.e. A is replaced with A_foo). This final example mixes in some varying operators: A = "1" A_append = "2" A_append = "3" A += "4" A .= "5" For this case, the type of append operators are affecting the order of assignments as BitBake passes through the code multiple times. Initially, A is set to "1 45" because of the three statements that use immediate operators. After these assignments are made, BitBake applies the _append operations. Those operations result in A becoming "1 4523".

BitBake allows for metadata sharing through include files (.inc) and class files (.bbclass). For example, suppose you have a piece of common functionality such as a task definition that you want to share between more than one recipe. In this case, creating a .bbclass file that contains the common functionality and then using the inherit directive in your recipes to inherit the class would be a common way to share the task. This section presents the mechanisms BitBake provides to allow you to share functionality between recipes. Specifically, the mechanisms include include, inherit, INHERIT, and require directives.

BitBake uses the BBPATH variable to locate needed include and class files. The BBPATH variable is analogous to the environment variable PATH. In order for include and class files to be found by BitBake, they need to be located in a "classes" subdirectory that can be found in BBPATH.

When writing a recipe or class file, you can use the inherit directive to inherit the functionality of a class (.bbclass). BitBake only supports this directive when used within recipe and class files (i.e. .bb and .bbclass). The inherit directive is a rudimentary means of specifying what classes of functionality your recipes require. For example, you can easily abstract out the tasks involved in building a package that uses Autoconf and Automake and put those tasks into a class file that can be used by your package. As an example, your recipes could use the following directive to inherit an autotools.bbclass file. The class file would contain common functionality for using Autotools that could be shared across recipes: inherit autotools In this case, BitBake would search for the directory classes/autotools.bbclass in BBPATH. You can override any values and functions of the inherited class within your recipe by doing so after the "inherit" statement.

BitBake understands the include directive. This directive causes BitBake to parse whatever file you specify, and to insert that file at that location. The directive is much like Make except that if the path specified on the include line is a relative path, BitBake locates the first file it can find within BBPATH. As an example, suppose you needed a recipe to include some self-test definitions: include test_defs.inc The include directive does not produce an error when the file cannot be found. Consequently, it is recommended that if the file you are including is expected to exist, you should use require instead of include. Doing so makes sure that an error is produced if the file cannot be found.

BitBake understands the require directive. This directive behaves just like the include directive with the exception that BitBake raises a parsing error if the file to be included cannot be found. Thus, any file you require is inserted into the file that is being parsed at the location of the directive. Similar to how BitBake uses include, if the path specified on the require line is a relative path, BitBake locates the first file it can find within BBPATH. As an example, suppose you have two versions of a recipe (e.g. foo_1.2.2.bb and foo_2.0.0.bb) where each version contains some identical functionality that could be shared. You could create an include file named foo.inc that contains the common definitions needed to build "foo". You need to be sure foo.inc is located in the same directory as your two recipe files as well. Once these conditions are set up, you can share the functionality using a require directive from within each recipe: require foo.inc

When creating a configuration file (.conf), you can use the INHERIT directive to inherit a class. BitBake only supports this directive when used within a configuration file. As an example, suppose you needed to inherit a class file called abc.bbclass from a configuration file as follows: INHERIT += "abc" This configuration directive causes the named class to be inherited at the point of the directive during parsing. As with the inherit directive, the .bbclass file must be located in a "classes" subdirectory in one of the directories specified in BBPATH. Because .conf files are parsed first during BitBake's execution, using INHERIT to inherit a class effectively inherits the class globally (i.e. for all recipes).

As with most languages, functions are the building blocks that are used to build up operations into tasks. BitBake supports three types of functions: Shell Functions: Functions written in shell script and executed either directly as functions, tasks, or both. They can also be called by other shell functions. BitBake Style Python Functions: Functions written in Python and executed by BitBake or other Python functions using bb.build.exec_func(). Python Functions: Functions written in Python and executed by Python. Regardless of the type of function, you can only define them in class (.bbclass) and recipe (.bb or .inc) files.

Functions written in shell script and executed either directly as functions, tasks, or both. They can also be called by other shell functions. Here is an example shell function definition: some_function () { echo "Hello World" } When you create these types of functions in your recipe or class files, you need to follow the shell programming rules. The scripts are executed by /bin/sh, which may not be a bash shell but might be something such as dash. You should not use Bash-specific script (bashisms).

These functions are written in Python and executed by BitBake or other Python functions using bb.build.exec_func(). An example BitBake function is: python some_python_function () { d.setVar("TEXT", "Hello World") print d.getVar("TEXT", True) } Because the Python "bb" and "os" modules are already imported, you do not need to import these modules. Also in these types of functions, the datastore ("d") is a global variable and is always automatically available.

These functions are written in Python and are executed by other Python code. Examples of Python functions are utility functions that you intend to call from in-line Python or from within other Python functions. Here is an example: def get_depends(d): if d.getVar('SOMECONDITION', True): return "dependencywithcond" else: return "dependency" SOMECONDITION = "1" DEPENDS = "${@get_depends(d)}" This would result in DEPENDS containing dependencywithcond. Here are some things to know about Python functions: Python functions can take parameters. The BitBake datastore is not automatically available. Consequently, you must pass it in as a parameter to the function. The "bb" and "os" Python modules are automatically available. You do not need to import them.

Tasks are BitBake execution units that originate as functions and make up the steps that BitBake needs to run for given recipe. Tasks are only supported in recipe (.bb or .inc) and class (.bbclass) files. By convention, tasks begin with the string "do_". Here is an example of a task that prints out the date: python do_printdate () { import time print time.strftime('%Y%m%d', time.gmtime()) }

Any function can be promoted to a task by applying the addtask command. The addtask command also describes inter-task dependencies. Here is the function from the previous section but with the addtask command promoting it to a task and defining some dependencies: python do_printdate () { import time print time.strftime('%Y%m%d', time.gmtime()) } addtask printdate after do_fetch before do_build In the example, the function is defined and then promoted as a task. The do_printdate task becomes a dependency of the do_build task, which is the default task. And, the do_printdate task is dependent upon the do_fetch task. Execution of the do_build task results in the do_printdate task running first.

When running a task, BitBake tightly controls the execution environment of the build tasks to make sure unwanted contamination from the build machine cannot influence the build. Consequently, if you do want something to get passed into the build task environment, you must take these two steps: Tell BitBake to load what you want from the environment into the datastore. You can do so through the BB_ENV_EXTRAWHITE variable. For example, assume you want to prevent the build system from accessing your $HOME/.ccache directory. The following command tells BitBake to load CCACHE_DIR from the environment into the datastore: export BB_ENV_EXTRAWHITE="$BB_ENV_EXTRAWHITE CCACHE_DIR" Tell BitBake to export what you have loaded into the datastore to the task environment of every running task. Loading something from the environment into the datastore (previous step) only makes it available in the datastore. To export it to the task environment of every running task, use a command similar to the following in your local configuration file local.conf or your distribution configuration file: export CCACHE_DIR A side effect of the previous steps is that BitBake records the variable as a dependency of the build process in things like the setscene checksums. If doing so results in unnecessary rebuilds of tasks, you can whitelist the variable so that the setscene code ignores the dependency when it creates checksums.

Variable flags (varflags) help control a task's functionality and dependencies. BitBake reads and writes varflags to the datastore using the following command forms: <variable> = d.getVarFlags("<variable>") self.d.setVarFlags("FOO", {"func": True}) When working with varflags, the same syntax, with the exception of overrides, applies. In other words, you can set, append, and prepend varflags just like variables. See the "Variable Flag Syntax" section for details. BitBake has a defined set of varflags available for recipes and classes. You can discover the complete set by using grep within a shell and search on the string "VarFlags". Tasks support a number of these flags which control various functionality of the task: dirs: Directories that should be created before the task runs. cleandirs: Empty directories that should created before the task runs. noexec: Marks the tasks as being empty and no execution required. These flags are used as dependency placeholders or used when added tasks need to be subsequently disabled. nostamp: Tells BitBake to not generate a stamp file for a task, which implies the task is always executed. fakeroot: Causes a task to be run in a fakeroot environment, obtained by adding the variables in FAKEROOTENV to the environment. umask: The umask to run the task under. deptask: Controls task build-time dependencies. See the DEPENDS variable and the "Build Dependencies" section for more information. rdeptask: Controls task runtime dependencies. See the RDEPENDS variable, the RRECOMMENDS variable, and the "Runtime Dependencies" section for more information. recrdeptask: Controls task recursive runtime dependencies. See the RDEPENDS variable, the RRECOMMENDS variable, and the "Recursive Dependencies" section for more information. depends: Controls inter-task dependencies. See the DEPENDS variable and the "Inter-Task Dependencies" section for more information. rdepends: Controls inter-task runtime dependencies. See the RDEPENDS variable, the RRECOMMENDS variable, and the "Inter-Task Dependencies" section for more information.

BitBake allows installation of event handlers within recipe and class files. Events are triggered at certain points during operation, such as the beginning of operation against a given .bb, the start of a given task, task failure, task success, and so forth. The intent is to make it easy to do things like email notification on build failure. Following is an example event handler that prints the name of the event and the content of the FILE variable: addhandler myclass_eventhandler python myclass_eventhandler() { from bb.event import getName from bb import data print("The name of the Event is %s" % getName(e)) print("The file we run for is %s" % data.getVar('FILE', e.data, True)) } This event handler gets called every time an event is triggered. A global variable "e" is defined and "e.data" contains an instance of "bb.data". With the getName(e) method, one can get the name of the triggered event. During all builds, the following common events occur: bb.event.ConfigParsed() bb.event.ParseStarted() bb.event.ParseProgress() bb.event.ParseCompleted() bb.event.BuildStarted() bb.build.TaskStarted() bb.build.TaskInvalid() bb.build.TaskFailedSilent() bb.build.TaskFailed() bb.build.TaskSucceeded() bb.event.BuildCompleted() bb.cooker.CookerExit() Here is a list of other events that occur based on specific requests to the server: bb.event.TreeDataPreparationStarted() bb.event.TreeDataPreparationProgress bb.event.TreeDataPreparationCompleted bb.event.DepTreeGenerated bb.event.CoreBaseFilesFound bb.event.ConfigFilePathFound bb.event.FilesMatchingFound bb.event.ConfigFilesFound bb.event.TargetsTreeGenerated

BitBake supports two features that facilitate creating from a single recipe file multiple incarnations of that recipe file where all incarnations are buildable. These features are enabled through the BBCLASSEXTEND and BBVERSIONS variables: BBCLASSEXTEND: This variable is a space separated list of classes used to "extend" the recipe for each variant. Here is an example that results in a second incarnation of the current recipe being available. This second incarnation will have the "native" class inherited. BBCLASSEXTEND = "native" BBVERSIONS: This variable allows a single recipe to build multiple versions of a project from a single recipe file. You can also specify conditional metadata (using the OVERRIDES mechanism) for a single version, or an optionally named range of versions. Here is an example: BBVERSIONS = "1.0 2.0 git" SRC_URI_git = "git://someurl/somepath.git" BBVERSIONS = "1.0.[0-6]:1.0.0+ \ 1.0.[7-9]:1.0.7+" SRC_URI_append_1.0.7+ = "file://some_patch_which_the_new_versions_need.patch;patch=1" The name of the range defaults to the original version of the recipe. For example, in OpenEmbedded, the recipe file foo_1.0.0+.bb creates a default name range of 1.0.0+. This is useful because the range name is not only placed into overrides, but it is also made available for the metadata to use in the variable that defines the base recipe versions for use in file:// search paths (FILESPATH).

To allow for efficient operation given multiple processes executing in parallel, BitBake handles dependencies at the task level. BitBake supports a robust method to handle these dependencies. This section describes several types of dependency mechanisms.

BitBake uses the addtask directive to manage dependencies that are internal to a given recipe file. You can use the addtask directive to indicate when a task is dependent on other tasks or when other tasks depend on that recipe. Here is an example: addtask printdate after do_fetch before do_build In this example, the printdate task is depends on the completion of the do_fetch task. And, the do_build depends on the completion of the printdate task.

BitBake uses the DEPENDS variable to manage build time dependencies. The "deptask" varflag for tasks signifies the task of each item listed in DEPENDS that must complete before that task can be executed. Here is an example: do_configure[deptask] = "do_populate_staging" In this example, the do_populate_staging task of each item in DEPENDS must complete before do_configure can execute.

BitBake uses the PACKAGES, RDEPENDS, and RRECOMMENDS variables to manage runtime dependencies. The PACKAGES variable lists runtime packages. Each of those packages can have RDEPENDS and RRECOMMENDS runtime dependencies. The "rdeptask" flag for tasks is used to signify the task of each item runtime dependency which must have completed before that task can be executed. do_package_write[rdeptask] = "do_package" In the previous example, the do_package task of each item in RDEPENDS must have completed before do_package_write can execute.

BitBake uses the "recrdeptask" flag to manage recursive task dependencies. BitBake looks through the build-time and runtime dependencies of the current recipe, looks through the task's inter-task dependencies, and then adds dependencies for the listed task. Once BitBake has accomplished this, it recursively works through the dependencies of those tasks. Iterative passes continue until all dependencies are discovered and added. You might want to not only have BitBake look for dependencies of those tasks, but also have BitBake look for build-time and runtime dependencies of the dependent tasks as well. If that is the case, you need to reference the task name itself in the task list: do_a[recrdeptask] = "do_a do_b"

BitBake uses the "depends" flag in a more generic form to manage inter-task dependencies. This more generic form allows for inter-dependency checks for specific tasks rather than checks for the data in DEPENDS. Here is an example: do_patch[depends] = "quilt-native:do_populate_staging" In this example, the do_populate_staging task of the target quilt-native must have completed before the do_patch task can execute. The "rdepends" flag works in a similar way but takes targets in the runtime namespace instead of the build-time dependency namespace.

It is often necessary to access variables in the BitBake datastore using Python functions. The Bitbake datastore has an API that allows you this access. Here is a list of available operations: Operation Description d.getVar("X", expand=False) Returns the value of variable "X". Using "expand=True" expands the value. d.setVar("X", value) Sets the variable "X" to "value". d.appendVar("X", value) Adds "value" to the end of the variable "X". d.prependVar("X", value) Adds "value" to the start of the variable "X". d.delVar("X") Deletes the variable "X" from the datastore. d.renameVar("X", "Y") Renames the variable "X" to "Y". d.getVarFlag("X", flag, expand=False) Gets "flag" from the variable "X". Using "expand=True" expands the flag. d.setVarFlag("X", flag, value) Sets "flag" for variable "X" to "value". setVarFlags does not clear previous flags. Think of this operation as addVarFlags. d.appendVarFlag("X", flag, value) Need description. d.prependVarFlag("X", flag, value) Need description. d.delVarFlag("X", flag) Need description. d.setVarFlags("X", flagsdict) Sets the flags specified in the dict() parameter. d.getVarFlags("X") Returns a dict of the flags for the variable "X". d.delVarFlags Deletes all the flags for a variable.

BitBake uses checksums (or signatures) along with the setscene to determine if a task needs to be run. This section describes the process. To help understand how BitBake does this, the section assumes an OpenEmbedded metadata-based example. This list is a place holder of content existed from previous work on the manual. Some or all of it probably needs integrated into the subsections that make up this section. For now, I have just provided a short glossary-like description for each variable. Ultimately, this list goes away. STAMP: The base path to create stamp files. STAMPCLEAN Again, the base path to create stamp files but can use wildcards for matching a range of files for clean operations. BB_STAMP_WHITELIST Lists stamp files that are looked at when the stamp policy is "whitelist". BB_STAMP_POLICY Defines the mode for comparing timestamps of stamp files. BB_HASHCHECK_FUNCTION Specifies the name of the function to call during the "setscene" part of the task's execution in order to validate the list of task hashes. BB_SETSCENE_VERIFY_FUNCTION Specifies a function to call that verifies the list of planned task execution before the main task execution happens. BB_SETSCENE_DEPVALID Specifies a function BitBake calls that determines whether BitBake requires a setscene dependency to be met. BB_TASKHASH Within an executing task, this variable holds the hash of the task as returned by the currently enabled signature generator.

This section needs to get the concept of the setscene across. The reader needs to know what it is and what it is used for during the build process.

A checksum is a unique signature of a task's inputs. The setscene code uses a checksum to determine if a task needs to be run. Because it is a change in a task's inputs that triggers running the task, the process needs to detect all the inputs to a given task. For shell tasks, this turns out to be fairly easy because BitBake generates a "run" shell script for each task and it is possible to create a checksum that gives you a good idea of when the task's data changes. To complicate the problem, some things should not be included in the checksum. First, there is the actual specific build path of a given task - the working directory. It does not matter if the working directory changes because it should not affect the output for target packages. The simplistic approach for excluding the working directory is to set it to some fixed value and create the checksum for the "run" script. Another problem results from the "run" scripts containing functions that might or might not get called. The incremental build solution contains code that figures out dependencies between shell functions. This code is used to prune the "run" scripts down to the minimum set, thereby alleviating this problem and making the "run" scripts much more readable as a bonus. So far we have solutions for shell scripts. What about Python tasks? The same approach applies even though these tasks are more difficult. The process needs to figure out what variables a Python function accesses and what functions it calls. Again, the incremental build solution contains code that first figures out the variable and function dependencies, and then creates a checksum for the data used as the input to the task. Like the working directory case, situations exist where dependencies should be ignored. For these cases, you can instruct the build process to ignore a dependency by using a line like the following: PACKAGE_ARCHS[vardepsexclude] = "MACHINE" This example ensures that the PACKAGE_ARCHS variable does not depend on the value of MACHINE, even if it does reference it. Equally, there are cases where we need to add dependencies BitBake is not able to find. You can accomplish this by using a line like the following: PACKAGE_ARCHS[vardeps] = "MACHINE" This example explicitly adds the MACHINE variable as a dependency for PACKAGE_ARCHS. Consider a case with in-line Python, for example, where BitBake is not able to figure out dependencies. When running in debug mode (i.e. using -DDD), BitBake produces output when it discovers something for which it cannot figure out dependencies. Thus far, this section has limited discussion to the direct inputs into a task. Information based on direct inputs is referred to as the "basehash" in the code. However, there is still the question of a task's indirect inputs - the things that were already built and present in the build directory. The checksum (or signature) for a particular task needs to add the hashes of all the tasks on which the particular task depends. Choosing which dependencies to add is a policy decision. However, the effect is to generate a master checksum that combines the basehash and the hashes of the task's dependencies. At the code level, there are a variety of ways both the basehash and the dependent task hashes can be influenced. Within the BitBake configuration file, we can give BitBake some extra information to help it construct the basehash. The following statement effectively results in a list of global variable dependency excludes - variables never included in any checksum. This example uses variables from OpenEmbedded to help illustrate the concept: BB_HASHBASE_WHITELIST ?= "TMPDIR FILE PATH PWD BB_TASKHASH BBPATH DL_DIR \ SSTATE_DIR THISDIR FILESEXTRAPATHS FILE_DIRNAME HOME LOGNAME SHELL TERM \ USER FILESPATH STAGING_DIR_HOST STAGING_DIR_TARGET COREBASE PRSERV_HOST \ PRSERV_DUMPDIR PRSERV_DUMPFILE PRSERV_LOCKDOWN PARALLEL_MAKE \ CCACHE_DIR EXTERNAL_TOOLCHAIN CCACHE CCACHE_DISABLE LICENSE_PATH SDKPKGSUFFIX" The previous example excludes the work directory, which is part of TMPDIR. The rules for deciding which hashes of dependent tasks to include through dependency chains are more complex and are generally accomplished with a Python function. The code in meta/lib/oe/sstatesig.py shows two examples of this and also illustrates how you can insert your own policy into the system if so desired. This file defines the two basic signature generators OpenEmbedded Core uses: "OEBasic" and "OEBasicHash". By default, there is a dummy "noop" signature handler enabled in BitBake. This means that behavior is unchanged from previous versions. OE-Core uses the "OEBasicHash" signature handler by default through this setting in the bitbake.conf file: BB_SIGNATURE_HANDLER ?= "OEBasicHash" The "OEBasicHash" BB_SIGNATURE_HANDLER is the same as the "OEBasic" version but adds the task hash to the stamp files. This results in any metadata change that changes the task hash, automatically causing the task to be run again. This removes the need to bump PR values, and changes to metadata automatically ripple across the build. It is also worth noting that the end result of these signature generators is to make some dependency and hash information available to the build. This information includes: BB_BASEHASH_task-<taskname>: The base hashes for each task in the recipe. BB_BASEHASH_<filename:taskname>: The base hashes for each dependent task. BBHASHDEPS_<filename:taskname>: The task dependencies for each task. BB_TASKHASH: The hash of the currently running task.