dune files

dune files are the main part of dune. They are used to describe libraries, executables, tests, and everything dune needs to know about.

The syntax of dune files is described in Metadata format section.

Stanzas

dune files are composed of stanzas. For instance a typical dune looks like:

(library
 (name mylib)
 (libraries base lwt))

(rule
 (targets foo.ml)
 (deps    generator/gen.exe)
 (action  (run %{deps} -o %{targets})))

The following sections describe the available stanzas and their meaning.

jbuild_version

Deprecated. This stanza is no longer used and will be removed in the future.

library

The library stanza must be used to describe OCaml libraries. The format of library stanzas is as follows:

library
  (name <library-name>
   <optional-fields>
  )

<library-name> is the real name of the library. It determines the names of the archive files generated for the library as well as the module name under which the library will be available, unless (wrapped false) is used (see below). It must be a valid OCaml module name but doesn’t need to start with a uppercase letter.

For instance, the modules of a library named foo will be available as Foo.XXX outside of foo itself. It is however allowed to write an explicit Foo module, in which case this will be the interface of the library and you are free to expose only the modules you want.

Note that by default libraries and other things that consume OCaml/Reason modules only consume modules from the directory where the stanza appear. In order to declare a multi-directory library, you need to use the include_subdirs stanza.

<optional-fields> are:

  • (public_name <name>) this is the name under which the library can be referred to as a dependency when it is not part of the current workspace, i.e. when it is installed. Without a (public_name ...) field, the library will not be installed by dune. The public name must start by the package name it is part of and optionally followed by a dot and anything else you want. The package name must be one of the packages that dune knows about, as determined by the dune-project files
  • (synopsis <string>) should give a one-line description of the library. This is used by tools that list installed libraries
  • (modules <modules>) specifies what modules are part of the library. By default dune will use all the .ml/.re files in the same directory as the dune file. This include ones that are present in the file system as well as ones generated by user rules. You can restrict this list by using a (modules <modules>) field. <modules> uses the Ordered set language where elements are module names and don’t need to start with a uppercase letter. For instance to exclude module Foo: (modules (:standard \ foo))
  • (libraries <library-dependencies>) is used to specify the dependencies of the library. See the section about Library dependencies for more details
  • (wrapped <boolean>) specifies whether the modules of the library should be available only through the top-level library module, or should all be exposed at the top level. The default is true and it is highly recommended to keep it this way. Because OCaml top-level modules must all be unique when linking an executables, polluting the top-level namespace will make your library unusable with other libraries if there is a module name clash. This option is only intended for libraries that manually prefix all their modules by the library name and to ease porting of existing projects to dune
  • (wrapped (transition <message>)) Is the same as (wrapped true) except that it will also generate unwrapped (not prefixed by the library name) modules to preserve compatibility. This is useful for libraries that would like to transition from (wrapped false) to (wrapped true) without breaking compatibility for users. The <message> will be included in the deprecation notice for the unwrapped modules.
  • (preprocess <preprocess-spec>) specifies how to preprocess files if needed. The default is no_processing. Other options are described in the Preprocessing specification section
  • (preprocessor_deps (<deps-conf list>)) specifies extra dependencies of the preprocessor, for instance if the preprocessor reads a generated file. The specification of dependencies is described in the Dependency specification section
  • (optional), if present it indicates that the library should only be built and installed if all the dependencies are available, either in the workspace or in the installed world. You can use this to provide extra features without adding hard dependencies to your project
  • (c_names (<names>)), if your library has stubs, you must list the C files in this field, without the .c extension
  • (cxx_names (<names>)) is the same as c_names but for C++ stubs
  • (install_c_headers (<names>)), if your library has public C header files that must be installed, you must list them in this field, without the .h extension
  • (modes <modes>) modes which should be built by default. The most common use for this feature is to disable native compilation when writing libraries for the OCaml toplevel. The following modes are available: byte, native and best. best is native or byte when native compilation is not available
  • (no_dynlink) is to disable dynamic linking of the library. This is for advanced use only, by default you shouldn’t set this option
  • (kind <kind>) is the kind of the library. The default is normal, other available choices are ppx_rewriter and ppx_deriver and must be set when the library is intended to be used as a ppx rewriter or a [@@deriving ...] plugin. The reason why ppx_rewriter and ppx_deriver are split is historical and hopefully we won’t need two options soon
  • (ppx_runtime_libraries (<library-names>)) is for when the library is a ppx rewriter or a [@@deriving ...] plugin and has runtime dependencies. You need to specify these runtime dependencies here
  • (virtual_deps (<opam-packages>). Sometimes opam packages enable a specific feature only if another package is installed. This is for instance the case of ctypes which will only install ctypes.foreign if the dummy ctypes-foreign package is installed. You can specify such virtual dependencies here. You don’t need to do so unless you use dune to synthesize the depends and depopts sections of your opam file
  • js_of_ocaml. See the section about js_of_ocaml
  • flags, ocamlc_flags and ocamlopt_flags. See the section about OCaml flags
  • (library_flags (<flags>)) is a list of flags that are passed as it to ocamlc and ocamlopt when building the library archive files. You can use this to specify -linkall for instance. <flags> is a list of strings supporting Variables expansion
  • (c_flags <flags>) specifies the compilation flags for C stubs, using the Ordered set language. This field supports (:include ...) forms
  • (cxx_flags <flags>) is the same as c_flags but for C++ stubs
  • (c_library_flags <flags>) specifies the flags to pass to the C compiler when constructing the library archive file for the C stubs. <flags> uses the Ordered set language and supports (:include ...) forms. When you are writing bindings for a C library named bar, you should typically write -lbar here, or whatever flags are necessary to to link against this library
  • (self_build_stubs_archive <c-libname>) indicates to dune that the library has stubs, but that the stubs are built manually. The aim of the field is to embed a library written in foreign language and/or building with another build system. It is not for casual uses, see the re2 library for an example of use
  • (modules_without_implementation <modules>) specifies a list of modules that have only a .mli or .rei but no .ml or .re file. Such modules are usually referred as mli only modules. They are not officially supported by the OCaml compiler, however they are commonly used. Such modules must only define types. Since it is not reasonably possible for dune to check that this is the case, dune requires the user to explicitly list such modules to avoid surprises. <modules> must be a subset of the modules listed in the (modules ...) field.
  • (private_modules <modules>) species a list of modules that will be marked as private. Private modules are inaccessible from outside the libraries they are defined in.
  • (allow_overlapping_dependencies) allows external dependencies to overlap with libraries that are present in the workspace
  • (no_keep_locs) undocumented, it is a necessary hack until this is implemented: https://github.com/ocaml/dune/issues/921

Note that when binding C libraries, dune doesn’t provide special support for tools such as pkg-config, however it integrates easily with configurator by using (c_flags (:include ...)) and (c_library_flags (:include ...)).

executable

The executable stanza must be used to describe an executable. The format of executable stanzas is as follows:

(executable
  (name <name>)
  <optional-fields>
)

<name> is a module name that contains the main entry point of the executable. There can be additional modules in the current directory, you only need to specify the entry point. Given an executable stanza with (name <name>), dune will know how to build <name>.exe, <name>.bc and <name>.bc.js. <name>.exe is a native code executable, <name>.bc is a bytecode executable which requires ocamlrun to run and <name>.bc.js is a JavaScript generated using js_of_ocaml.

Note that in case native compilation is not available, <name>.exe will in fact be a custom byte-code executable. Custom in the sense of ocamlc -custom, meaning that it is a native executable that embeds the ocamlrun virtual machine as well as the byte code. As such you can always rely on <name>.exe being available. Moreover, it is usually preferable to use <name>.exe in custom rules or when calling the executable by hand. This is because running a byte-code executable often requires loading shared libraries that are locally built, and so requires additional setup such as setting specific environment variables and dune doesn’t do at the moment.

Native compilation is considered not available when there is no ocamlopt binary at the same place as where ocamlc was found.

Executables can also be linked as object or shared object files. See linking modes for more information.

<optional-fields> are:

  • (public_name <public-name>) specifies that the executable should be installed under that name. It is the same as adding the following stanza to your dune file:

    (install
     (section bin)
     (files (<name>.exe as <public-name>)))
    
  • (package <package>) if there is a (public_name ...) field, this specifies the package the executables are part of
  • (libraries <library-dependencies>) specifies the library dependencies. See the section about Library dependencies for more details
  • (link_flags <flags>) specifies additional flags to pass to the linker. This field supports (:include ...) forms
  • (link_deps (<deps-conf list>)) specifies the dependencies used only by the linker, for example when using a version script. See the Dependency specification section for more details.
  • (modules <modules>) specifies which modules in the current directory dune should consider when building this executable. Modules not listed here will be ignored and cannot be used inside the executable described by the current stanza. It is interpreted in the same way as the (modules ...) field of library
  • (modes (<modes>)) sets the linking modes. The default is (byte exe)
  • (preprocess <preprocess-spec>) is the same as the (preprocess ...) field of library
  • (preprocessor_deps (<deps-conf list>)) is the same as the (preprocessor_deps ...) field of library
  • js_of_ocaml. See the section about js_of_ocaml
  • flags, ocamlc_flags and ocamlopt_flags. See the section about
    specifying OCaml flags
  • (modules_without_implementation <modules>) is the same as the corresponding field of library
  • (allow_overlapping_dependencies) is the same as the corresponding field of library

Linking modes

The modes field allows to select what linking modes should be used to link executables. Each mode is a pair (<compilation-mode> <binary-kind>) where <compilation-mode> describes whether the byte code or native code backend of the OCaml compiler should be used and <binary-kind> describes what kind of file should be produced.

<compilation-mode> must be byte, native or best, where best is native with a fallback to byte-code when native compilation is not available.

<binary-kind> is one of:

= c for producing OCaml bytecode embedded in a C file - exe for normal executables - object for producing static object files that can be manually

linked into C applications
  • shared_object for producing object files that can be dynamically loaded into an application. This mode can be used to write a plugin in OCaml for a non-OCaml application.

For instance the following executables stanza will produce byte code executables and native shared objects:

(executables
 ((names (a b c))
  (modes ((byte exe) (native shared_object)))))

Additionally, you can use the following short-hands:

  • c for (byte c)
  • exe for (best exe)
  • object for (best object)
  • shared_object for (best shared_object)
  • byte for (byte exe)
  • native for (native exe)

For instance the following modes fields are all equivalent:

(modes (exe object shared_object))
(modes ((best exe)
        (best object)
        (best shared_object)))

The extensions for the various linking modes are chosen as follows:

compilation mode binary kind extensions
byte exe .bc and .bc.js
native/best exe .exe
byte object .bc%{ext_obj}
native/best object .exe%{ext_obj}
byte shared_object .bc%{ext_dll}
native/best shared_object %{ext_dll}
byte c .bc.c

Where %{ext_obj} and %{ext_dll} are the extensions for object and shared object files. Their value depends on the OS, for instance on Unix %{ext_obj} is usually .o and %{ext_dll} is usually .so while on Windows %{ext_obj} is .obj and %{ext_dll} is .dll.

Note that when (byte exe) is specified but neither (best exe) nor (native exe) are specified, Jbuilkd still knows how to build an executable with the extension .exe. In such case, the .exe version is the same as the .bc one except that it is linked with the -custom option of the compiler. You should always use the .exe rather that the .bc inside build rules.

executables

The executables stanza is the same as the executable stanza, except that it is used to describe several executables sharing the same configuration.

It shares the same fields as the executable stanza, except that instead of (name ...) and (public_name ...) you must use:

  • (names <names>) where <names> is a list of entry point names. As for executable you only need to specify the modules containing the entry point of each executable
  • (public_names <names>) describes under what name each executable should be installed. The list of names must be of the same length as the list in the (names ...) field. Moreover you can use - for executables that shouldn’t be installed

rule

The rule stanza is used to create custom user rules. It tells dune how to generate a specific set of files from a specific set of dependencies.

The syntax is as follows:

(rule
  (targets <filenames>)
  (action  <action>)
  <optional-fields>)

<filenames> is a list of file names. Note that currently dune only support user rules with targets in the current directory.

<action> is the action to run to produce the targets from the dependencies. See the User actions section for more details.

<optional-fields> are:

  • (deps <deps-conf list>) to specify the dependencies of the rule. See the Dependency specification section for more details.
  • (mode <mode>) to specify how to handle the targets, see modes for details
  • (fallback) is deprecated and is the same as (mode fallback)
  • (locks (<lock-names>)) specify that the action must be run while holding the following locks. See the Locks section for more details.

Note that contrary to makefiles or other build systems, user rules currently don’t support patterns, such as a rule to produce %.y from %.x for any given %. This might be supported in the future.

modes

By default, the target of a rule must not exist in the source tree and dune will error out when this is the case.

However, it is possible to change this behavior using the mode field. The following modes are available:

  • standard, this is the standard mode
  • fallback, in this mode, when the targets are already present in the source tree, dune will ignore the rule. It is an error if only a subset of the targets are present in the tree. The common use of fallback rules is to generate default configuration files that may be generated by a configure script.
  • promote, in this mode, the files in the source tree will be ignored. Once the rule has been executed, the targets will be copied back to the source tree
  • promote-until-clean is the same as promote except than dune clean will remove the promoted files from the source tree

There are two use cases for promote rules. The first one is when the generated code is easier to review than the generator, so it’s easier to commit the generated code and review it. The second is to cut down dependencies during releases: by passing --ignore-promoted-rules to dune, rules will (mode promote) will be ignored and the source files will be used instead. The -p/--for-release-of-packages flag implies --ignore-promote-rules.

inferred rules

When using the action DSL (see User actions), it is most of the time obvious what are the dependencies and targets.

For instance:

(rule
 (targets b)
 (deps    a)
 (action  (copy %{deps} %{targets})))

In this example it is obvious by inspecting the action what the dependencies and targets are. When this is the case you can use the following shorter syntax, where dune infers dependencies and targets for you:

(rule <action>)

For instance:

(rule (copy a b))

Note that in dune, targets must always be known statically. Especially, this mean that dune must be able to statically determine all targets. For instance, this (rule ...) stanza is rejected by dune:

(rule (copy a b.%{read:file}))

ocamllex

(ocamllex <names>) is essentially a shorthand for:

(rule
  (targets <name>.ml)
  (deps    <name>.mll)
  (action  (chdir %{workspace_root}
            (run %{bin:ocamllex} -q -o %{targets} %{deps}))))

To use a different rule mode, use the long form:

(ocamllex
  (modules <names>)
  (mode    <mode>))

ocamlyacc

(ocamlyacc <names>) is essentially a shorthand for:

(rule
  (targets <name>.ml <name>.mli)
  (deps    <name>.mly)
  (action  (chdir %{workspace_root}
            (run %{bin:ocamlyacc} %{deps}))))

To use a different rule mode, use the long form:

(ocamlyacc
  (modules <names>)
  (mode    <mode>))

menhir

A menhir stanza is available to support the menhir parser generator. See the Menhir section for details.

alias

The alias stanza lets you add dependencies to an alias, or specify an action to run to construct the alias.

The syntax is as follows:

(alias
 (name    <alias-name>)
 (deps    <deps-conf list>)
  <optional-fields>)

<name> is an alias name such as runtest.

<deps-conf list> specifies the dependencies of the alias. See the Dependency specification section for more details.

<optional-fields> are:

  • <action>, an action to run when constructing the alias. See the User actions section for more details.
  • (package <name>) indicates that this alias stanza is part of package <name> and should be filtered out if <name> is filtered out from the command line, either with --only-packages <pkgs> or -p <pkgs>
  • (locks (<lock-names>)) specify that the action must be run while holding the following locks. See the Locks section for more details.
  • (enabled_if <blang expression>) specifies the boolean condition that must be true for the tests to run. The condition is specified using the blang, and the field allows for variables to appear in the expressions.

The typical use of the alias stanza is to define tests:

(alias
 (name   runtest)
 (action (run %{exe:my-test-program.exe} blah)))

See the section about Running tests for details.

Note that if your project contains several packages and you run test the tests from the opam file using a build-test field, then all your runtest alias stanzas should have a (package ...) field in order to partition the set of tests.

install

The install stanza is what lets you describe what dune should install, either when running dune install or through opam.

Libraries and executables don’t need an install stanza to be installed, just a public_name field. Everything else needs an install stanza.

The syntax is as follows:

(install
  (section <section>)
   (files   <filenames>)
   <optional-fields>)

<section> is the installation section, as described in the opam manual. The following sections are available:

  • lib
  • lib_root
  • libexec
  • libexec_root
  • bin
  • sbin
  • toplevel
  • share
  • share_root
  • etc
  • doc
  • stublibs
  • man
  • misc

<files> is the list of files to install. Each element in the list must be either a literal filename or a S-expression of the form:

(<filename> as <destination>)

where <destination> describe how the file will be installed. For instance, to install a file mylib.el as emacs/site-lisp/mylib.el in the share_root section:

(install
 (section share_root)
 (files   (mylib.el as emacs/site-lisp/mylib.el)))

<optional-fields> are:

  • (package <name>). If there are no ambiguities, you can omit this field. Otherwise you need it to specify which package these files are part of. The package is not ambiguous when the first parent directory to contain a <package>.opam file contains exactly one <package>.opam file

Handling of the .exe extension on Windows

Under Microsoft Windows, executables must be suffixed with .exe. Dune tries to make sure that executables are always installed with this extension on Windows.

More precisely, when installing a file via an (install ...) stanza, if the source file has extension .exe or .bc, then dune implicitly adds the .exe extension to the destination, if not already present.

copy_files

The copy_files and copy_files# stanzas allow to specify that files from another directory could be copied if needed to the current directory.

The syntax is as follows:

(copy_files <glob>)

<glob> represents the set of files to copy, see the glob for details.

The difference between copy_files and copy_files# is the same as the difference between the copy and copy# action. See the User actions section for more details.

include

The include stanza allows to include the contents of another file into the current dune file. Currently, the included file cannot be generated and must be present in the source tree. This feature is intended to be used in conjunction with promotion, when parts of a dune file are to be generated.

For instance:

(include dune.inc)

(rule (with-stdout-to dune.inc.gen (run ./gen-dune.exe)))

(alias
 (name   runtest)
 (action (diff dune.inc dune.inc.gen)))

With this dune file, running dune as follow will replace the dune.inc file in the source tree by the generated one:

$ dune build @runtest --auto-promote

tests

The tests stanza allows one to easily define multiple tests. For example we can define two tests at once with:

(tests
 (names mytest expect_test)
 <optional fields>)

This will define an executable named mytest.exe that will be executed as part of the runtest alias. If the directory also contains an expect_test.expected file, then expect_test will be used to define an expect test. That is, the test will be executed and its output will be compared to expect_test.expected.

The optional fields that are supported are a subset of the alias and executables fields. In particular, all fields except for public_names are supported from the executables stanza. Alias fields apart from name are allowed.

test

The test stanza is the singular form of tests. The only difference is that it’s of the form:

(test
 (name foo)
 <optional fields>)

where the name field is singular. The same optional fields are supported.

env

The env stanza allows to modify the environment. The syntax is as follow:

(env
 (<profile1> <settings1>)
 (<profile2> <settings2>)
 ...
 (<profilen> <settingsn>))

The first form (<profile> <settings>) that correspond to the selected build profile will be used to modify the environment in this directory. You can use _ to match any build profile.

Currently <settings> can be any OCaml flags field, see OCaml flags for more details.

ignored_subdirs

The ignored_subdirs stanza allows to tell Dune to ignore one or more sub-directories. The syntax is as follow:

(ignored_subdirs (<sub-dir1> <sub-dir2> ...))

A directory that is ignored will not be eagerly scanned by Dune. Any dune or other special files in it won’t be interpreted either and will be treated as raw data. It is however possible to depend on files inside ignored sub-directories.

include_subdirs

The include_subdirs stanza is used to control how dune considers sub-directories of the current directory. The syntax is as follow:

(include_subdirs <mode>)

Where <mode> maybe be one of:

  • no, the default
  • unqualified

When the include_subdirs stanza is not present or <mode> is no, dune considers sub-directories as independent. When <mode> is unqualified, dune will assume that the sub-directories of the current directory are part of the same group of directories. In particular, dune will scan all these directories at once when looking for OCaml/Reason files. This allows you to split a library between several directories. unqualified means that modules in sub-directories are seen as if they were all in the same directory. In particular, you cannot have two modules with the same name in two different directories. It is planned to add a qualified mode in the future.

Note that sub-directories are included recursively, however the recursion will stop when encountering a sub-directory that contains another include_subdirs stanza. Additionally, it is not allowed for a sub-directory of a directory with (include_subdirs <x>) where <x> is not no to contain one of the following stanzas:

  • library
  • executable(s)
  • test(s)

Common items

Ordered set language

A few fields takes as argument an ordered set and can be specified using a small DSL.

This DSL is interpreted by dune into an ordered set of strings using the following rules:

  • :standard denotes the standard value of the field when it is absent
  • an atom not starting with a : is a singleton containing only this atom
  • a list of sets is the concatenation of its inner sets
  • (<sets1> \ <sets2>) is the set composed of elements of <sets1> that do not appear in <sets2>

In addition, some fields support the inclusion of an external file using the syntax (:include <filename>). This is useful for instance when you need to run a script to figure out some compilation flags. <filename> is expected to contain a single S-expression and cannot contain (:include ...) forms.

Note that inside an ordered set, the first element of a list cannot be an atom except if it starts with - or :. The reason for this is that we are planning to add simple programmatic features in the futures so that one may write:

(flags (if (>= %{ocaml_version} 4.06) ...))

This restriction will allow to add this feature without introducing a breaking changes. If you want to write a list where the first element doesn’t start by -, you can simply quote it: ("x" y z).

Most fields using the ordered set language also support Variables expansion. Variables are expanded after the set language is interpreted.

Boolean Language

The boolean language allows the user to define simple boolean expressions that dune can evaluate. Here’s a semi formal specification of the language:

op := '=' | '<' | '>' | '<>' | '>=' | '<='

expr := (and <expr>+)
      | (or <expr>+)
      | (<op> <template> <template>)
      | <template>

After an expression is evaluated, it must be exactly the string true or false to be considered as a boolean. Any other value will be treated as an error.

Here’s a simple example of a condition that expresses running on OSX and having an flambda compiler with the help of variable expansion:

(and %{ocamlc-config:flambda} (= %{ocamlc-config:system} macosx))

Variables expansion

Some fields can contains variables of the form %{var} that are expanded by dune.

Dune supports the following variables:

  • project_root is the root of the current project. It is typically the toplevel directory of your project and as long as you have a dune-project file there, project_root is independent of the workspace configuration
  • workspace_root is the root of the current workspace. Note that the value of workspace_root is not constant and depends on whether your project is vendored or not
  • CC is the C compiler command line (list made of the compiler name followed by its flags) that was used to compile OCaml in the current build context
  • CXX is the C++ compiler command line being used in the current build context
  • ocaml_bin is the path where ocamlc lives
  • ocaml is the ocaml binary
  • ocamlc is the ocamlc binary
  • ocamlopt is the ocamlopt binary
  • ocaml_version is the version of the compiler used in the current build context
  • ocaml_where is the output of ocamlc -where
  • arch_sixtyfour is true if using a compiler targeting a 64 bit architecture and false otherwise
  • null is /dev/null on Unix or nul on Windows
  • ext_obj, ext_asm, ext_lib, ext_dll and ext_exe are the file extension used for various artifacts
  • ocaml-config:v for every variable v in the output of ocamlc -config. Note that dune processes the output of ocamlc -config in order to make it a bit more stable across versions, so the exact set of variables accessible this way might not be exactly the same as what you can see in the output of ocamlc -config. In particular, variables added in new versions of OCaml needs to be registered in dune before they can be used
  • profile the profile selected via --profile

In addition, (action ...) fields support the following special variables:

  • targets expands to the list of target
  • deps expands to the list of dependencies
  • ^ expands to the list of dependencies, separated by spaces
  • dep:<path> expands to <path> (and adds <path> as a dependency of the action)
  • exe:<path> is the same as <path>, except when cross-compiling, in which case it will expand to <path> from the host build context
  • bin:<program> expands to a path to program. If program is installed by a package in the workspace (see install stanzas), the locally built binary will be used, otherwise it will be searched in the PATH of the current build context. Note that (run %{bin:program} ...) and (run program ...) behave in the same way. %{bin:...} is only necessary when you are using (bash ...) or (system ...)
  • lib:<public-library-name>:<file> expands to a path to file <file> of library <public-library-name>. If <public-library-name> is available in the current workspace, the local file will be used, otherwise the one from the installed world will be used
  • libexec:<public-library-name>:<file> is the same as lib:... except when cross-compiling, in which case it will expand to the file from the host build context
  • lib-available:<library-name> expands to true or false depending on whether the library is available or not. A library is available iff at least one of the following condition holds:
    • it is part the installed worlds
    • it is available locally and is not optional
    • it is available locally and all its library dependencies are available
  • version:<package> expands to the version of the given package. Note that this is only supported for packages that are being defined in the current scope
  • read:<path> expands to the contents of the given file
  • read-lines:<path> expands to the list of lines in the given file
  • read-strings:<path> expands to the list of lines in the given file, unescaped using OCaml lexical convention

The %{<kind>:...} forms are what allows you to write custom rules that work transparently whether things are installed or not.

Note that aliases are ignored by %{deps}

The intent of this last form is to reliably read a list of strings generated by an OCaml program via:

List.iter (fun s -> print_string (String.escaped s)) l
  1. Expansion of lists

Forms that expands to list of items, such as %{cc}, %{deps}, %{targets} or %{read-lines:...}, are suitable to be used in, say, (run <prog> <arguments>). For instance in:

(run foo %{deps})

if there are two dependencies a and b, the produced command will be equivalent to the shell command:

$ foo "a" "b"

If you want the two dependencies to be passed as a single argument, you have to quote the variable as in:

(run foo "%{deps}")

which is equivalent to the following shell command:

$ foo "a b"

(the items of the list are concatenated with space). Note that, since %{deps} is a list of items, the first one may be used as a program name, for instance:

(rule
 (targets result.txt)
 (deps    foo.exe (glob_files *.txt))
 (action  (run %{deps})))

Here is another example:

(rule
 (targets foo.exe)
 (deps    foo.c)
 (action  (run %{cc} -o %{targets} %{deps} -lfoolib)))

Library dependencies

Dependencies on libraries are specified using (libraries ...) fields in library and executables stanzas.

For libraries defined in the current scope, you can use either the real name or the public name. For libraries that are part of the installed world, or for libraries that are part of the current workspace but in another scope, you need to use the public name. For instance: (libraries base re).

When resolving libraries, libraries that are part of the workspace are always preferred to ones that are part of the installed world.

Alternative dependencies

In addition to direct dependencies you can specify alternative dependencies. This is described in the Alternative dependencies section

It is sometimes the case that one wants to not depend on a specific library, but instead on whatever is already installed. For instance to use a different backend depending on the target.

Dune allows this by using a (select ... from ...) form inside the list of library dependencies.

Select forms are specified as follows:

(select <target-filename> from
  (<literals> -> <filename>)
  (<literals> -> <filename>)
   ...)

<literals> are lists of literals, where each literal is one of:

  • <library-name>, which will evaluate to true if <library-name> is available, either in the workspace or in the installed world
  • !<library-name>, which will evaluate to true if <library-name> is not available in the workspace or in the installed world

When evaluating a select form, dune will create <target-filename> by copying the file given by the first (<literals> -> <filename>) case where all the literals evaluate to true. It is an error if none of the clauses are selectable. You can add a fallback by adding a clause of the form (-> <file>) at the end of the list.

Preprocessing specification

Dune accepts three kinds of preprocessing:

  • no_preprocessing, meaning that files are given as it to the compiler, this is the default
  • (action <action>) to preprocess files using the given action
  • (pps <ppx-rewriters-and-flags>) to preprocess files using the given list of ppx rewriters
  • (staged_pps <ppx-rewriters-and-flags>) is similar to (pps ...) but behave slightly differently and is needed for certain ppx rewriters (see below for details)

Dune normally assumes that the compilation pipeline is sequenced as follow:

  • code generation (including preprocessing)
  • dependency analysis
  • compilation

Dune uses this fact to optimize the pipeline and in particular share the result of code generation and preprocessing between the dependency analysis and compilation phases. However, some specific code generators or preprocessors require feedback from the compilation phase. As a result they must be applied in stages as follows:

  • first stage of code geneneration
  • dependency analysis
  • second step of code generation in parallel with compilation

This is the case for ppx rewriters using the OCaml typer for instance. When using such ppx rewriters, you must use staged_pps instead of pps in order to force Dune to use the second pipeline, which is slower but necessary in this case.

Preprocessing with actions

<action> uses the same DSL as described in the User actions section, and for the same reason given in that section, it will be executed from the root of the current build context. It is expected to be an action that reads the file given as only dependency named input-file and outputs the preprocessed file on its standard output.

More precisely, (preprocess (action <action>)) acts as if you had setup a rule for every file of the form:

(rule
 (targets file.pp.ml)
 (deps    file.ml)
 (action  (with-stdout-to %{targets}
           (chdir %{workspace_root} <action>))))

The equivalent of a -pp <command> option passed to the OCaml compiler is (system "<command> %{input-file}").

Preprocessing with ppx rewriters

<ppx-rewriters-and-flags> is expected to be a sequence where each element is either a command line flag if starting with a - or the name of a library. If you want to pass command line flags that do not start with a -, you can separate library names from flags using --. So for instance from the following preprocess field:

(preprocess (pps ppx1 -foo ppx2 -- -bar 42))

The list of libraries will be ppx1 and ppx2 and the command line arguments will be: -foo -bar 42.

Libraries listed here should be libraries implementing an OCaml AST rewriter and registering themselves using the ocaml-migrate-parsetree.driver API.

Dune will build a single executable by linking all these libraries and their dependencies. Note that it is important that all these libraries are linked with -linkall. Dune automatically uses -linkall when the (kind ...) field is set to ppx_rewriter or ppx_deriver.

Per module preprocessing specification

By default a preprocessing specification will apply to all modules in the library/set of executables. It is possible to select the preprocessing on a module-by-module basis by using the following syntax:

(preprocess (per_module
               (<spec1> (<module-list1>))
               (<spec2> (<module-list2>))
               ...))

Where <spec1>, <spec2>, … are preprocessing specifications and <module-list1>, <module-list2>, … are list of module names.

For instance:

(preprocess (per_module
               (((action (run ./pp.sh X=1 %{input-file})) (foo bar)))
               (((action (run ./pp.sh X=2 %{input-file})) (baz)))))

Dependency specification

Dependencies in dune files can be specified using one of the following syntax:

  • (:name <dependencies>) will bind the the list of dependencies to the name variable. This variable will be available as %{name} in actions.
  • (file <filename>) or simply <filename>: depend on this file
  • (alias <alias-name>): depend on the construction of this alias, for instance: (alias src/runtest)
  • (alias_rec <alias-name>): depend on the construction of this alias recursively in all children directories wherever it is defined. For instance: (alias_rec src/runtest) might depend on (alias src/runtest), (alias src/foo/bar/runtest), …
  • (glob_files <glob>): depend on all files matched by <glob>, see the glob for details
  • (source_tree <dir>): depend on all source files in the subtree with root <dir>
  • (universe): depend on everything in the universe. This is for cases where dependencies are too hard to specify. Note that dune will not be able to cache the result of actions that depend on the universe. In any case, this is only for dependencies in the installed world, you must still specify all dependencies that come from the workspace.
  • (package <pkg>) depend on all files installed by <package>, as well as on the transitive package dependencies of <package>. This can be used to test a command against the files that will be installed
  • (env <var>): depend on the value of the environment variable <var>. If this variable becomes set, becomes unset, or changes value, the target will be rebuilt.

In all these cases, the argument supports Variables expansion.

Named Dependencies

dune allows a user to organize dependency lists by naming them. The user is allowed to assign a group of dependencies a name that can later be referred to in actions (like the %{deps} and %{targets} built in variables).

One instance where this is useful is for naming globs. Here’s an example of an imaginary bundle command:

(rule
 (targets archive.tar)
 (deps
  index.html
  (:css (glob_files *.css))
  (:js foo.js bar.js)
  (:img (glob_files *.png) (glob_files *.jpg)))
 (action
  (run %{bin:bundle} index.html -css %{css} -js %{js} -img %{img} -o %{targets})))

Note that such named dependency list can also include unnamed dependencies (like index.html in the example above). Also, such user defined names wil shadow built in variables. So (:workspace_root x) will shadow the built in %{workspace_root} variable.

Glob

You can use globs to declare dependencies on a set of files. Note that globs will match files that exist in the source tree as well as buildable targets, so for instance you can depend on *.cmi.

Currently dune only support globbing files in a single directory. And in particular the glob is interpreted as follows:

  • anything before the last / is taken as a literal path
  • anything after the last /, or everything if the glob contains no /, is interpreted using the glob syntax

The glob syntax is interpreted as follows:

  • \<char> matches exactly <char>, even if it is a special character (*, ?, …)
  • * matches any sequence of characters, except if it comes first in which case it matches any character that is not . followed by anything
  • ** matches any character that is not . followed by anything, except if it comes first in which case it matches anything
  • ? matches any single character
  • [<set>] matches any character that is part of <set>
  • [!<set>] matches any character that is not part of <set>
  • {<glob1>,<glob2>,...,<globn>} matches any string that is matched by one of <glob1>, <glob2>, …

OCaml flags

In library, executable, executables and env stanzas, you can specify OCaml compilation flags using the following fields:

  • (flags <flags>) to specify flags passed to both ocamlc and ocamlopt
  • (ocamlc_flags <flags>) to specify flags passed to ocamlc only
  • (ocamlopt_flags <flags>) to specify flags passed to ocamlopt only

For all these fields, <flags> is specified in the Ordered set language. These fields all support (:include ...) forms.

The default value for (flags ...) is taken from the environment, as a result it is recommended to write (flags ...) fields as follows:

(flags (:standard <my options>))

js_of_ocaml

A js_of_ocaml field exists in executable and libraries stanzas that allows one to customize options relevant to jsoo.

User actions

(action ...) fields describe user actions.

User actions are always run from the same subdirectory of the current build context as the dune file they are defined in. So for instance an action defined in src/foo/dune will be run from _build/<context>/src/foo.

The argument of (action ...) fields is a small DSL that is interpreted by dune directly and doesn’t require an external shell. All atoms in the DSL support Variables expansion. Moreover, you don’t need to specify dependencies explicitly for the special %{<kind>:...} forms, these are recognized and automatically handled by dune.

The DSL is currently quite limited, so if you want to do something complicated it is recommended to write a small OCaml program and use the DSL to invoke it. You can use shexp to write portable scripts or configurator for configuration related tasks.

The following constructions are available:

  • (run <prog> <args>) to execute a program. <prog> is resolved locally if it is available in the current workspace, otherwise it is resolved using the PATH
  • (chdir <dir> <DSL>) to change the current directory
  • (setenv <var> <value> <DSL>) to set an environment variable
  • (with-<outputs>-to <file> <DSL>) to redirect the output to a file, where <outputs> is one of: stdout, stderr or outputs (for both stdout and stderr)
  • (ignore-<outputs> <DSL) to ignore the output, where <outputs> is one of: stdout, stderr or outputs
  • (progn <DSL>...) to execute several commands in sequence
  • (echo <string>) to output a string on stdout
  • (write-file <file> <string>) writes <string> to <file>
  • (cat <file>) to print the contents of a file to stdout
  • (copy <src> <dst>) to copy a file
  • (copy# <src> <dst>) to copy a file and add a line directive at the beginning
  • (system <cmd>) to execute a command using the system shell: sh on Unix and cmd on Windows
  • (bash <cmd>) to execute a command using /bin/bash. This is obviously not very portable
  • (diff <file1> <file2>) is similar to (run diff <file1> <file2>) but is better and allows promotion. See Diffing and promotion for more details
  • (diff? <file1> <file2>) is the same as (diff <file1> <file2>) except that it is ignored when <file1> or <file2> doesn’t exists
  • (cmp <file1> <file2>) is similar to (run cmp <file1> <file2>) but allows promotion. See Diffing and promotion for more details

As mentioned copy# inserts a line directive at the beginning of the destination file. More precisely, it inserts the following line:

# 1 "<source file name>"

Most languages recognize such lines and update their current location, in order to report errors in the original file rather than the copy. This is important as the copy exists only under the _build directory and in order for editors to jump to errors when parsing the output of the build system, errors must point to files that exist in the source tree. In the beta versions of dune, copy# was called copy-and-add-line-directive. However, most of time one wants this behavior rather than a bare copy, so it was renamed to something shorter.

Note: expansion of the special %{<kind>:...} is done relative to the current working directory of the part of the DSL being executed. So for instance if you have this action in a src/foo/dune:

(action (chdir ../../.. (echo %{path:dune})))

Then %{path:dune} will expand to src/foo/dune. When you run various tools, they often use the filename given on the command line in error messages. As a result, if you execute the command from the original directory, it will only see the basename.

To understand why this is important, let’s consider this dune file living in src/foo:

(rule
 (targets blah.ml)
 (deps    blah.mll)
 (action  (run ocamllex -o %{targets} %{deps})))

Here the command that will be executed is:

ocamllex -o blah.ml blah.mll

And it will be executed in _build/<context>/src/foo. As a result, if there is an error in the generated blah.ml file it will be reported as:

File "blah.ml", line 42, characters 5-10:
Error: ...

Which can be a problem as you editor might think that blah.ml is at the root of your project. What you should write instead is:

(rule
 (targets blah.ml)
 (deps    blah.mll)
 (action  (chdir %{workspace_root} (run ocamllex -o %{targets} %{deps}))))

Locks

Given two rules that are independent, dune will assume that there associated action can be run concurrently. Two rules are considered independent if none of them depend on the other, either directly or through a chain of dependencies. This basic assumption allows to parallelize the build.

However, it is sometimes the case that two independent rules cannot be executed concurrently. For instance this can happen for more complicated tests. In order to prevent dune from running the actions at the same time, you can specify that both actions take the same lock:

(alias
 (name   runtest)
 (deps   foo)
 (locks  m)
 (action (run test.exe %{deps})))

(alias
 (name   runtest)
 (deps   bar)
 (locks  m)
 (action (run test.exe %{deps})))

Dune will make sure that the executions of test.exe foo and test.exe bar are serialized.

Although they don’t live in the filesystem, lock names are interpreted as file names. So for instance (with-lock m ...) in src/dune and (with-lock ../src/m) in test/dune refer to the same lock.

Note also that locks are per build context. So if your workspace has two build contexts setup, the same rule might still be executed concurrently between the two build contexts. If you want a lock that is global to all build contexts, simply use an absolute filename:

(alias
 (name   runtest)
 (deps   foo)
 (locks  /tcp-port/1042)
 (action (run test.exe %{deps})))

Diffing and promotion

(diff <file1> <file2>) is very similar to (run diff <file1> <file2>). In particular it behaves in the same way:

  • when <file1> and <file2> are equal, it doesn’t nothing
  • when they are not, the differences are shown and the action fails

However, it is different for the following reason:

  • the exact command used to diff files can be configured via the --diff-command command line argument. Note that it is only called when the files are not byte equals

  • by default, it will use patdiff if it is installed. patdiff is a better diffing program. You can install it via opam with:

    $ opam install patdiff
    
  • on Windows, both (diff a b) and (diff? a b) normalize the end of lines before comparing the files

  • since (diff a b) is a builtin action, dune knowns that a and b are needed and so you don’t need to specify them explicitly as dependencies

  • you can use (diff? a b) after a command that might or might not produce b. For cases where commands optionally produce a corrected file

  • it allows promotion. See below

Note that (cmp a b) does no end of lines normalization and doesn’t print a diff when the files differ. cmp is meant to be used with binary files.

Promotion

Whenever an action (diff <file1> <file2>) or (diff?  <file1> <file2>) fails because the two files are different, dune allows you to promote <file2> as <file1> if <file1> is a source file and <file2> is a generated file.

More precisely, let’s consider the following dune file:

(rule
 (with-stdout-to data.out (run ./test.exe)))

(alias
 (name   runtest)
 (action (diff data.expected data.out)))

Where data.expected is a file committed in the source repository. You can use the following workflow to update your test:

  • update the code of your test
  • run dune runtest. The diff action will fail and a diff will be printed
  • check the diff to make sure it is what you expect
  • run dune promote. This will copy the generated data.out file to data.expected directly in the source tree

You can also use dune runtest --auto-promote which will automatically do the promotion.

OCaml syntax

If a dune file starts with (* -*- tuareg -*- *), then it is interpreted as an OCaml script that generates the dune file as described in the rest of this section. The code in the script will have access to a Jbuild_plugin module containing details about the build context it is executed in.

The OCaml syntax gives you an escape hatch for when the S-expression syntax is not enough. It is not clear whether the OCaml syntax will be supported in the long term as it doesn’t work well with incremental builds. It is possible that it will be replaced by just an include stanza where one can include a generated file.

Consequently you must not build complex systems based on it.