The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.

NAME

Clownfish::Docs::BuildingProjects - Building Apache Clownfish projects in C environments

DESCRIPTION

The following steps are required to build a project that uses Apache Clownfish:

  • Run the cfc command-line application (Clownfish compiler) to generate C headers, C source files, and documentation.

  • Include the C headers generated by the Clownfish compiler in your own code to work with Clownfish objects.

  • Compile an additional C source file generated by CFC.

  • Link your object code, the object code from the generated file, and the Clownfish runtime library.

Running cfc

    cfc [--source=<dir>] [--include=<dir>] [--parcel=<name>]
        --dest=<dir>
        [--header=<file>] [--footer=<file>]

–source

Every --source argument adds a directory to the list of source directories to search for source parcels. For every source parcel found, CFC generates C headers, a C source file, and documentation.

The source directories are scanned for Clownfish parcel definitions (.cfp), Clownfish headers (.cfh), and standalone documentation in Markdown format (.md).

This option may be specified multiple times.

–include

Every --include argument adds a directory to the list of include directories. CFC generates C headers for parcels from an include directory only if they’re required by a source parcel or if they’re specified with the --parcel option. Only C headers are generated for included parcels, no C source code or documentation.

The include directories are scanned for Clownfish parcel definitions (.cfp) and Clownfish headers (.cfh).

After the directories specified on the command-line, the following directories are processed:

  • Directories from the environment variable CLOWNFISH_INCLUDE. This variable contains a colon-separated list of directories.

  • On UNIX-like systems /usr/local/share/clownfish/include and /usr/share/clownfish/include.

This option may be specified multiple times.

–parcel

Adds a parcel to the list of prerequisite parcels to make sure that its C headers are generated. This is useful when running without the --source option. For example, a project that doesn’t define its own Clownfish classes can generate the C headers for the Clownfish runtime with:

    cfc --parcel=Clownfish --dest=autogen

This option may be specified multiple times.

–dest

The destination directory for generated files. By convention, the name autogen is used.

CFC creates the following subdirectories in the destination directory:

  • include contains generated C headers.

  • sources contains generated C source files.

  • man contains generated man pages.

  • share/doc/clownfish contains generated HTML documentation.

This option is required.

–header

Specifies a file whose contents are added as a comment on top of each generated file.

–footer

Specifies a file whose contents are added as a comment on the bottom of each generated file.

Including the generated C headers

The C header files generated with cfc can be found in autogen/include. You should add this directory to your compiler’s search path, for example using -Iautogen/include under GCC.

One C header file is generated for every Clownfish header (.cfh) file. C code that makes use of a class defined in the .cfh file must include the respective C header. The Clownfish compiler also creates a few other internal C header files.

Compiling the generated source files

cfc creates one source file for every parcel in autogen/sources/{parcel_nick}_parcel.c. These files must be compiled with autogen/include added to the header search path.

Linking

When linking, add the object files of the CFC-generated code created in the previous step. You must also link the shared library of the Clownfish runtime (-lclownfish under GCC).