chpldoc

chpldoc is a tool for generating HTML based documentation from Chapel source code and embedded comments. It is similar in spirit to Godoc, Javadoc, and Pydoc. chpldoc is used to produce Chapel’s web documentation for the Standard Library and some of the internal types and functions documented in the Language Spec Pages.

See the online man page for all available flags.

Building chpldoc

To build chpldoc, use:

[g]make chpldoc

This downloads the required Python package dependencies and creates the chpldoc program in the same directory as the chpl compiler.

To ensure chpldoc is installed properly, optionally run the automatic sanity check:

[g]make check-chpldoc

Prerequisites

chpldoc depends on several Python packages. The full list is available in $CHPL_HOME/third-party/chpl-venv/README.md including details on each package.

Documenting Chapel code

To document Chapel code, place a multi-line comment just prior to the definition of the symbol — module, class, function, method, global variable, etc. — that you wish to comment. All multi-line comments are considered potential documentation. Documentation is output in the same order that it exists in the source. Documentation text within comments is parsed as ReStructured Text.

Documenting Modules

To document individual modules, inside the Chapel standard library or elsewhere, use the chpldoc command directory. For example:

chpldoc modules/standard/Random.chpl

By default, documentation created by chpldoc will end up in a subdirectory of the current working directory named ‘docs/’. You can specify a different directory name using the --output-dir option.

As an example, there is a primer on the chpldoc capability in $CHPL_HOME/examples/primers/ that can be turned into documentation using:

cd $CHPL_HOME/examples/primers
chpldoc chpldoc.doc.chpl

The output documentation will be located in:

$CHPL_HOME/examples/primers/docs/index.html

Comment style

For the Chapel standard modules, it is recommended to use multiline comments without any per-line prefixes. This results in the most consistent, best looking output. For example:

/* Support for buffers - ...

   Provides bytes and buffer types, which can be used to...
 */
module Buffers {
  ...

Argument and return/yield types

Arguments and their types can optionally be documented using the :arg: and :type: fields. For the Chapel standard modules, :arg: should be a concise description of the argument using a sentence, or two. This should include any constraints and default information. :type: should be the literal type for the parameter. A link will be created to the type, if possible. For example:

/* Fill and array with pseudorandom values ...

   :arg arr: Array to be filled, where T is real(64), imag(64), or complex(128).
   :type arr: [] T

   :arg seed: Integer seed to use for the PRNG. Defaults to :proc:`SeedGenerator.currentTime`.
   :type seed: int(64)
 */
proc fillRandom(arr: [], seed: int(64) = SeedGenerator.currentTime) {
  ...

Return or yield types can optionally be documented using the :returns: and :yields: fields, and the :rtype: and :ytype: fields. :returns: and :rtype: are for procedures and methods, while :yields: and :ytype: are for iterators.

Similar to arguments, the :returns: and :yields: should be a concise description of the return value or yield value, including any constraints, using a sentence, or two. :rtype: and :ytype: should be the literal types. A link to the type will be created, if possible. For example:

/* Returns the next value in the random stream.

   :returns: Next value in the random integer stream.
   :rtype: int
 */
proc RandomStream.getNext(): int {
  ...

/* Simple find-like utility for iterating over files.

   :arg startdir: Directory to start looking for files.
   :arg recursive: Whether or not to descend recursively into directories.
   :arg hidden: Whether or not to yield hidden files.

   :yields: Filenames found in `startdir` based on input arguments.
   :ytype: string
 */
iter findfiles(startdir=".", recursive=false, hidden=false): string {
  ...

Note

  • Leave an empty line above and below all these fields for best results.

  • All of these fields must be left-aligned with the outer most paragraph(s).

For more information see the Documenting argument, return, and yield values and types section.

Documenting functions that throw

Errors that are thrown from functions can be documented using the :throw: or :throws: field. The field will include the Error subtype being thrown, for example :throws FileNotFoundError:. The description should be a short summary of when that error could occur, using a sentence, or two. For example:

/* Set the permissions of the file or directory specified by the argument
   `name` to that indicated by the argument `mode`.

   :arg name: The name of the file or directory whose permissions should be
              altered.
   :type name: `string`
   :arg mode: The permissions desired for the file or directory in question.
              See description of :const:`S_IRUSR`, for instance, for
              potential values.
   :type mode: `int`

   :throws FileNotFoundError: If :arg:`name` does not correspond to a file
                              or directory that exists.
   :throws PermissionError: If the current user does not have permission to
                            change the permissions.
*/
proc chmod(name: string, mode: int) throws {
  ...

chpldoc will not validate the list of Errors documented for a function. Documentation writers are responsible for keeping their documentation in line with the behavior of their function.

When writing :throws: fields, developers do not need to list every possible error that could be thrown - however, specificity will allow users to appropriately anticipate and respond to exceptional situations. Their response to a PermissionError when accessing a remote file will be different than their response to a ConnectionResetError, for instance. Documenting such errors ahead of time will allow users to write appropriate try - catch branches.

In some cases, it may be unreasonable to list all the possible Error subclasses that could be thrown. Developers should feel free to only list a common superclass instead - for instance, :throws SystemError: as opposed to ConnectionAbortedError, ConnectionRefusedError, and ConnectionResetError.

Note

  • Leave an empty line above and below all these fields for best results.

  • All of these fields must be left-aligned with the outer most paragraph(s).

Stifling documentation

To mark a particular symbol to not be output as part of the documentation, preface the symbol with the attribute @chpldoc.nodoc. For example:

@chpldoc.nodoc
proc foo() { ... }

Private symbols are not documented by default.

private proc foo() { ... }

Note

Private cannot be applied to all symbols at this time. It is currently limited to functions, iterators, globals, and modules. Not supported are type aliases, enum declarations, classes/records/unions, methods, and fields.

reStructuredText primer

This section is a brief introduction to reStructuredText (aka reST). It is intended to provide Chapel developers with enough information to write documentation in comments in the source code.

The Python and Sphinx projects have thorough primers of reST. Please see those for more detail.

This primer is based on the information and text in the Sphinx and Python reST primers (some of the text is copied verbatim).

The authoritative reStructuredText User Guide is also helpful.

Paragraphs

Simple chunks of text. Paragraphs can have line breaks to improve source readability. Separate paragraphs with an empty line. Indentation is significant in reST. All lines of paragraph must be left-aligned.

Inline markup

  • one asterisk: *text* for emphasis (italics),

  • two asterisks: **text** for strong emphasis (boldface), and

  • backquotes: ``text`` for code samples.

If asterisks or backquotes appear in running text and could be confused with inline markup delimiters, they have to be escaped with a backslash.

Lists and Quotes

To create a list, put an asterisk at the start of a paragraph and indent accordingly. Create numbered lists by using the literal numbers, e.g. 1., 2.. Automatically numbered lists begin with #.:

* This is a bulleted list.
* It has two items, the second
  item uses two lines.

1. This is a numbered list.
2. It has two items too.

#. This is a numbered list.
#. It has two items too.

Nested lists are possible, but be aware that they must be separated from the parent list items by blank lines:

* a
* b

  * nested
  * blah

* my final item, in the parent list

Source Code

Literal code blocks are introduced by ending a paragraph with the special marker ::. The literal block must be indented:

This is a normal text paragraph. The next paragraph is a code sample::

   It is not processed in any way, except
   that the indentation is removed.

   It can span multiple lines.

This is a normal text paragraph again.

The handling of the :: marker is smart:

  • If it occurs as a paragraph of its own, that paragraph is completely left out of the document.

  • If it is preceded by whitespace, the marker is removed.

  • If it is preceded by non-whitespace, the marker is replaced by a single colon.

That way, the second sentence in the above example’s first paragraph would be rendered as “The next paragraph is a code sample:”.

The highlight language is configured with the highlight directive. The configured language is used for all literal blocks until the next highlight directive. For example:

.. highlight:: chapel

Chapel code::

   writeln("Hello from Chapel!");

More chapel::

   x <=> y;

.. highlight:: c

::

   printf("Hello from C!\n");

Showing code examples

The code-block directive can be used to specify the highlight language of a single code block. For example:

.. code-block:: chapel

   use Foo;

   proc bar() {
     writeln("Fooy!");
   }

If highlighting with the specified language fails, e.g. if the syntax is not parsable, the block is not highlighted in anyway. Note that there should be a blank line between the code-block directive and the indented code snippet.

Sections

Section headers are created by underlining (and optionally overlining) the section title with a punctuation character, at least as long as the text:

This is a heading
=================

There are no heading levels assigned to certain characters. The structure is determined from the succession of headings.

Comments

Every explicit markup block which is not a valid markup construct is regarded as a comment. For example:

This is a normal paragraph.
It is interesting.

.. TODO: Make it more interesting.

Another paragraph goes here.

.. add another paragraph below

You can indent text after a comment start to form multiline comments:

..
   This whole indented block
   is a comment.

   Still in the comment.

Inline markup

As said before, Sphinx uses interpreted text roles to insert semantic markup in documents.

Names of builtins, like true, false, types like int(64), and local variables, such as function/method arguments, are an exception, they should be marked simply with `myVar`.

For all other roles, you have to write :rolename:`content`.

There are some additional facilities that make cross-referencing roles more versatile:

  • You may supply an explicit title and reference target, like in reST direct hyperlinks: :role:`title <target>` will refer to target, but the link text will be title.

  • If you prefix the content with !, no reference/hyperlink will be created.

  • For the Chapel roles, if you prefix the content with ~, the link text will only be the last component of the target. For example, :proc:`~Random.RandomStream.fillRandom` will refer to Random.RandomStream.fillRandom but only display fillRandom as the link text.

    In HTML output, the link’s title attribute (that is e.g. shown as a tool-tip on mouse-hover) will always be the full target name.

The following roles refer to objects in modules and are possibly hyperlinked if a matching identifier is found:

:mod:

Reference a module; a dotted name may be used. See Inline markup for details on dotted and non-dotted names.

:proc: :iter:

Reference a Chapel function or iterator. The role text needs not include trailing parentheses to enhance readability.

These can also be used to reference a method or iterator on an object (class or record instance). The role text can include the type name and the method, in those cases. If it occurs within the description of a type, the type name can be omitted.

Dotted names may be used for any form.

:data: :const: :var: :param: :type: :enum: :enumconstant:

Reference a module-level variable, constant, compiler param, type, enum, or enum constant.

:class: :record:

Reference a class or record; a dotted name may be used.

:attr:

Reference a data attribute (const, var, param, generic type) of an object. Note that this does _not_ refer to a Chapel attribute; Sphinx uses the term “attribute” in a different way than Chapel does, and we pick the Sphinx terminology here.

For attributes, you should use the :annotation: role instead.

:annotation:

Reference to the documentation for an attribute like @assertOnGpu. The name of this role is annotation because Sphinx uses the term “attribute” in a different way than Chapel does; see also the documentation for :attr: above.

For example:

Uses :proc:`Random.RandomStream.fillRandom` and real->int casts to
generate a vector of random integers. See :attr:`RandomStream.seed`
and description of :mod:`Random` for details on PRNG.

Relies on :iter:`MyModule.Set.these` to iterate over all values in the
given :record:`MyModule.Set`.

Documenting argument, return, and yield values and types

Inside Chapel description directives, reST field lists with these fields are recognized and formatted nicely:

  • arg, argument: Description of a parameter.

  • type: Type of a parameter. Creates a link if possible.

  • returns, return: Description of the return value.

  • rtype: Return type. Creates a link if possible.

  • yields, yield: Description of the yield value, often used for iterators.

  • ytype: Yield type. Creates a link if possible.

type, rtype, and ytype should be concise and literal type definitions, like int, int(64), bool, [] int, RandomStream, etc. More verbose descriptions, qualifications, and limitations of those types should go in the corresponding arg, returns, or yields field.

For example, when documenting a Chapel proc:

/*
 * Calculates number of pipes and returns fooy.
 *
 * :arg bars: Number of bars. Must be more than 1 and less than 1000.
 * :type bars: int
 *
 * :arg hours: Hours available. Default is 1.0.
 * :type hours: real
 *
 * :returns: Amount of fooy available.
 * :rtype: Foo
 */
proc foo(bars, hours=1.0): Foo
{
  ...
}

Note

  • Leave an empty line above and below all these fields for best results.

  • All of these fields must be left-aligned with the outer most paragraph(s).

Paragraph level markup

These directives create short paragraphs and can be used inside information units as well as normal text:

note

An especially important bit of information about an API that a user should be aware of when using whatever bit of API the note pertains to. The content of the directive should be written in complete sentences and include all appropriate punctuation.

Example:

.. note::

   This function is not suitable for high precision calculations.

warning

An important bit of information about an API that a user should be aware of when using whatever bit of API the warning pertains to. The content of the directive should be written in complete sentences and include all appropriate punctuation. In the interest of not scaring users away from pages filled with warnings, this directive should only be chosen over note for information regarding the possibility of crashes, data loss, or security implications.

versionadded

This directive documents the version of Chapel which added the described feature, or a part of it, to the library or API. When this applies to an entire module, it should be placed at the top of the module section before any prose.

The first argument must be given and is the version in question; if the addition is only part of the described API element, you should add a second argument consisting of a brief explanation of the change.

Example:

.. versionadded:: 2.1
   Multi-precision integer support added.

Note that there must be no blank line between the directive head and the explanation; this is to make these blocks visually continuous in the markup.

versionchanged

Similar to versionadded, but describes when and what changed in the named feature in some way (changed side effects, platform support, etc.). This one must have the second argument (explanation of the change).

seealso

Many sections include a list of references to module documentation or external documents. These lists are created using the seealso directive.

The seealso directive is typically placed in a section just before any sub-sections. For the HTML output, it is shown boxed off from the main flow of the text.

The content of the seealso directive should be a reST definition list. Example:

.. seealso::

   Module :mod:`Random`
      Documentation of the :mod:`Random` standard module.

   `Mersenne Twister pseudo random number generator <http://link>`_
      Documentation for the PRNG.

Advanced chpldoc options

If you would like to restrict documentation to multi-line comments starting only with a special character sequence (say, /***) use the --docs-comment-style flag to indicate the desired prefix (e.g., --docs-comment-style='/***'). Setting a comment style in this way also establishes that the closing comment style should have the same number of characters (though they can be different ones).

How chpldoc works

The Chapel standard library documentation is generated by running chpldoc over all Chapel source files in $CHPL_HOME/modules/standard/.

The markup used in the comments is reStructuredText. reStructuredText is developed by the docutils projects and is amended by custom directives to support documenting Chapel code. Sphinx is used by chpldoc to render reStructuredText as HTML.

Future directions

If there are other features you would like, please let us know. These are currently on our backlog:

  • Expand visibility control (public/private) to remaining Chapel symbols (with options for including private elements in output if desired).

  • Ability to include doctests, which would be code snippets in documentation that can be tested. This is similar to Python’s doctest feature.