blob: feb3dba3234ef9d7326561248648a9268bbd24c6 [file] [log] [blame]
Modules
-------
.. index:: index, type index, function index, table index, memory index, global index, local index, label index
pair: text format; type index
pair: text format; function index
pair: text format; table index
pair: text format; memory index
pair: text format; global index
pair: text format; local index
pair: text format; label index
.. _text-typeidx:
.. _text-funcidx:
.. _text-tableidx:
.. _text-memidx:
.. _text-globalidx:
.. _text-localidx:
.. _text-labelidx:
.. _text-index:
Indices
~~~~~~~
:ref:`Indices <syntax-index>` can be given either in raw numeric form or as symbolic :ref:`identifiers <text-id>` when bound by a respective construct.
Such identifiers are looked up in the suitable space of the :ref:`identifier context <text-context>` :math:`I`.
.. math::
\begin{array}{llcllllllll}
\production{type index} & \Ttypeidx_I &::=&
x{:}\Tu32 &\Rightarrow& x \\&&|&
v{:}\Tid &\Rightarrow& x & (\iff I.\ITYPES[x] = v) \\
\production{function index} & \Tfuncidx_I &::=&
x{:}\Tu32 &\Rightarrow& x \\&&|&
v{:}\Tid &\Rightarrow& x & (\iff I.\IFUNCS[x] = v) \\
\production{table index} & \Ttableidx_I &::=&
x{:}\Tu32 &\Rightarrow& x \\&&|&
v{:}\Tid &\Rightarrow& x & (\iff I.\ITABLES[x] = v) \\
\production{memory index} & \Tmemidx_I &::=&
x{:}\Tu32 &\Rightarrow& x \\&&|&
v{:}\Tid &\Rightarrow& x & (\iff I.\IMEMS[x] = v) \\
\production{global index} & \Tglobalidx_I &::=&
x{:}\Tu32 &\Rightarrow& x \\&&|&
v{:}\Tid &\Rightarrow& x & (\iff I.\IGLOBALS[x] = v) \\
\production{local index} & \Tlocalidx_I &::=&
x{:}\Tu32 &\Rightarrow& x \\&&|&
v{:}\Tid &\Rightarrow& x & (\iff I.\ILOCALS[x] = v) \\
\production{label index} & \Tlabelidx_I &::=&
l{:}\Tu32 &\Rightarrow& l \\&&|&
v{:}\Tid &\Rightarrow& l & (\iff I.\ILABELS[l] = v) \\
\end{array}
.. index:: type definition, identifier
pair: text format; type definition
.. _text-typedef:
Types
~~~~~
Type definitions can bind a symbolic :ref:`type identifier <text-id>`.
.. math::
\begin{array}{llclll}
\production{type definition} & \Ttype &::=&
\text{(}~\text{type}~~\Tid^?~~\X{ft}{:}\Tfunctype~\text{)}
&\Rightarrow& \X{ft} \\
\end{array}
.. index:: type use
pair: text format; type use
.. _text-typeuse:
Type Uses
~~~~~~~~~
A *type use* is a reference to a :ref:`type definition <text-type>`.
It may optionally be augmented by explicit inlined :ref:`parameter <text-param>` and :ref:`result <text-result>` declarations.
That allows binding symbolic :ref:`identifiers <text-id>` to name the :ref:`local indices <text-localidx>` of parameters.
If inline declarations are given, then their types must match the referenced :ref:`function type <text-type>`.
.. math::
\begin{array}{llcllll}
\production{type use} & \Ttypeuse_I &::=&
\text{(}~\text{type}~~x{:}\Ttypeidx_I~\text{)}
\quad\Rightarrow\quad x, I' \\ &&& \qquad
(\iff \begin{array}[t]{@{}l@{}}
I.\ITYPEDEFS[x] = [t_1^n] \to [t_2^\ast] \wedge
I' = \{\ILOCALS~(\epsilon)^n\}) \\
\end{array} \\[1ex] &&|&
\text{(}~\text{type}~~x{:}\Ttypeidx_I~\text{)}
~~(t_1{:}\Tparam)^\ast~~(t_2{:}\Tresult)^\ast
\quad\Rightarrow\quad x, I' \\ &&& \qquad
(\iff \begin{array}[t]{@{}l@{}}
I.\ITYPEDEFS[x] = [t_1^\ast] \to [t_2^\ast] \wedge
I' = \{\ILOCALS~\F{id}(\Tparam)^\ast\} \idcwellformed) \\
\end{array} \\
\end{array}
The synthesized attribute of a |Ttypeuse| is a pair consisting of both the used :ref:`type index <syntax-typeidx>` and the updated :ref:`identifier context <text-context>` including possible parameter identifiers.
The following auxiliary function extracts optional identifiers from parameters:
.. math::
\begin{array}{lcl@{\qquad\qquad}l}
\F{id}(\text{(}~\text{param}~\Tid^?~\dots~\text{)}) &=& \Tid^? \\
\end{array}
.. note::
Both productions overlap for the case that the function type is :math:`[] \to []`.
However, in that case, they also produce the same results, so that the choice is immaterial.
The :ref:`well-formedness <text-context-wf>` condition on :math:`I'` ensures that the parameters do not contain duplicate identifier.
.. _text-typeuse-abbrev:
Abbreviations
.............
A |Ttypeuse| may also be replaced entirely by inline :ref:`parameter <text-param>` and :ref:`result <text-result>` declarations.
In that case, a :ref:`type index <syntax-typeidx>` is automatically inserted:
.. math::
\begin{array}{llclll}
\production{type use} &
(t_1{:}\Tparam)^\ast~~(t_2{:}\Tresult)^\ast &\equiv&
\text{(}~\text{type}~~x~\text{)}~~\Tparam^\ast~~\Tresult^\ast \\
\end{array}
where :math:`x` is the smallest existing :ref:`type index <syntax-typeidx>` whose definition in the current module is the :ref:`function type <syntax-functype>` :math:`[t_1^\ast] \to [t_2^\ast]`.
If no such index exists, then a new :ref:`type definition <text-type>` of the form
.. math::
\text{(}~\text{type}~~\text{(}~\text{func}~~\Tparam^\ast~~\Tresult~\text{)}~\text{)}
is inserted at the end of the module.
Abbreviations are expanded in the order they appear, such that previously inserted type definitions are reused by consecutive expansions.
.. index:: import, name, function type, table type, memory type, global type
pair: text format; import
.. _text-importdesc:
.. _text-import:
Imports
~~~~~~~
The descriptors in imports can bind a symbolic function, table, memory, or global :ref:`identifier <text-id>`.
.. math::
\begin{array}{llclll}
\production{import} & \Timport_I &::=&
\text{(}~\text{import}~~\X{mod}{:}\Tname~~\X{nm}{:}\Tname~~d{:}\Timportdesc_I~\text{)} \\ &&& \qquad
\Rightarrow\quad \{ \IMODULE~\X{mod}, \INAME~\X{nm}, \IDESC~d \} \\[1ex]
\production{import description} & \Timportdesc_I &::=&
\text{(}~\text{func}~~\Tid^?~~x,I'{:}\Ttypeuse_I~\text{)}
&\Rightarrow& \IDFUNC~x \\ &&|&
\text{(}~\text{table}~~\Tid^?~~\X{tt}{:}\Ttabletype~\text{)}
&\Rightarrow& \IDTABLE~\X{tt} \\ &&|&
\text{(}~\text{memory}~~\Tid^?~~\X{mt}{:}\Tmemtype~\text{)}
&\Rightarrow& \IDMEM~~\X{mt} \\ &&|&
\text{(}~\text{global}~~\Tid^?~~\X{gt}{:}\Tglobaltype~\text{)}
&\Rightarrow& \IDGLOBAL~\X{gt} \\
\end{array}
Abbreviations
.............
As an abbreviation, imports may also be specified inline with :ref:`function <text-func>`, :ref:`table <text-table>`, :ref:`memory <text-mem>`, or :ref:`global <text-global>` definitions; see the respective sections.
.. index:: function, type index, function type, identifier, local
pair: text format; function
pair: text format; local
.. _text-local:
.. _text-func:
Functions
~~~~~~~~~
Function definitions can bind a symbolic :ref:`function identifier <text-id>`, and :ref:`local identifiers <text-id>` for its :ref:`parameters <text-typeuse>` and locals.
.. math::
\begin{array}{llclll}
\production{function} & \Tfunc_I &::=&
\text{(}~\text{func}~~\Tid^?~~x,I'{:}\Ttypeuse_I~~
(t{:}\Tlocal)^\ast~~(\X{in}{:}\Tinstr_{I''})^\ast~\text{)} \\ &&& \qquad
\Rightarrow\quad \{ \FTYPE~x, \FLOCALS~t^\ast, \FBODY~\X{in}^\ast~\END \} \\ &&& \qquad\qquad\qquad
(\iff I'' = I' \compose \{\ILOCALS~\F{id}(\Tlocal)^\ast\} \idcwellformed) \\[1ex]
\production{local} & \Tlocal &::=&
\text{(}~\text{local}~~\Tid^?~~t{:}\Tvaltype~\text{)}
\quad\Rightarrow\quad t \\
\end{array}
The definition of the local :ref:`identifier context <text-context>` :math:`I''` uses the following auxiliary function to extract optional identifiers from locals:
.. math::
\begin{array}{lcl@{\qquad\qquad}l}
\F{id}(\text{(}~\text{local}~\Tid^?~\dots~\text{)}) &=& \Tid^? \\
\end{array}
.. note::
The :ref:`well-formedness <text-context-wf>` condition on :math:`I''` ensures that parameters and locals do not contain duplicate identifiers.
.. index:: import, name
pair: text format; import
.. index:: export, name, index, function index
pair: text format; export
.. _text-func-abbrev:
Abbreviations
.............
Multiple anonymous locals may be combined into a single declaration:
.. math::
\begin{array}{llclll}
\production{local} &
\text{(}~~\text{local}~~\Tvaltype^\ast~~\text{)} &\equiv&
(\text{(}~~\text{local}~~\Tvaltype~~\text{)})^\ast \\
\end{array}
Functions can be defined as :ref:`imports <text-import>` or :ref:`exports <text-export>` inline:
.. math::
\begin{array}{llclll}
\production{module field} &
\text{(}~\text{func}~~\Tid^?~~\text{(}~\text{import}~~\Tname_1~~\Tname_2~\text{)}~~\Ttypeuse~\text{)} \quad\equiv \\ & \qquad
\text{(}~\text{import}~~\Tname_1~~\Tname_2~~\text{(}~\text{func}~~\Tid^?~~\Ttypeuse~\text{)}~\text{)} \\[1ex] &
\text{(}~\text{func}~~\Tid^?~~\text{(}~\text{export}~~\Tname~\text{)}~~\dots~\text{)} \quad\equiv \\ & \qquad
\text{(}~\text{export}~~\Tname~~\text{(}~\text{func}~~\Tid'~\text{)}~\text{)}~~
\text{(}~\text{func}~~\Tid'~~\dots~\text{)}
\\ & \qquad\qquad
(\iff \Tid' = \Tid^? \neq \epsilon \vee \Tid' \idfresh) \\
\end{array}
The latter abbreviation can be applied repeatedly, with ":math:`\dots`" containing another import or export.
.. index:: table, table type, identifier
pair: text format; table
.. _text-table:
Tables
~~~~~~
Table definitions can bind a symbolic :ref:`table identifier <text-id>`.
.. math::
\begin{array}{llclll}
\production{table} & \Ttable_I &::=&
\text{(}~\text{table}~~\Tid^?~~\X{tt}{:}\Ttabletype~\text{)}
&\Rightarrow& \{ \TTYPE~\X{tt} \} \\
\end{array}
.. index:: import, name
pair: text format; import
.. index:: export, name, index, table index
pair: text format; export
.. index:: element, table index, function index
pair: text format; element
single: table; element
single: element; segment
.. _text-table-abbrev:
Abbreviations
.............
An :ref:`element segment <text-elem>` can be given inline with a table definition, in which case its offset is :math:`0` and the :ref:`limits <text-limits>` of the :ref:`table type <text-tabletype>` are inferred from the length of the given segment:
.. math::
\begin{array}{llclll}
\production{module field} &
\text{(}~\text{table}~~\Tid^?~~\Telemtype~~\text{(}~\text{elem}~~x^n{:}\Tvec(\Tfuncidx)~\text{)}~~\text{)} \quad\equiv \\ & \qquad
\text{(}~\text{table}~~\Tid'~~n~~n~~\Telemtype~\text{)}~~
\text{(}~\text{elem}~~\Tid'~~\text{(}~\text{i32.const}~~\text{0}~\text{)}~~\Tvec(\Tfuncidx)~\text{)}
\\ & \qquad\qquad
(\iff \Tid' = \Tid^? \neq \epsilon \vee \Tid' \idfresh) \\
\end{array}
Tables can be defined as :ref:`imports <text-import>` or :ref:`exports <text-export>` inline:
.. math::
\begin{array}{llclll}
\production{module field} &
\text{(}~\text{table}~~\Tid^?~~\text{(}~\text{import}~~\Tname_1~~\Tname_2~\text{)}~~\Ttabletype~\text{)} \quad\equiv \\ & \qquad
\text{(}~\text{import}~~\Tname_1~~\Tname_2~~\text{(}~\text{table}~~\Tid^?~~\Ttabletype~\text{)}~\text{)} \\[1ex] &
\text{(}~\text{table}~~\Tid^?~~\text{(}~\text{export}~~\Tname~\text{)}~~\dots~\text{)} \quad\equiv \\ & \qquad
\text{(}~\text{export}~~\Tname~~\text{(}~\text{table}~~\Tid'~\text{)}~\text{)}~~
\text{(}~\text{table}~~\Tid'~~\dots~\text{)}
\\ & \qquad\qquad
(\iff \Tid' = \Tid^? \neq \epsilon \vee \Tid' \idfresh) \\
\end{array}
The latter abbreviation can be applied repeatedly, with ":math:`\dots`" containing another import or export or an inline elements segment.
.. index:: memory, memory type, identifier
pair: text format; memory
.. _text-mem:
Memories
~~~~~~~~
Memory definitions can bind a symbolic :ref:`memory identifier <text-id>`.
.. math::
\begin{array}{llclll}
\production{memory} & \Tmem_I &::=&
\text{(}~\text{memory}~~\Tid^?~~\X{mt}{:}\Tmemtype~\text{)}
&\Rightarrow& \{ \MTYPE~\X{mt} \} \\
\end{array}
.. index:: import, name
pair: text format; import
.. index:: export, name, index, memory index
pair: text format; export
.. index:: data, memory, memory index, expression, byte, page size
pair: text format; data
single: memory; data
single: data; segment
.. _text-mem-abbrev:
Abbreviations
.............
A :ref:`data segment <text-data>` can be given inline with a memory definition, in which case its offset is :math:`0` the :ref:`limits <text-limits>` of the :ref:`memory type <text-memtype>` are inferred from the length of the data, rounded up to :ref:`page size <page-size>`:
.. math::
\begin{array}{llclll}
\production{module field} &
\text{(}~\text{memory}~~\Tid^?~~\text{(}~\text{data}~~b^n{:}\Tdatastring~\text{)}~~\text{)} \quad\equiv \\ & \qquad
\text{(}~\text{memory}~~\Tid'~~m~~m~\text{)}~~
\text{(}~\text{data}~~\Tid'~~\text{(}~\text{i32.const}~~\text{0}~\text{)}~~\Tdatastring~\text{)}
\\ & \qquad\qquad
(\iff \Tid' = \Tid^? \neq \epsilon \vee \Tid' \idfresh, m = \F{ceil}(n / 64\F{Ki})) \\
\end{array}
Memories can be defined as :ref:`imports <text-import>` or :ref:`exports <text-export>` inline:
.. math::
\begin{array}{llclll}
\production{module field} &
\text{(}~\text{memory}~~\Tid^?~~\text{(}~\text{import}~~\Tname_1~~\Tname_2~\text{)}~~\Tmemtype~\text{)} \quad\equiv \\ & \qquad
\text{(}~\text{import}~~\Tname_1~~\Tname_2~~\text{(}~\text{memory}~~\Tid^?~~\Tmemtype~\text{)}~\text{)}
\\[1ex] &
\text{(}~\text{memory}~~\Tid^?~~\text{(}~\text{export}~~\Tname~\text{)}~~\dots~\text{)} \quad\equiv \\ & \qquad
\text{(}~\text{export}~~\Tname~~\text{(}~\text{memory}~~\Tid'~\text{)}~\text{)}~~
\text{(}~\text{memory}~~\Tid'~~\dots~\text{)}
\\ & \qquad\qquad
(\iff \Tid' = \Tid^? \neq \epsilon \vee \Tid' \idfresh) \\
\end{array}
The latter abbreviation can be applied repeatedly, with ":math:`\dots`" containing another import or export or an inline data segment.
.. index:: global, global type, identifier, expression
pair: text format; global
.. _text-global:
Globals
~~~~~~~
Global definitions can bind a symbolic :ref:`global identifier <text-id>`.
.. math::
\begin{array}{llclll}
\production{global} & \Tglobal_I &::=&
\text{(}~\text{global}~~\Tid^?~~\X{gt}{:}\Tglobaltype~~e{:}\Texpr_I~\text{)}
&\Rightarrow& \{ \GTYPE~\X{gt}, \GINIT~e \} \\
\end{array}
.. index:: import, name
pair: text format; import
.. index:: export, name, index, global index
pair: text format; export
.. _text-global-abbrev:
Abbreviations
.............
Globals can be defined as :ref:`imports <text-import>` or :ref:`exports <text-export>` inline:
.. math::
\begin{array}{llclll}
\production{module field} &
\text{(}~\text{global}~~\Tid^?~~\text{(}~\text{import}~~\Tname_1~~\Tname_2~\text{)}~~\Tglobaltype~\text{)} \quad\equiv \\ & \qquad
\text{(}~\text{import}~~\Tname_1~~\Tname_2~~\text{(}~\text{global}~~\Tid^?~~\Tglobaltype~\text{)}~\text{)}
\\[1ex] &
\text{(}~\text{global}~~\Tid^?~~\text{(}~\text{export}~~\Tname~\text{)}~~\dots~\text{)} \quad\equiv \\ & \qquad
\text{(}~\text{export}~~\Tname~~\text{(}~\text{global}~~\Tid'~\text{)}~\text{)}~~
\text{(}~\text{global}~~\Tid'~~\dots~\text{)}
\\ & \qquad\qquad
(\iff \Tid' = \Tid^? \neq \epsilon \vee \Tid' \idfresh) \\
\end{array}
The latter abbreviation can be applied repeatedly, with ":math:`\dots`" containing another import or export.
.. index:: export, name, index, function index, table index, memory index, global index
pair: text format; export
.. _text-exportdesc:
.. _text-export:
Exports
~~~~~~~
The syntax for exports mirrors their :ref:`abstract syntax <syntax-export>` directly.
.. math::
\begin{array}{llclll}
\production{export} & \Texport_I &::=&
\text{(}~\text{export}~~\X{nm}{:}\Tname~~d{:}\Texportdesc_I~\text{)}
&\Rightarrow& \{ \ENAME~\X{nm}, \EDESC~d \} \\
\production{export description} & \Texportdesc_I &::=&
\text{(}~\text{func}~~x{:}\Bfuncidx_I~\text{)}
&\Rightarrow& \EDFUNC~x \\ &&|&
\text{(}~\text{table}~~x{:}\Btableidx_I~\text{)}
&\Rightarrow& \EDTABLE~x \\ &&|&
\text{(}~\text{memory}~~x{:}\Bmemidx_I~\text{)}
&\Rightarrow& \EDMEM~x \\ &&|&
\text{(}~\text{global}~~x{:}\Bglobalidx_I~\text{)}
&\Rightarrow& \EDGLOBAL~x \\
\end{array}
Abbreviations
.............
As an abbreviation, exports may also be specified inline with :ref:`function <text-func>`, :ref:`table <text-table>`, :ref:`memory <text-mem>`, or :ref:`global <text-global>` definitions; see the respective sections.
.. index:: start function, function index
pair: text format; start function
.. _text-start:
Start Function
~~~~~~~~~~~~~~
A :ref:`start function <syntax-start>` is defined in terms of its index.
.. math::
\begin{array}{llclll}
\production{start function} & \Tstart_I &::=&
\text{(}~\text{start}~~x{:}\Tfuncidx_I~\text{)}
&\Rightarrow& \{ \SFUNC~x \} \\
\end{array}
.. note::
At most one start function may occur in a module,
which is ensured by a suitable side condition on the |Tmodule| grammar.
.. index:: element, table index, expression, function index
pair: text format; element
single: table; element
single: element; segment
.. _text-elem:
Element Segments
~~~~~~~~~~~~~~~~
Element segments allow for an optional :ref:`table index <text-tableidx>` to identify the table to initialize.
.. math::
\begin{array}{llclll}
\production{element segment} & \Telem_I &::=&
\text{(}~\text{elem}~~x{:}\Ttableidx_I~~\text{(}~\text{offset}~~e{:}\Texpr_I~\text{)}~~y^\ast{:}\Tvec(\Tfuncidx_I)~\text{)} \\ &&& \qquad
\Rightarrow\quad \{ \ETABLE~x, \EOFFSET~e, \EINIT~y^\ast \} \\
\end{array}
.. note::
In the current version of WebAssembly, the only valid table index is 0
or a symbolic :ref:`table identifier <text-id>` resolving to the same value.
Abbreviations
.............
As an abbreviation, a single instruction may occur in place of the offset:
.. math::
\begin{array}{llcll}
\production{element offset} &
\Tinstr &\equiv&
\text{(}~\text{offset}~~\Tinstr~\text{)}
\end{array}
Also, the table index can be omitted, defaulting to :math:`\T{0}`.
.. math::
\begin{array}{llclll}
\production{element segment} &
\text{(}~\text{elem}~~\text{(}~\text{offset}~~\Texpr_I~\text{)}~~\dots~\text{)}
&\equiv&
\text{(}~\text{elem}~~0~~\text{(}~\text{offset}~~\Texpr_I~\text{)}~~\dots~\text{)}
\end{array}
As another abbreviation, element segments may also be specified inline with :ref:`table <text-table>` definitions; see the respective section.
.. index:: data, memory, memory index, expression, byte
pair: text format; data
single: memory; data
single: data; segment
.. _text-datastring:
.. _text-data:
Data Segments
~~~~~~~~~~~~~
Data segments allow for an optional :ref:`memory index <text-memidx>` to identify the memory to initialize.
The data is written as a :ref:`string <text-string>`, which may be split up into a possibly empty sequence of individual string literals.
.. math::
\begin{array}{llclll}
\production{data segment} & \Tdata_I &::=&
\text{(}~\text{data}~~x{:}\Tmemidx_I~~\text{(}~\text{offset}~~e{:}\Texpr_I~\text{)}~~b^\ast{:}\Tdatastring~\text{)} \\ &&& \qquad
\Rightarrow\quad \{ \DMEM~x', \DOFFSET~e, \DINIT~b^\ast \} \\[1ex]
\production{data string} & \Tdatastring &::=&
(b^\ast{:}\Tstring)^\ast \quad\Rightarrow\quad \concat((b^\ast)^\ast) \\
\end{array}
.. note::
In the current version of WebAssembly, the only valid memory index is 0
or a symbolic :ref:`memory identifier <text-id>` resolving to the same value.
Abbreviations
.............
As an abbreviation, a single instruction may occur in place of the offset:
.. math::
\begin{array}{llcll}
\production{data offset} &
\Tinstr &\equiv&
\text{(}~\text{offset}~~\Tinstr~\text{)}
\end{array}
Also, the memory index can be omitted, defaulting to :math:`\T{0}`.
.. math::
\begin{array}{llclll}
\production{data segment} &
\text{(}~\text{data}~~\text{(}~\text{offset}~~\Texpr_I~\text{)}~~\dots~\text{)}
&\equiv&
\text{(}~\text{data}~~0~~\text{(}~\text{offset}~~\Texpr_I~\text{)}~~\dots~\text{)}
\end{array}
As another abbreviation, data segments may also be specified inline with :ref:`memory <text-mem>` definitions; see the respective section.
.. index:: module, type definition, function type, function, table, memory, global, element, data, start function, import, export, identifier context, identifier, name section
pair: text format; module
single: section; name
.. _text-modulefield:
.. _text-module:
Modules
~~~~~~~
A module consists of a sequence of fields that can occur in any order.
All definitions and their respective bound :ref:`identifiers <text-id>` scope over the entire module, including the text preceding them.
A module may optionally bind an :ref:`identifier <text-id>` that names the module.
The name serves a documentary role only.
.. note::
Tools may include the module name in the :ref:`name section <binary-namesec>` of the :ref:`binary format <binary>`.
.. math::
\begin{array}{lll}
\production{module} & \Tmodule &
\begin{array}[t]{@{}cllll}
::=&
\text{(}~\text{module}~~\Tid^?~~(m{:}\Tmodulefield_I)^\ast~\text{)}
\quad\Rightarrow\quad \bigcompose m^\ast \\
&\qquad (\iff I = \bigcompose \F{idc}(\Tmodulefield)^\ast \idcwellformed) \\
\end{array} \\[1ex]
\production{module field} & \Tmodulefield_I &
\begin{array}[t]{@{}clll}
::=&
\X{ty}{:}\Ttype &\Rightarrow& \{\MTYPES~\X{ty}\} \\ |&
\X{im}{:}\Timport_I &\Rightarrow& \{\MIMPORTS~\X{im}\} \\ |&
\X{fn}{:}\Tfunc_I &\Rightarrow& \{\MFUNCS~\X{fn}\} \\ |&
\X{ta}{:}\Ttable_I &\Rightarrow& \{\MTABLES~\X{ta}\} \\ |&
\X{me}{:}\Tmem_I &\Rightarrow& \{\MMEMS~\X{me}\} \\ |&
\X{gl}{:}\Tglobal_I &\Rightarrow& \{\MGLOBALS~\X{gl}\} \\ |&
\X{ex}{:}\Texport_I &\Rightarrow& \{\MEXPORTS~\X{ex}\} \\ |&
\X{st}{:}\Tstart_I &\Rightarrow& \{\MSTART~\X{st}\} \\ |&
\X{el}{:}\Telem_I &\Rightarrow& \{\MELEM~\X{el}\} \\ |&
\X{da}{:}\Tdata_I &\Rightarrow& \{\MDATA~\X{da}\} \\
\end{array}
\end{array}
The following restrictions are imposed on the composition of :ref:`modules <syntax-module>`: :math:`m_1 \compose m_2` is defined if and only if
* :math:`m_1.\MSTART = \epsilon \vee m_2.\MSTART = \epsilon`
* :math:`m_1.\MFUNCS = m_1.\MTABLES = m_1.\MMEMS = m_1.\MGLOBALS = \epsilon \vee m_2.\MIMPORTS = \epsilon`
.. note::
The first condition ensures that there is at most one start function.
The second condition enforces that all :ref:`imports <text-import>` must occur before any regular definition of a :ref:`function <text-func>`, :ref:`table <text-table>`, :ref:`memory <text-mem>`, or :ref:`global <text-global>`,
thereby maintaining the ordering of the respective :ref:`index spaces <syntax-index>`.
The :ref:`well-formedness <text-context-wf>` condition on :math:`I` in the grammar for |Tmodule| ensures that no namespace contains duplicate identifiers.
The definition of the initial :ref:`identifier context <text-context>` :math:`I` uses the following auxiliary definition which maps each relevant definition to a singular context with one (possibly empty) identifier:
.. math::
\begin{array}{@{}lcl@{\qquad\qquad}l}
\F{idc}(\text{(}~\text{type}~\Tid^?~\X{ft}{:}\Tfunctype~\text{)}) &=&
\{\ITYPES~(\Tid^?), \ITYPEDEFS~\X{ft}\} \\
\F{idc}(\text{(}~\text{func}~\Tid^?~\dots~\text{)}) &=&
\{\IFUNCS~(\Tid^?)\} \\
\F{idc}(\text{(}~\text{table}~\Tid^?~\dots~\text{)}) &=&
\{\ITABLES~(\Tid^?)\} \\
\F{idc}(\text{(}~\text{memory}~\Tid^?~\dots~\text{)}) &=&
\{\IMEMS~(\Tid^?)\} \\
\F{idc}(\text{(}~\text{global}~\Tid^?~\dots~\text{)}) &=&
\{\IGLOBALS~(\Tid^?)\} \\
\F{idc}(\text{(}~\text{import}~\dots~\text{(}~\text{func}~\Tid^?~\dots~\text{)}~\text{)}) &=&
\{\IFUNCS~(\Tid^?)\} \\
\F{idc}(\text{(}~\text{import}~\dots~\text{(}~\text{table}~\Tid^?~\dots~\text{)}~\text{)}) &=&
\{\ITABLES~(\Tid^?)\} \\
\F{idc}(\text{(}~\text{import}~\dots~\text{(}~\text{memory}~\Tid^?~\dots~\text{)}~\text{)}) &=&
\{\IMEMS~(\Tid^?)\} \\
\F{idc}(\text{(}~\text{import}~\dots~\text{(}~\text{global}~\Tid^?~\dots~\text{)}~\text{)}) &=&
\{\IGLOBALS~(\Tid^?)\} \\
\F{idc}(\text{(}~\dots~\text{)}) &=&
\{\} \\
\end{array}
Abbreviations
.............
In a source file, the toplevel :math:`\T{(module}~\dots\T{)}` surrounding the module body may be omitted.
.. math::
\begin{array}{llcll}
\production{module} &
\Tmodulefield^\ast &\equiv&
\text{(}~\text{module}~~\Tmodulefield^\ast~\text{)}
\end{array}