For example, if we want to generate a Haskell module called `FCoBase` from the FCo specification, using string-based variables, we would issue the command:
For example, if we want to generate a Haskell module called `FCoBase` from the
FCo specification, using string-based variables, we would issue the command:
@@ -22,7 +59,8 @@ The following options are currently available:
## Language descriptions
Language descriptions are simple text files. They consist of four different parts:
Language descriptions are simple text files. They consist of four different
parts:
- Import declarations
- Namespace declarations
...
...
@@ -35,13 +73,16 @@ For examples of language descriptions, see the `Specifications` directory.
### Import declarations
If your native code uses other modules, you can declare the import statements with the following syntax at the top of your language description:
If your native code uses other modules, you can declare the import statements
with the following syntax at the top of your language description:
```
import (MODULE_NAME) [ ( ENTITY [ ENTITY]* ) ]
```
Where the first parentheses contain the module name, and the optional second pair contains the specifically imported entities from the given module (separated by spaces). One line contains one module import.
Where the first parentheses contain the module name, and the optional second
pair contains the specifically imported entities from the given module
(separated by spaces). One line contains one module import.
Example:
...
...
@@ -50,11 +91,14 @@ import (Data.Map)
import (Data.List) (genericLength genericTake)
```
The Haskell implementation imports `Data.List` by default, because some of the generated functions use its functions.
The Haskell implementation imports `Data.List` by default, because some of the
generated functions use its functions.
### Namespace declarations
A namespace declaration consists of the namespace's name, a sort name, and optionally additional comma separated sort names (the environment of the namespace).
A namespace declaration consists of the namespace's name, a sort name, and
optionally additional comma separated sort names (the environment of the
namespace).
```
namespace NAMESPACENAME: SORTNAME[, SORTNAME]*
...
...
@@ -81,17 +125,23 @@ Sort declarations have the following structure:
sort SORTNAME [rewrite]
```
Declares a new sort. The optional rewrite keyword specifies whether the result of substitution functions for the given sort should be passed to an external rewrite function. The rewrite functions need to have the name `rewriteSORTNAME` and either be defined in the native code section or imported.
Declares a new sort. The optional rewrite keyword specifies whether the result
of substitution functions for the given sort should be passed to an external
rewrite function. The rewrite functions need to have the name `rewriteSORTNAME`
and either be defined in the native code section or imported.
#### Context declarations
A sort can have multiple (or no) context attributes. These are declared after the sort name with the following syntax:
A sort can have multiple (or no) context attributes. These are declared after
the sort name with the following syntax:
```
CTXTYPE CTXINSTANCE NAMESPACENAME
```
Where `CTXTYPE` can be `inh` for inherited and `syn` for synthesized contexts. `CTXINSTANCE` is the context instance's name, and `NAMESPACENAME` is the name of the namespace of the contained variables.
Where `CTXTYPE` can be `inh` for inherited and `syn` for synthesized contexts.
`CTXINSTANCE` is the context instance's name, and `NAMESPACENAME` is the name
of the namespace of the contained variables.
Example:
```
...
...
@@ -104,7 +154,9 @@ sort Pat
#### Constructor declarations
Sorts can contain a number of constructor declarations. These follow the context declarations. A constructor declaration consists of a constructor name, parameter declarations, and context attribute assignments.
Sorts can contain a number of constructor declarations. These follow the context
declarations. A constructor declaration consists of a constructor name,
parameter declarations, and context attribute assignments.
The parameter declarations can be of the following forms:
- List of a sort
...
...
@@ -113,19 +165,23 @@ The parameter declarations can be of the following forms:
- Native type
- Binder parameter of a namespace
You can have zero or more parameters of each type, but the order of their declaration must correspond to the order in this list.
You can have zero or more parameters of each type, but the order of their
declaration must correspond to the order in this list.
The final part of the constructor declaration is the context attribute assignments. Zero or more of them can follow the parameter list in new lines.
The final part of the constructor declaration is the context attribute
assignments. Zero or more of them can follow the parameter list in new lines.
```
PARAM.CTXINSTANCE = PARAM.CTXINSTANCE[, PARAM]
```
Where `PARAM` can be one of the constructor parameters or `lhs` which represents the constructor's context. You can optionally also append a parameter to the assigned context instance.
Where `PARAM` can be one of the constructor parameters or `lhs` which represents
the constructor's context. You can optionally also append a parameter to the
assigned context instance.
Example:
...
...
@@ -140,9 +196,12 @@ sort Term
### Native code
After the sort declarations you can include code written in the target language. This needs to be preceded by the `NativeCode` keyword. Everything after this keyword will be copied verbatim to the output module.
After the sort declarations you can include code written in the target language.
This needs to be preceded by the `NativeCode` keyword. Everything after this
keyword will be copied verbatim to the output module.
For example, if we want to generate a Haskell module, we can put the following in the language specification:
For example, if we want to generate a Haskell module, we can put the following
