branch: externals/parser-generator
commit 4fe014a47fbd1daaf48b696e48b5236364c20ddb
Author: Christian Johansson <christ...@cvj.se>
Commit: Christian Johansson <christ...@cvj.se>
Updating documentation regarding operator precedence
---
README.md | 4 +---
docs/Syntax-Analysis.md | 20 ++++++++++++++++++++
docs/Syntax-Analysis/LR0.md | 36 ++++++++++++++++++++++++++++++++++++
docs/Syntax-Analysis/LRk.md | 36 ++++++++++++++++++++++++++++++++++++
4 files changed, 93 insertions(+), 3 deletions(-)
diff --git a/README.md b/README.md
index e2b7477..67fc98c 100644
--- a/README.md
+++ b/README.md
@@ -3,9 +3,7 @@
[](https://www.gnu.org/licenses/gpl-3.0.txt)
[](https://travis-ci.org/cjohansson/emacs-parser-generator)
-The idea of this plugin is to provide functions for various kinds of
context-free grammar parser generations with support for
syntax-directed-translations (SDT) and semantic actions (SA) and the
possibility of exporting parsers and translators (as elisp code) to enable
plugin-agnostic usage. This project is also about implementing algorithms
described in the book `The Theory of Parsing, Translation and Compiling (Volume
1)` by `Alfred V. Aho and Jeffrey D. Ullman` (1972). Also this proje [...]
-
-This is just started, so most stuff are *WIP*.
+The idea of this plugin is to provide functions for various kinds of
context-free grammar parser generations with support for
syntax-directed-translations (SDT) and semantic actions (SA) and the
possibility of exporting parsers and translators (as generated elisp code) to
enable Emacs plugin-agnostic usage. This project is also about implementing
algorithms described in the book `The Theory of Parsing, Translation and
Compiling (Volume 1)` by `Alfred V. Aho and Jeffrey D. Ullman` (1972). [...]
## Lexical Analysis
diff --git a/docs/Syntax-Analysis.md b/docs/Syntax-Analysis.md
index b384185..9659f1e 100644
--- a/docs/Syntax-Analysis.md
+++ b/docs/Syntax-Analysis.md
@@ -39,10 +39,30 @@ Grammar consists of `N`, `T`, `P` and `S`, where `N` is
non-terminals, `T` is te
* P = `'((S (A B)) (A (B a) e) (B (C b) C) (C c e))`
* S = `'S`
+Example:
+
``` emacs-lisp
(parser-generator-set-grammar '((S A B C) (a b c) ((S (A B)) (A (B a) e) (B (C
b) C) (C c e)) S))
```
+Productions can include context-sensitive attributes like this:
+
+``` emacs-lisp
+((S (A B %prec first)) (A (B a %weight) e) (B (C b) C) (C c e))
+```
+
+### Global attributes
+
+A list of valid attributes can be set in the variable
`parser-generator--global-attributes`.
+
+### Context-sensitive attributes
+
+A list of valid attributes can be set in the variable
`parser-generator--context-sensitive-attributes`.
+
+### Global declaration
+
+Can be set in variable `parser-generator--global-declaration`. This may be
used differently in different parsing algorithms.
+
### e-identifier
The symbol defined in variable `parser-generator--e-identifier`, with
default-value: 'e`, symbolizes the e symbol. The symbol is allowed in some
grammars and not in others.
diff --git a/docs/Syntax-Analysis/LR0.md b/docs/Syntax-Analysis/LR0.md
index 39c8a4d..fed7694 100644
--- a/docs/Syntax-Analysis/LR0.md
+++ b/docs/Syntax-Analysis/LR0.md
@@ -21,6 +21,42 @@ Example with grammar with production: S -> SaSb and S is
non-terminal and a, b a
* A is the production LHS
* B, C is parts of the production RHS, if the dot is at the left B is nil and
C is the entire RHS. If the dot is at the right then B is the production RHS
and C is nil, otherwise B and C contains parts of the RHS
+## Declaring operator precedence
+
+You can set global symbol operator precedence and also context-sensitive
precedence, like in GNU Bison. Example
+
+``` emacs-lisp
+(setq
+ parser-generator--global-attributes
+ '(%left %precedence %right))
+ (setq
+ parser-generator-lr--global-precedence-attributes
+ '(%left %precedence %right))
+ (setq
+ parser-generator--context-sensitive-attributes
+ '(%prec))
+ (setq
+ parser-generator-lr--context-sensitive-precedence-attribute
+ '%prec)
+ (setq
+ parser-generator--global-declaration
+ '((%left a)
+ (%right b)
+ (%left c)
+ (%precedence FIRST)))
+(parser-generator-set-grammar
+ '(
+ (Sp S A B)
+ (a b c)
+ (
+ (Sp S)
+ (S (A c) B)
+ (A (a b %prec a))
+ (B (a b c %prec FIRST))
+ )
+ Sp))
+```
+
## Parse
Perform a right-parse of input-stream. Example from
[Wikipedia](https://en.wikipedia.org/wiki/LR_parser#Constructing_LR(0)_parsing_tables).
diff --git a/docs/Syntax-Analysis/LRk.md b/docs/Syntax-Analysis/LRk.md
index 25fa359..db12254 100644
--- a/docs/Syntax-Analysis/LRk.md
+++ b/docs/Syntax-Analysis/LRk.md
@@ -22,6 +22,42 @@ Example with grammar with production: S -> SaSb and S is
non-terminal and a, b a
* B, C is parts of the production RHS, if the dot is at the left B is nil and
C is the entire RHS. If the dot is at the right then B is the production RHS
and C is nil, otherwise B and C contains parts of the RHS
* L is the item look-ahead
+## Declaring operator precedence
+
+You can set global symbol operator precedence and also context-sensitive
precedence, like in GNU Bison. Example
+
+``` emacs-lisp
+(setq
+ parser-generator--global-attributes
+ '(%left %precedence %right))
+ (setq
+ parser-generator-lr--global-precedence-attributes
+ '(%left %precedence %right))
+ (setq
+ parser-generator--context-sensitive-attributes
+ '(%prec))
+ (setq
+ parser-generator-lr--context-sensitive-precedence-attribute
+ '%prec)
+ (setq
+ parser-generator--global-declaration
+ '((%left a)
+ (%right b)
+ (%left c)
+ (%precedence FIRST)))
+(parser-generator-set-grammar
+ '(
+ (Sp S A B)
+ (a b c)
+ (
+ (Sp S)
+ (S (A c) B)
+ (A (a b %prec a))
+ (B (a b c %prec FIRST))
+ )
+ Sp))
+```
+
## LR items for prefix (S)
Calculate the set of LR items valid for any viable prefix S.
