1. Language Processor Implementation
using Python

2. Objectiv
e

Implement parser, syntax analyzer, semantic
analyzer for Pascal language

Show main approaches to implementation of
semantic analysis as well as intermediary code
generation

Parameterize language processor

3. Choosing Programming
Language
I investigated strong and week points of
several programming languages:
 Java
 Perl
 Python
 C++
 Delphi
As a result I chose Python due to several very good
reasons.

4. Python
Benefits
High level programming language
 Supports object oriented paradigm
 Convenient data types
 Relatively fast due to using C based libraries
 Easy readable syntax
 Cross-platform
 Convenient tools for parsing YAML
 Supports regular expressions out of the box


5. Language Processor Work
Scheme
Entry
Parser
Syntax
Semantics
Code gen
Exit
Вихід
In memory tables

6. Parser.
Tokens
Identifier
^[A-Za-z][A-Za-z_0-9]{0,255}$
Integer Const
^[+-]?d{1,10}$
Float Const
^([+-]?((d+.d+)|(d+.d+e[+-]d+)))$
String Literal
^'.{0,65535}'$

7. Algorythm of Analyzing
Lexems
def getClass(self, word):
c = None
if ((word in self.KeyWords)
or (word in metadata["delimiters"])
or (word in metadata["double"])
or (word in metadata["conditional_delimiters"])
or (word in metadata["multiplicative"])
or (word in metadata["additive"])
or (word in metadata["Relation"])):
c = word
else:
for r in self.RegExp.keys():
if re.compile(r).match(word):
c = self.RegExp[r]
return c

8. Attribute Table

Separate class

Design pattern Singleton

Uses hash table as internal structure

Fast access

Convenient format

9. Attribute Table Look

10. Syntax
Analyzer

Context free grammar

Left associated grammar

EBNF

Configuration format is yaml

Language grammar is easily changed without
source code modification

11. YAML Configuration Format
Benefits

Short and readable

yaml uses data structures that are native to
programming languages like Perl, Python

12. Configuration Format of Language
Grammar
EBNF
rule
program ::=
EBNF
Program ID ; block .
rule in configuration
programme:
- [program, id, ;, Block, "."]

13. Semantic Analyzer. Semantic
Actions
Rule
without semantic actions
complex_action:
- [begin, action_list, end]
Rule
with semantic actions
complex_action:
- [begin, "#200", action_list, end, "#220"]

14. Attribute
classes
Making
code generation process easier, I created a
bunch of classes like AttrFor.
class
AttrFor (AttrObject):
self.parameter = None
self.first = None
self.last = None
self.step = None
self.body = None

15. Abstract Parse
Tree

16. Intermediary code
generation
We
are using tetrads language in order to generate
intermediary code
How tetrads language looks:
Z := X op Y
Z := op X
Z := Y
Z := Y[X]
Z:
GOTO Z
If condition GOTO Z

17. Example of input file
program q;
var
a, b: integer;
i: integer;
d: integer;
begin
d := 4;
for i:= 1 to (2+2*2)*2 do
begin
b:=b + 1;
a:=a * 2;
end;
d:=a;
end.

18. Output of language processor using intermediary code
d:=4
i:=1
@Lid1:
if i > 12 goto @Lid2
b:=b + 1
a:=a * 2
i := i + 1
goto @Lid1
@Lid2:
d:=a

19. Bottom line

I implemented parser, syntax analyzer, semantic analyzer,
intermediary code generation of Pascal programming
language

I showed main concepts of semantic analysis as well as
intermediary code generation

Language processor has been parameterized