CHANGES revision 2632:1bb2f91485ea 
1Version 1.3
2------------------------------
312/10/02: jmdyck
4          Various minor adjustments to the code that Dave checked in today.
5          Updated test/yacc_{inf,unused}.exp to reflect today's changes.
6
712/10/02: beazley
8          Incorporated a variety of minor bug fixes to empty production
9          handling and infinite recursion checking.  Contributed by
10          Michael Dyck.
11
1212/10/02: beazley
13          Removed bogus recover() method call in yacc.restart()
14
15Version 1.2
16------------------------------
1711/27/02: beazley
18          Lexer and parser objects are now available as an attribute
19          of tokens and slices respectively. For example:
20 
21             def t_NUMBER(t):
22                 r'\d+'
23                 print t.lexer
24
25             def p_expr_plus(t):
26                 'expr: expr PLUS expr'
27                 print t.lexer
28                 print t.parser
29
30          This can be used for state management (if needed).
31 
3210/31/02: beazley
33          Modified yacc.py to work with Python optimize mode.  To make
34          this work, you need to use
35
36              yacc.yacc(optimize=1)
37
38          Furthermore, you need to first run Python in normal mode
39          to generate the necessary parsetab.py files.  After that,
40          you can use python -O or python -OO.  
41
42          Note: optimized mode turns off a lot of error checking.
43          Only use when you are sure that your grammar is working.
44          Make sure parsetab.py is up to date!
45
4610/30/02: beazley
47          Added cloning of Lexer objects.   For example:
48
49              import copy
50              l = lex.lex()
51              lc = copy.copy(l)
52
53              l.input("Some text")
54              lc.input("Some other text")
55              ...
56
57          This might be useful if the same "lexer" is meant to
58          be used in different contexts---or if multiple lexers
59          are running concurrently.
60                
6110/30/02: beazley
62          Fixed subtle bug with first set computation and empty productions.
63          Patch submitted by Michael Dyck.
64
6510/30/02: beazley
66          Fixed error messages to use "filename:line: message" instead
67          of "filename:line. message".  This makes error reporting more
68          friendly to emacs. Patch submitted by Fran�ois Pinard.
69
7010/30/02: beazley
71          Improvements to parser.out file.  Terminals and nonterminals
72          are sorted instead of being printed in random order.
73          Patch submitted by Fran�ois Pinard.
74
7510/30/02: beazley
76          Improvements to parser.out file output.  Rules are now printed
77          in a way that's easier to understand.  Contributed by Russ Cox.
78
7910/30/02: beazley
80          Added 'nonassoc' associativity support.    This can be used
81          to disable the chaining of operators like a < b < c.
82          To use, simply specify 'nonassoc' in the precedence table
83
84          precedence = (
85            ('nonassoc', 'LESSTHAN', 'GREATERTHAN'),  # Nonassociative operators
86            ('left', 'PLUS', 'MINUS'),
87            ('left', 'TIMES', 'DIVIDE'),
88            ('right', 'UMINUS'),            # Unary minus operator
89          )
90
91          Patch contributed by Russ Cox.
92
9310/30/02: beazley
94          Modified the lexer to provide optional support for Python -O and -OO
95          modes.  To make this work, Python *first* needs to be run in
96          unoptimized mode.  This reads the lexing information and creates a
97          file "lextab.py".  Then, run lex like this:
98
99                   # module foo.py
100                   ...
101                   ...
102                   lex.lex(optimize=1)
103
104          Once the lextab file has been created, subsequent calls to
105          lex.lex() will read data from the lextab file instead of using 
106          introspection.   In optimized mode (-O, -OO) everything should
107          work normally despite the loss of doc strings.
108
109          To change the name of the file 'lextab.py' use the following:
110
111                  lex.lex(lextab="footab")
112
113          (this creates a file footab.py)
114         
115
116Version 1.1   October 25, 2001
117------------------------------
118
11910/25/01: beazley
120          Modified the table generator to produce much more compact data.
121          This should greatly reduce the size of the parsetab.py[c] file.
122          Caveat: the tables still need to be constructed so a little more
123          work is done in parsetab on import. 
124
12510/25/01: beazley
126          There may be a possible bug in the cycle detector that reports errors
127          about infinite recursion.   I'm having a little trouble tracking it
128          down, but if you get this problem, you can disable the cycle
129          detector as follows:
130
131                 yacc.yacc(check_recursion = 0)
132
13310/25/01: beazley
134          Fixed a bug in lex.py that sometimes caused illegal characters to be
135          reported incorrectly.  Reported by Sverre J�rgensen.
136
1377/8/01  : beazley
138          Added a reference to the underlying lexer object when tokens are handled by
139          functions.   The lexer is available as the 'lexer' attribute.   This
140          was added to provide better lexing support for languages such as Fortran
141          where certain types of tokens can't be conveniently expressed as regular 
142          expressions (and where the tokenizing function may want to perform a 
143          little backtracking).  Suggested by Pearu Peterson.
144
1456/20/01 : beazley
146          Modified yacc() function so that an optional starting symbol can be specified.
147          For example:
148            
149                 yacc.yacc(start="statement")
150
151          Normally yacc always treats the first production rule as the starting symbol.
152          However, if you are debugging your grammar it may be useful to specify
153          an alternative starting symbol.  Idea suggested by Rich Salz.
154                      
155Version 1.0  June 18, 2001
156--------------------------
157Initial public offering
158
159