Max Brunsfeld
9686c57e90
Allow ubiquitous tokens to also be used in grammar rules
2014-06-26 08:52:42 -07:00
Max Brunsfeld
174f306e2a
Fix precedence of comments vs '/' operator
2014-06-11 12:27:58 -07:00
Max Brunsfeld
4ad6278334
Add finally, instance of, typeof, in to js grammar
2014-06-11 11:49:06 -07:00
Max Brunsfeld
c30055ba18
Fix symbol names for extracted tokens
2014-05-20 08:30:58 -07:00
Max Brunsfeld
649f200831
Expand regex/string rules as part of grammar preparation
...
This makes it possible to report errors in regex parsing
2014-05-19 20:54:59 -07:00
Max Brunsfeld
34137be12d
Represent state ids as unsigned shorts
...
This fixes some signedness conversion warnings
2014-05-08 13:23:46 -07:00
Max Brunsfeld
0a21eee3f0
Remove magic number from generated symbols enums
...
The symbol numbers 0 and 1 are reserved for 'error' and 'eof',
so the grammar's start symbol is always 2.
2014-05-08 13:14:45 -07:00
Max Brunsfeld
4700e33746
Introduce 'ubiquitous_tokens' concept, for parsing comments and such
2014-05-06 12:54:04 -07:00
Max Brunsfeld
1d314d71c2
Separate functions for building parse and lex tables
...
Now, instead of adding states to the lex table as they are needed
by the parse states, we iterate over the parse states after the fact
and set up their corresponding lex states. This has the nice side
effect that the lex states are in a more readable order.
2014-05-04 22:07:52 -07:00
Max Brunsfeld
68d44fd565
Intern symbols during grammar preparation
2014-04-22 23:38:26 -07:00
Max Brunsfeld
5145bba53d
Silence missing-initializer warnings for gcc
2014-04-12 20:16:16 -07:00
Max Brunsfeld
53285f7d0c
Give rule-specific names to repeat helper rules
2014-04-09 18:41:43 -07:00
Max Brunsfeld
bd5ec68c96
Get generated parsers building under gcc
2014-04-08 22:11:20 -07:00
Max Brunsfeld
f71d7bae16
Record parse conflicts when compiling grammars
...
Need to remove duplicate conflicts
2014-04-08 18:47:42 -07:00
Max Brunsfeld
129d2b9314
Remove extra EOF actions in lexer
2014-04-04 08:44:35 -07:00
Max Brunsfeld
1cc7e32e2d
Fix handling of tokens consisting of separator characters
...
The parser is no longer hard-coded to skip whitespace. Tokens
such as newlines, whose characters overlap with the separator
characters, can now be correctly recognized.
2014-04-03 19:10:09 -07:00
Max Brunsfeld
a79a7435de
Remove remaining trailing whitespace from generated c code
2014-03-29 19:21:42 -07:00
Max Brunsfeld
8e1b78ca8e
Remove trailing whitespace from generated c code
2014-03-29 19:00:31 -07:00
Max Brunsfeld
e1ac62edc5
Give better symbol names to generated tokens
...
This should make debugging easier
2014-03-27 12:54:54 -07:00
Max Brunsfeld
4454925b5a
Clean up parser macros more
2014-03-26 13:03:12 -07:00
Max Brunsfeld
f601322956
Clean up macros in parser.h
2014-03-25 19:51:34 -07:00
Max Brunsfeld
80b19cbb83
Construct entire parse table statically
...
This removes the need for the 'init_parse_table' function,
which was not really thread safe
2014-03-25 19:34:17 -07:00
Max Brunsfeld
25861b7f03
Remove reduction-specific collapse flags in favor of globally hidden symbols
2014-03-25 09:05:55 -07:00
Max Brunsfeld
2df56f01c8
Remove public hidden symbol function
...
Now, you can't a particular occurrence of a symbol
in a grammar. You can only hide a symbol globally
(right now, by beginning its name with an underscore).
2014-03-25 08:16:26 -07:00
Max Brunsfeld
48baf056b4
Store grammars' rules in vectors, not maps
...
This way, we can keep of the order in which the rules
were given, for resolving reduce/reduce conflicts.
2014-03-24 13:17:38 -07:00
Max Brunsfeld
50a90e456b
Handle * quantifier in regex patterns
2014-03-22 20:08:11 -07:00
Max Brunsfeld
c40411b4d1
Handle unexpected tokens properly
...
Also, add null and boolean values to json grammar
2014-03-22 15:46:58 -07:00
Max Brunsfeld
8e7f59db1c
Allow room in parse table for error and EOF entries
2014-03-18 08:04:17 -07:00
Max Brunsfeld
67b33a615b
Refactor generated parsers to used explicit table
...
This is slightly slower than encoding the parse table in
flow control, but allows the parser to inspect the parse
table more flexibly. This is needed for incremental parsing.
2014-03-17 18:43:17 -07:00
Max Brunsfeld
da992b77ff
Rename 'parse_config' export -> 'parser'
2014-03-15 14:48:28 -07:00
Max Brunsfeld
0dc3a95d0c
Refactor parser header
...
Make separate lexer, stack and parser structs.
2014-03-15 14:43:50 -07:00
Max Brunsfeld
926b4ef242
Clean up parser.h
2014-02-27 00:06:22 -08:00
Max Brunsfeld
0d4687bdad
Make start and EOF symbols into builtin symbols
2014-02-26 23:18:49 -08:00
Max Brunsfeld
8b7ed13c8b
Add more error recovery tests for JSON
2014-02-26 22:58:45 -08:00
Max Brunsfeld
63718d5bad
Add basic support for error recovery
2014-02-26 19:03:43 -08:00
Max Brunsfeld
4541332c2b
Put definition of EOF symbol back in parse table builder (not in parser.h)
2014-02-25 13:25:58 -08:00
Max Brunsfeld
e58a6d8ba7
Start work on error recovery
...
- In runtime, make parse errors part of the parse tree
- Add error state to lexers in which they can accept any token
2014-02-24 18:42:54 -08:00
Max Brunsfeld
6ed6aa05cc
Update runtime code naming
...
Also, add macros so that code generator doesn't
need to know about any runtime variables names
or types
2014-02-20 13:30:43 -08:00
Max Brunsfeld
acad97cfd2
Separate concepts of hidden and auxiliary symbols
...
This way, we can hide auxiliary symbols from library users,
but still allow them to use hidden symbols
2014-02-19 13:05:54 -08:00
Max Brunsfeld
fd757c20d0
Move example grammars from fixtures -> examples dir
2014-02-17 13:18:09 -08:00
