2 \brief lexical analyzer implementation for APC
3 \details The lexer manages two FIFO stacks. One for maintaining tokens, the
4 other for maintaining a list of files to be scanned. During
5 execution, the lexer will return a token from its token queue if any
6 are present. If not, the lexer will will pop an element from its
7 file queue to 'scanner' to be tokenized. If the file queue is empty,
8 the lexer will instead call 'parsedir' to traverse the directory tree
9 and tokenize the results. If 'parsedir' does not generate any new
11 \author Jordan Lavatai
13 ----------------------------------------------------------------------------*/
23 //#include "parser.tab.h"
25 #define DP_STACKSIZE 1024
28 #define TK_STACKSIZE 1024
38 void lexer_pushtok(int int);
39 struct dirent
* lexer_direntpa
[DP_STACKSIZE
];
44 int dredge_current_depth(void);
48 struct tok token_stack
[TK_STACKSIZE
];
55 /* Directory Entity Array/Stack
56 Simple array for keeping track of dirents yet to be processed by the scanner.
57 If this list is empty and there are no tokens, the lexer is done.
59 #define DP_STACK (lexer_direntpa)
60 #define DP_STACKP (dps)
61 #define DP_LEN() (DP_STACKP - DP_STACK)
62 #define DP_INIT() (DP_STACKP = DP_STACK)
63 #define DP_POP() (*--DP_STACKP)
66 This is a FIFO stack whose pointers are a union of either a pointer to an
67 integer, or a pointer to two integers (a struct tok). This way, integers may
68 be added or removed from the stack either singularly (IPUSH/IPOP), or as a
69 full token of two integers (PUSH/POP).
70 An alignment error will occur if IPOP or IPUSH are used a non-even number of
73 #define TK_STACK (token_stack)
74 #define TK_STACKP (tks.t)
75 #define TK_STACKPI (tks.i)
76 #define TK_STACKX (tkx.t)
77 #define TK_STACKXI (tkx.i)
78 #define TK_LEN() (TK_STACKP - TK_STACKX)
79 #define TK_INIT() (TK_STACKP = TK_STACKX = TK_STACK)
80 #define TK_POP() (*TK_STACKP++)
81 #define TK_POPI() (*TK_STACKPI++);
82 #define TK_PUSH(T) (*TKSTACKX++ = T)
83 #define TK_PUSHI(I) (*TKSTACKXI++ = (I))
86 const char* cargs
['Z'];
89 int scanner_init(void);
91 int scanner(struct dirent
**);
94 The initializer returns boolean true if an error occurs, which may be handled with standard errno.
100 return scanner_init();
104 If the token buffer is empty, 'lexer' will initialize the token buffer and
105 call 'lexer_scandir'. If #SCANDIR_ERROR is returned, an error is printed
106 before sending a null return to bison. If 0 tokens are generated, the error
107 printing is skipped. In all other cases, 'yylval' is set, and the token's
108 integer representation is returned.
111 #define SCAN_ERROR -1
112 #define TK_EMPTY (TK_STACKP == TK_STACKX)
115 { switch (parsedir())
117 perror("lexer_scan");
130 This receiver takes a struct tok and pushes it to the FIFO stack.
134 #define ERR_TK "Fatal: Generated over " S(TK_STACKSIZE) " tokens in one pass."
136 { if (TK_LEN
>= TK_STACKSIZE
)
137 { fprintf(stderr
, ERR_TK
);
144 Ragel state machine for tokenizing text.
151 lexer_pushtok(token
);
157 while ((c = *lsp++) == *csp)
160 delimeters_skipped++;
162 csp++; //delayed to ensure csp is the start of scannable text
166 last_string = string;
168 return scanner_tokenize(csp);