X-Git-Url: https://www.kengrimes.com/gitweb/?p=henge%2Fwebcc.git;a=blobdiff_plain;f=src%2Fapc%2Flexer.c;h=f11c6a21d1736743b9059aa0e9fa110c1e353654;hp=4e0a611faa06bf16bd4501dd0bed73fa850f34ce;hb=0f505368fa8abbc2e9ab0296b9a5e6bd4869345f;hpb=882513ad6569a640cb5c90b83e27e742ace71b63 diff --git a/src/apc/lexer.c b/src/apc/lexer.c index 4e0a611..f11c6a2 100644 --- a/src/apc/lexer.c +++ b/src/apc/lexer.c @@ -14,53 +14,73 @@ /* Standard */ #include #include +#include #include /* Posix */ #include +#include +#include +#include +#include +#include #include +#include //realpath, NAME_MAX, PATH_MAX #include +/* Redefinitions of NAME_MAX and PATH_MAX */ +//#define NAME_MAX NAME_MAX/4 +//#define PATH_MAX PATH_MAX/4 + /* Local */ -//#include "parser.tab.h" -#ifndef DP_STACKSIZE -#define DP_STACKSIZE 1024 +#include "parser.tab.h" +#ifndef DE_STACKSIZE +#define DE_STACKSIZE 1024 #endif #ifndef TK_STACKSIZE #define TK_STACKSIZE 1024 #endif +#ifndef MAX_SETNAME_LEN //max setname length +#define MAX_SETNAME_LEN 32 +#endif + /* Public */ -struct tok -{ int lval; - int tok; -}; -int lexer_init(void); -int lexer(void); -inline -void lexer_pushtok(int int); -struct dirent* lexer_direntpa[DP_STACKSIZE]; +int lexer_init(void); +int lexer(void); +int lexer_lexfile(const uint8_t*); +void lexer_pushtok(int, YYSTYPE); +uint8_t const* lexer_get_current_filepath(void); +int lexer_lexfilename(uint8_t*); +struct dirent* lexer_direntpa[DE_STACKSIZE],** lexer_direntpp,** lexer_direntpb; /* Private */ +extern //lexer_fsm.rl +int lexer_lexstring(uint8_t*, int); +extern //scanner.c +int scanner_init(void); +extern //scanner.c +int scanner(void); static inline -int scan(void); -static inline -int dredge_current_depth(void); +int dredge_current_depth(void); +extern //bison +YYSTYPE yylval; static -struct dirent** dps; +uint8_t const* current_filename; static -struct tok token_stack[TK_STACKSIZE]; -static -union tokp -{ int* i; - struct tok* t; -} tks, tkx; +struct tok +{ YYSTYPE lval; //token val + int tok_t; //token type +} token_stack[TK_STACKSIZE], *tsp, *tsx; /* Directory Entity Array/Stack Simple array for keeping track of dirents yet to be processed by the scanner. If this list is empty and there are no tokens, the lexer is done. + This array is populated by the scanner as an array, and popped locally by the + lexer as a stack, and is popped as a FIFO stack. */ -#define DP_STACK (lexer_direntpa) -#define DP_STACKP (dps) -#define DP_LEN() (DP_STACKP - DP_STACK) -#define DP_INIT() (DP_STACKP = DP_STACK) -#define DP_POP() (*--DP_STACKP) +#define DE_STACK (lexer_direntpa) +#define DE_STACKP (lexer_direntpp) +#define DE_STACKB (lexer_direntpb) +#define DE_LEN() (DE_STACKP - DE_STACKB) +#define DE_INIT() (DE_STACKP = DE_STACKB = DE_STACK) +#define DE_POP() (*DE_STACKB++) /* Token Stack This is a FIFO stack whose pointers are a union of either a pointer to an @@ -70,100 +90,215 @@ union tokp An alignment error will occur if IPOP or IPUSH are used a non-even number of times in a sequence! */ -#define TK_STACK (token_stack) -#define TK_STACKP (tks.t) -#define TK_STACKPI (tks.i) -#define TK_STACKX (tkx.t) -#define TK_STACKXI (tkx.i) -#define TK_LEN() (TK_STACKP - TK_STACKX) -#define TK_INIT() (TK_STACKP = TK_STACKX = TK_STACK) -#define TK_POP() (*TK_STACKP++) -#define TK_POPI() (*TK_STACKPI++); -#define TK_PUSH(T) (*TKSTACKX++ = T) -#define TK_PUSHI(I) (*TKSTACKXI++ = (I)) - -extern //main.c -const char* cargs['Z']; - -extern //scanner.c -int scanner_init(void); -extern //scanner.c -int scanner(struct dirent**); +#define TK_STACK (token_stack) +#define TK_STACKP (tsp) +#define TK_STACKX (tsx) +#define TK_LEN() (TK_STACKX - TK_STACKP) +#define TK_INIT() (TK_STACKP = TK_STACKX = TK_STACK) +#define TK_POP() (*TK_STACKP++) +#define TK_PUSH(T,L) (*TK_STACKX++ = (struct tok){L,T}) /* Initializer - The initializer returns boolean true if an error occurs, which may be handled with standard errno. + The initializer returns boolean true if an error occurs, which may be handled + with standard errno. */ int lexer_init () { TK_INIT(); - DP_INIT(); + DE_INIT(); return scanner_init(); } /* Lexer If the token buffer is empty, 'lexer' will initialize the token buffer and - call 'lexer_scandir'. If #SCANDIR_ERROR is returned, an error is printed + call 'lexer_scandir'. If SCAN_ERROR is returned, an error is printed before sending a null return to bison. If 0 tokens are generated, the error printing is skipped. In all other cases, 'yylval' is set, and the token's integer representation is returned. */ int lexer +#define $($)#$ #define SCAN_ERROR -1 #define TK_EMPTY (TK_STACKP == TK_STACKX) +#define FAIL(...) \ + do { \ + fprintf(stderr,__VA_ARGS__); \ + goto done; \ + } while (0) () -{ if (TK_EMPTY) - { switch (parsedir()) - { case SCAN_ERROR: - perror("lexer_scan"); - case 0: - yylval = 0; - return 0; - default: - break; +{ struct tok token; + start: + while (DE_LEN() > 0)//lex any directory entries in our stack + { + if (lexer_lexfile(DE_POP()->d_name) == 0) + FAIL("Lexer failed to tokenize [%s]\n",(*DE_STACKB)->d_name); + } + if (TK_EMPTY) //if there are no tokens, + { TK_INIT(); //initialize the token stack back to 0 + switch (scanner()) + { case SCAN_ERROR: //if an error occurred, + FAIL("Scanner error\n"); + case 0: //if the the scanner finds no dirents, + goto done; //then we are done + default: //if we found some elements to scan, + goto start; //start over and lex them } } - yylval = TK_IPOP(); - return TK_IPOP(); + token = TK_POP(); + yylval = token.lval; + return token.tok_t; + done: + yylval.val = 0; + return 0; } + /* Token Receiver This receiver takes a struct tok and pushes it to the FIFO stack. */ -inline void lexer_pushtok -#define ERR_TK "Fatal: Generated over " S(TK_STACKSIZE) " tokens in one pass." -( struct tok token ) -{ if (TK_LEN >= TK_STACKSIZE) +#define $($)#$ //stringifier +#define ERR_TK "Fatal: Generated over " $(TK_STACKSIZE) " tokens in one pass." +( int tok, YYSTYPE lval ) +{ if (TK_LEN() >= TK_STACKSIZE) { fprintf(stderr, ERR_TK); exit(EXIT_FAILURE); } - TK_PUSH(token); + TK_PUSH(tok, lval); } -/* Lexical Analysis - Ragel state machine for tokenizing text. - */ -void lexer_lex -(const char* str) -{ struct tok token; - token.TOK = 1; - token.LVAL = 2; - lexer_pushtok(token); - printf (str); +/* Lexical analysis of a file + Strips a filename to its base name, then sends it to lexer_lex +*/ +int lexer_lexfile +#define HIDDEN_WARNING "%s is hidden and will not be parsed!\n", filename +( const uint8_t *filename +) +{ static uint8_t fname[NAME_MAX]; + uint8_t *last_period = NULL, *iter; + + if (*filename == '.') + { fprintf (stderr, HIDDEN_WARNING); + return 0; + } + /* Copy the filename and remove its suffix */ + u8_strncpy(fname,filename,NAME_MAX); + last_period = NULL; + for (iter = fname; *iter; iter++) //find the last '.' char + if (*iter == '.') + last_period = iter; + if (last_period) //if we found one, + *last_period = 0; //truncate the string there + /* Register the current_filename */ + current_filename = filename; + + return lexer_lexfilename(fname); } -/* init_file: - if (lsp != NULL) - while ((c = *lsp++) == *csp) - { switch (c) - { case DELIM: - delimeters_skipped++; - default: - csp++; //delayed to ensure csp is the start of scannable text - break; +uint8_t const* lexer_get_current_filepath +() +{ static uint8_t current_path[PATH_MAX]; + static uint8_t const* last_filename; + if ((!last_filename || last_filename != current_filename) && + (realpath(current_filename, current_path) != (char*) current_path)) + { perror("realpath: "); + return NULL; } + return (const char*)current_path; +} + +/* Scan filename and push the its tokens + onto the stack */ +int lexer_lexfilename +(uint8_t* str) +{ + int ntok, i, cmp, len, set_len, height, width; + char map_key[] = "_m_"; + static uint8_t set_name[MAX_SETNAME_LEN] = {0}; + uint8_t *first, *map_begin; + + printf("Starting lexerfilename on %s\n", str); + + + if(*str == 0) + printf("Lexfilename:: str is NULL so fail\n"); + printf("setname is %s\n", set_name); + + /* If last file was a mapfile, then its 5th to last token should + be a MOPEN. If this is the case, then we only pass MOPEN, height, + weight and name of the current file. */ + if( (TK_STACKX - 5)->tok_t == MOPEN ) + { printf("The last file was a mapfile\n"); + if( (map_begin = strstr(map_key, str)) ) //if the current file is a mapfile + { printf("The current file is a variant of the last mapfile\n"); + printf("Start lexing mapfile %s\n", str); + ntok += lexer_lexstring(map_begin, strlen(map_begin)); + } + printf("Current file is not a variant of the last mapfile\n"); } - last_string = string; - scan_text: - return scanner_tokenize(csp); -*/ + else //last file was not a mapfile + { printf("Last file was not a mapfile\n"); + + first = (uint8_t*) u8_strchr(str, '_'); //find the first '_' to find where str set_name ends + + if(set_name[0] != 0) //if there is a set_name from last str + { printf("There is a set_name (%s) present\n", set_name); + set_len = first - str; + + if(u8_strncmp(str, set_name, set_len) == 0) //check if it matches the current set_name + { str = str + set_len + 1; //if so, remove it from str + printf("str set_name matched last set_name, set str to %s\n", str); + } + else //update set_name to be str set_name + { u8_cpy(set_name, str, set_len); + set_name[set_len] = 0; + + } + } + else //set set_name + { u8_cpy(set_name, str, first-str); + } + /* Call lexer_lexstring to tokenize the string */ + printf("calling lexstring to tokenize str (%s) of len %d\n", str, u8_strlen(str)); + ntok += lexer_lexstring(str, u8_strlen(str)); + } + + /*TODO: if regfile, store full path for later */ + + printf("Ending lexer_lex on %s, %d tokens were lexed\n", str, ntok); + return ntok; +} + +/* int lexer_lexmapfile */ +/* #define INC_X() */ +/* (int height, int width) */ +/* { */ +/* int x, y; */ + +/* /\* Give scanner_scanpixels a buffer and a len. Iterate through */ +/* buf with buf[n]. If n == 0, do nothing. if n has a value, push x, */ +/* push y, push (z = n << 24), push (ref_id = n >> 8) *\/ */ +/* //scanner_scanpixels() */ + +/* for(i = 0; i < len; i++) */ +/* if(buf[i] == 0) */ +/* if(x == width) */ +/* x = 0; */ +/* else */ + + + + +/* } */ +/* fname_bytes = (uint8_t*)(DE_POP()->d_name); */ + /* printf("d_name is %s\n", fname_bytes); */ + /* for (fnp = filename, i = 0; i < NAME_MAX; i += unit_size, fnp++) */ + /* { unit_size = u8_mblen(fname_bytes + i, min(4, NAME_MAX - i)); */ + /* if (u8_mbtouc(fnp, fname_bytes + i, unit_size) == -1) //add ucs4 char to the filename */ + /* FAIL("Lexer failed to convert ^%s to unicode\n", (fname_bytes + i)); */ + /* if (*fnp == 0) //added a terminating char */ + /* break; */ + /* } */ + /* if(u8_mbtouc(filename, DE_POP()->d_name, NAME_MAX) == -1) */ + /* FAIL("Lexer failed to convert d_name into uint8_t\n"); */ + /* ulc_fprintf(stdout, "filename is %11U\n c", filename); */