-#include <stddef.h>
-#include <stdio.h>
+#include "php_psi_stdinc.h"
+#include <sys/mman.h>
#include <assert.h>
#include "parser.h"
-#include "parser_proc.h"
-void *PSI_ParserProcAlloc(void*(unsigned long));
-void PSI_ParserProcFree(void*, void(*)(void*));
-void PSI_ParserProc(void *, token_t, PSI_Token *, PSI_Parser *);
-void PSI_ParserProcTrace(FILE *, const char*);
+void *psi_parser_proc_Alloc(void*(unsigned long));
+void psi_parser_proc_Free(void*, void(*)(void*));
+void psi_parser_proc_(void *, token_t, struct psi_token *, struct psi_parser *);
+void psi_parser_proc_Trace(FILE *, const char*);
-PSI_Parser *PSI_ParserInit(PSI_Parser *P, const char *filename, psi_error_cb error, unsigned flags)
+struct psi_parser *psi_parser_init(struct psi_parser *P, psi_error_cb error, unsigned flags)
{
- FILE *fp;
-
if (!P) {
P = malloc(sizeof(*P));
}
memset(P, 0, sizeof(*P));
- fp = fopen(filename, "r");
+ psi_data_ctor_with_dtors(PSI_DATA(P), error, flags);
- if (!fp) {
- perror(filename);
- return NULL;
+ P->col = 1;
+ P->line = 1;
+ P->proc = psi_parser_proc_Alloc(malloc);
+
+ if (flags & PSI_DEBUG) {
+ psi_parser_proc_Trace(stderr, "PSI> ");
}
- if (!P) {
- P = malloc(sizeof(*P));
+ return P;
+}
+
+bool psi_parser_open_file(struct psi_parser *P, const char *filename)
+{
+ FILE *fp = fopen(filename, "r");
+
+ if (!fp) {
+ P->error(PSI_DATA(P), NULL, PSI_WARNING,
+ "Could not open '%s' for reading: %s",
+ filename, strerror(errno));
+ return false;
}
- memset(P, 0, sizeof(*P));
- P->psi.file.fn = strdup(filename);
- P->fp = fp;
- P->col = 1;
- P->line = 1;
- P->error = error;
- P->flags = flags;
+ P->input.type = PSI_PARSE_FILE;
+ P->input.data.file.handle = fp;
+
+#if HAVE_MMAP
+ struct stat sb;
+ int fd = fileno(fp);
+
+ if (fstat(fd, &sb)) {
+ P->error(PSI_DATA(P), NULL, PSI_WARNING,
+ "Could not stat '%s': %s",
+ filename, strerror(errno));
+ return false;
+ }
- P->proc = PSI_ParserProcAlloc(malloc);
- if (flags & PSI_PARSER_DEBUG) {
- PSI_ParserProcTrace(stderr, "PSI> ");
+ P->input.data.file.buffer = mmap(NULL, sb.st_size, PROT_READ, MAP_SHARED, fd, 0);
+ if (MAP_FAILED == P->input.data.file.buffer) {
+ P->error(PSI_DATA(P), NULL, PSI_WARNING,
+ "Could not map '%s' for reading: %s",
+ filename, strerror(errno));
+ return false;
}
+ P->input.data.file.length = sb.st_size;
+#else
+ P->input.data.file.buffer = malloc(BSIZE);
+#endif
- PSI_ParserFill(P, 0);
+ P->file.fn = strdup(filename);
- return P;
+ return true;
}
-size_t PSI_ParserFill(PSI_Parser *P, size_t n)
+bool psi_parser_open_string(struct psi_parser *P, const char *string, size_t length)
{
- if (P->flags & PSI_PARSER_DEBUG) {
- fprintf(stderr, "PSI> Fill: n=%zu\n", n);
+ P->input.type = PSI_PARSE_STRING;
+ P->input.data.string.length = length;
+ if (!(P->input.data.string.buffer = strndup(string, length))) {
+ return false;
}
+
+ P->file.fn = strdup("<input>");
+
+ return true;
+}
+
+static ssize_t psi_parser_fill(struct psi_parser *P, size_t n)
+{
+ PSI_DEBUG_PRINT(P, "PSI> Fill: n=%zu (input.type=%d)\n", n, P->input.type);
+
+ /* init if n==0 */
if (!n) {
- P->cur = P->tok = P->lim = P->mrk = P->buf;
- P->eof = NULL;
- }
+ switch (P->input.type) {
+ case PSI_PARSE_FILE:
+ P->cur = P->tok = P->mrk = P->input.data.file.buffer;
+#if HAVE_MMAP
+ P->eof = P->input.data.file.buffer + P->input.data.file.length;
+ P->lim = P->eof;
+#else
+ P->eof = NULL;
+ P->lim = P->input.data.file.buffer;
+#endif
+ break;
- if (!P->eof) {
- size_t consumed = P->tok - P->buf;
- size_t reserved = P->lim - P->tok;
- size_t available = BSIZE - reserved;
- size_t didread;
-
- if (consumed) {
- memmove(P->buf, P->tok, reserved);
- P->tok -= consumed;
- P->cur -= consumed;
- P->lim -= consumed;
- P->mrk -= consumed;
+ case PSI_PARSE_STRING:
+ P->cur = P->tok = P->mrk = P->input.data.string.buffer;
+ P->eof = P->input.data.string.buffer + P->input.data.string.length;
+ P->lim = P->eof;
+ break;
}
- didread = fread(P->lim, 1, available, P->fp);
- P->lim += didread;
- if (didread < available) {
- P->eof = P->lim;
- }
+ PSI_DEBUG_PRINT(P, "PSI> Fill: cur=%p lim=%p eof=%p\n", P->cur, P->lim, P->eof);
+ }
+
+ switch (P->input.type) {
+ case PSI_PARSE_STRING:
+ break;
- if (P->flags & PSI_PARSER_DEBUG) {
- fprintf(stderr, "PSI> Fill: consumed=%zu reserved=%zu available=%zu didread=%zu\n",
- consumed, reserved, available, didread);
+ case PSI_PARSE_FILE:
+#if !HAVE_MMAP
+ if (!P->eof) {
+ size_t consumed = P->tok - P->buf;
+ size_t reserved = P->lim - P->tok;
+ size_t available = BSIZE - reserved;
+ size_t didread;
+
+ if (consumed) {
+ memmove(P->buf, P->tok, reserved);
+ P->tok -= consumed;
+ P->cur -= consumed;
+ P->lim -= consumed;
+ P->mrk -= consumed;
+ }
+
+ didread = fread(P->lim, 1, available, P->fp);
+ P->lim += didread;
+ if (didread < available) {
+ P->eof = P->lim;
+ }
+ PSI_DEBUG_PRINT(P, "PSI> Fill: consumed=%zu reserved=%zu available=%zu didread=%zu\n",
+ consumed, reserved, available, didread);
}
+#endif
+ break;
}
- if (P->flags & PSI_PARSER_DEBUG) {
- fprintf(stderr, "PSI> Fill: avail=%zu\n", P->lim - P->cur);
- }
+
+ PSI_DEBUG_PRINT(P, "PSI> Fill: avail=%td\n", P->lim - P->cur);
+
return P->lim - P->cur;
}
-void PSI_ParserParse(PSI_Parser *P, PSI_Token *T)
+void psi_parser_parse(struct psi_parser *P, struct psi_token *T)
{
if (T) {
- PSI_ParserProc(P->proc, T->type, T, P);
+ psi_parser_proc_(P->proc, T->type, T, P);
} else {
- PSI_ParserProc(P->proc, 0, NULL, P);
+ psi_parser_proc_(P->proc, 0, NULL, P);
}
}
-void PSI_ParserDtor(PSI_Parser *P)
+void psi_parser_dtor(struct psi_parser *P)
{
- PSI_ParserProcFree(P->proc, free);
+ psi_parser_proc_Free(P->proc, free);
- if (P->fp) {
- fclose(P->fp);
+ switch (P->input.type) {
+ case PSI_PARSE_FILE:
+ if (P->input.data.file.buffer) {
+#if HAVE_MMAP
+ munmap(P->input.data.file.buffer, P->input.data.file.length);
+#else
+ free(P->input.data.file.buffer);
+#endif
+ }
+ if (P->input.data.file.handle) {
+ fclose(P->input.data.file.handle);
+ }
+ break;
+
+ case PSI_PARSE_STRING:
+ if (P->input.data.string.buffer) {
+ free(P->input.data.string.buffer);
+ }
+ break;
}
- PSI_DataDtor((PSI_Data *) P);
+ psi_data_dtor(PSI_DATA(P));
memset(P, 0, sizeof(*P));
}
-void PSI_ParserFree(PSI_Parser **P)
+void psi_parser_free(struct psi_parser **P)
{
if (*P) {
- PSI_ParserDtor(*P);
+ psi_parser_dtor(*P);
free(*P);
*P = NULL;
}
}
/*!max:re2c*/
-#define BSIZE 256
-
#if BSIZE < YYMAXFILL
# error BSIZE must be greater than YYMAXFILL
#endif
-#define PSI_T(n) \
-(n) == PSI_T_NAME ? "NAME" : \
-(n) == PSI_T_PLUS ? "PLUS" : \
-(n) == PSI_T_MINUS ? "MINUS" : \
-(n) == PSI_T_SLASH ? "SLASH" : \
-(n) == PSI_T_ASTERISK ? "ASTERISK" : \
-(n) == PSI_T_TEMP ? "TEMP" : \
-(n) == PSI_T_FREE ? "FREE" : \
-(n) == PSI_T_SET ? "SET" : \
-(n) == PSI_T_LET ? "LET" : \
-(n) == PSI_T_RETURN ? "RETURN" : \
-(n) == PSI_T_LIB ? "LIB" : \
-(n) == PSI_T_INT ? "INT" : \
-(n) == PSI_T_UNSIGNED ? "UNSIGNED" : \
-(n) == PSI_T_EOF ? "EOF" : \
-(n) == PSI_T_QUOTED_STRING ? "QUOTED_STRING" : \
-(n) == PSI_T_EOS ? "EOS" : \
-(n) == PSI_T_STRUCT ? "STRUCT" : \
-(n) == PSI_T_LBRACE ? "LBRACE" : \
-(n) == PSI_T_RBRACE ? "RBRACE" : \
-(n) == PSI_T_COLON ? "COLON" : \
-(n) == PSI_T_LPAREN ? "LPAREN" : \
-(n) == PSI_T_NUMBER ? "NUMBER" : \
-(n) == PSI_T_RPAREN ? "RPAREN" : \
-(n) == PSI_T_BOOL ? "BOOL" : \
-(n) == PSI_T_FLOAT ? "FLOAT" : \
-(n) == PSI_T_STRING ? "STRING" : \
-(n) == PSI_T_CONST ? "CONST" : \
-(n) == PSI_T_NSNAME ? "NSNAME" : \
-(n) == PSI_T_EQUALS ? "EQUALS" : \
-(n) == PSI_T_TYPEDEF ? "TYPEDEF" : \
-(n) == PSI_T_VOID ? "VOID" : \
-(n) == PSI_T_LBRACKET ? "LBRACKET" : \
-(n) == PSI_T_RBRACKET ? "RBRACKET" : \
-(n) == PSI_T_COMMA ? "COMMA" : \
-(n) == PSI_T_ELLIPSIS ? "ELLIPSIS" : \
-(n) == PSI_T_DOUBLE ? "DOUBLE" : \
-(n) == PSI_T_INT8 ? "INT8" : \
-(n) == PSI_T_UINT8 ? "UINT8" : \
-(n) == PSI_T_INT16 ? "INT16" : \
-(n) == PSI_T_UINT16 ? "UINT16" : \
-(n) == PSI_T_INT32 ? "INT32" : \
-(n) == PSI_T_UINT32 ? "UINT32" : \
-(n) == PSI_T_INT64 ? "INT64" : \
-(n) == PSI_T_UINT64 ? "UINT64" : \
-(n) == PSI_T_FUNCTION ? "FUNCTION" : \
-(n) == PSI_T_NULL ? "NULL" : \
-(n) == PSI_T_TRUE ? "TRUE" : \
-(n) == PSI_T_FALSE ? "FALSE" : \
-(n) == PSI_T_DOLLAR ? "DOLLAR" : \
-(n) == PSI_T_CALLOC ? "CALLOC" : \
-(n) == PSI_T_OBJVAL ? "OBJVAL" : \
-(n) == PSI_T_ARRVAL ? "ARRVAL" : \
-(n) == PSI_T_PATHVAL ? "PATHVAL" : \
-(n) == PSI_T_STRLEN ? "STRLEN" : \
-(n) == PSI_T_STRVAL ? "STRVAL" : \
-(n) == PSI_T_FLOATVAL ? "FLOATVAL" : \
-(n) == PSI_T_INTVAL ? "INTVAL" : \
-(n) == PSI_T_BOOLVAL ? "BOOLVAL" : \
-(n) == PSI_T_TO_OBJECT ? "TO_OBJECT" : \
-(n) == PSI_T_TO_ARRAY ? "TO_ARRAY" : \
-(n) == PSI_T_TO_STRING ? "TO_STRING" : \
-(n) == PSI_T_TO_INT ? "TO_INT" : \
-(n) == PSI_T_TO_FLOAT ? "TO_FLOAT" : \
-(n) == PSI_T_TO_BOOL ? "TO_BOOL" : \
-(n) == PSI_T_MIXED ? "MIXED" : \
-(n) == PSI_T_ARRAY ? "ARRAY" : \
-(n) == PSI_T_OBJECT ? "OBJECT" : \
-(n) == PSI_T_AMPERSAND ? "AMPERSAND" : \
-<UNKNOWN>
-
#define RETURN(t) do { \
P->num = t; \
- if (P->flags & PSI_PARSER_DEBUG) { \
- fprintf(stderr, "PSI> TOKEN: %d %.*s (EOF=%d %s:%u:%u)\n", \
+ PSI_DEBUG_PRINT(P, "PSI> TOKEN: %d %.*s (EOF=%d %s:%u:%u)\n", \
P->num, (int) (P->cur-P->tok), P->tok, P->num == PSI_T_EOF, \
- P->psi.file.fn, P->line, P->col); \
- } \
+ P->file.fn, P->line, P->col); \
return t; \
} while(1)
#define ADDCOLS \
P->col += P->cur - P->tok
-#define NEWLINE \
+#define NEWLINE(label) \
P->col = 1; \
++P->line; \
- goto nextline
+ goto label
-token_t PSI_ParserScan(PSI_Parser *P)
+token_t psi_parser_scan(struct psi_parser *P)
{
+ if (!P->cur) {
+ psi_parser_fill(P, 0);
+ }
for (;;) {
ADDCOLS;
nextline:
re2c:define:YYCURSOR = P->cur;
re2c:define:YYLIMIT = P->lim;
re2c:define:YYMARKER = P->mrk;
- re2c:define:YYFILL = "{ if (!PSI_ParserFill(P,@@)) RETURN(PSI_T_EOF); }";
+ re2c:define:YYFILL = "{ if (!psi_parser_fill(P,@@)) RETURN(PSI_T_EOF); }";
re2c:yyfill:parameter = 0;
B = [^a-zA-Z0-9_];
W = [a-zA-Z0-9_];
NAME = [a-zA-Z_]W*;
NSNAME = (NAME)? ("\\" NAME)+;
+ DOLLAR_NAME = '$' W+;
QUOTED_STRING = "\"" ([^\"])+ "\"";
NUMBER = [+-]? [0-9]* "."? [0-9]+ ([eE] [+-]? [0-9]+)?;
- ("#"|"//") .* "\n" { NEWLINE; }
+ "/*" { goto comment; }
+ ("#"|"//") .* "\n" { NEWLINE(nextline); }
"(" {RETURN(PSI_T_LPAREN);}
")" {RETURN(PSI_T_RPAREN);}
";" {RETURN(PSI_T_EOS);}
"[" {RETURN(PSI_T_LBRACKET);}
"]" {RETURN(PSI_T_RBRACKET);}
"=" {RETURN(PSI_T_EQUALS);}
- "$" {RETURN(PSI_T_DOLLAR);}
"*" {RETURN(PSI_T_ASTERISK);}
"&" {RETURN(PSI_T_AMPERSAND);}
"+" {RETURN(PSI_T_PLUS);}
"-" {RETURN(PSI_T_MINUS);}
"/" {RETURN(PSI_T_SLASH);}
"..." {RETURN(PSI_T_ELLIPSIS);}
- [\r\n] { NEWLINE; }
+ [\r\n] { NEWLINE(nextline); }
[\t ]+ { continue; }
'TRUE' {RETURN(PSI_T_TRUE);}
'FALSE' {RETURN(PSI_T_FALSE);}
'NULL' {RETURN(PSI_T_NULL);}
'MIXED' {RETURN(PSI_T_MIXED);}
+ 'CALLABLE' {RETURN(PSI_T_CALLABLE);}
'VOID' {RETURN(PSI_T_VOID);}
'BOOL' {RETURN(PSI_T_BOOL);}
'CHAR' {RETURN(PSI_T_CHAR);}
'STRING' {RETURN(PSI_T_STRING);}
'ARRAY' {RETURN(PSI_T_ARRAY);}
'OBJECT' {RETURN(PSI_T_OBJECT);}
+ 'CALLBACK' {RETURN(PSI_T_CALLBACK);}
+ 'STATIC' {RETURN(PSI_T_STATIC);}
'FUNCTION' {RETURN(PSI_T_FUNCTION);}
'TYPEDEF' {RETURN(PSI_T_TYPEDEF);}
'STRUCT' {RETURN(PSI_T_STRUCT);}
+ 'UNION' {RETURN(PSI_T_UNION);}
+ 'ENUM' {RETURN(PSI_T_ENUM);}
'CONST' {RETURN(PSI_T_CONST);}
'LIB' {RETURN(PSI_T_LIB);}
'LET' {RETURN(PSI_T_LET);}
'BOOLVAL' {RETURN(PSI_T_BOOLVAL);}
'ARRVAL' {RETURN(PSI_T_ARRVAL);}
'OBJVAL' {RETURN(PSI_T_OBJVAL);}
+ 'ZVAL' {RETURN(PSI_T_ZVAL);}
+ 'COUNT' {RETURN(PSI_T_COUNT);}
'CALLOC' {RETURN(PSI_T_CALLOC);}
'TO_OBJECT' {RETURN(PSI_T_TO_OBJECT);}
'TO_ARRAY' {RETURN(PSI_T_TO_ARRAY);}
NUMBER {RETURN(PSI_T_NUMBER);}
NAME {RETURN(PSI_T_NAME);}
NSNAME {RETURN(PSI_T_NSNAME);}
+ DOLLAR_NAME {RETURN(PSI_T_DOLLAR_NAME);}
QUOTED_STRING {RETURN(PSI_T_QUOTED_STRING);}
[^] {break;}
*/
+
+ comment:
+ P->tok = P->cur;
+ /*!re2c
+ "\n" { NEWLINE(comment); }
+ "*" "/" { continue; }
+ [^] { goto comment; }
+ */
}
return -1;
}
projects / m6w6 / ext-psi / blobdiff