void rep( uint8_t x ) {
printf( "%4d_{(10)} = ", x );
for( int i = ( BITSOF( x ) - 1 ); i >= 0; i-- ) {
printf( "%d", ( x >> i ) & 1 );
}
printf( "_{(2)}\n" );
}
/* TODO */
u32 run_testfile(struct Lexer *lexer, struct Printer *p)
{
u32 err = 0;
u32 lineno = 1;
/* TODO Arena */
static u32 c2s[256];
static u32 eqc_bitmat[DIV_CEIL(8 * ARRAY_COUNT(c2s), BITSOF(u64))];
struct regex_eqc eqc;
static struct regex_nfaparser_stackpost nfaparser_stack[4096];
static struct NFA_state nfa_states[4*4096];
struct NFA nfa;
u32 nfa_fin_state;
struct regex_nfaparser parser;
u32 regex_lex_state = REGEX_DFA_START_STATE;
static u32 dfa_acc[POW2(20)];
static u32 dfa_mat[ARRAY_COUNT(dfa_acc) * ARRAY_COUNT(c2s)];
struct DFA_mat dfa;
struct nfa2dfa_data nfa2dfa_dat;
static u32 nfa2dfa_dat_epsstack[ARRAY_COUNT(nfa_states)];
static u32 nfa2dfa_dat_visited[DIV_CEIL(ARRAY_COUNT(nfa_states),
BITSOF(u32))];
static u32 nfa2dfa_dat_setoff[ARRAY_COUNT(dfa_acc)];
static u32 nfa2dfa_dat_hashtable[ARRAY_COUNT(dfa_acc)];
static u32 nfa2dfa_dat_stack[ARRAY_COUNT(dfa_acc)];
static u32 nfa2dfa_dat_setdata[16 * ARRAY_COUNT(dfa_acc)];
{
nfa.states = nfa_states;
nfa.cap = ARRAY_COUNT(nfa_states);
nfa.len = 0;
nfa.start_state = MAX_U32;
eqc.bitmat = eqc_bitmat;
eqc.width = ARRAY_COUNT(c2s);
eqc.cap = 8;
eqc.len = 0;
parser.nfa = &nfa;
parser.stack = nfaparser_stack;
parser.stack_cap = ARRAY_COUNT(nfaparser_stack);
parser.stack_len = 0;
parser.eqc = &eqc;
assert(nfa.len < nfa.cap);
nfa_fin_state = nfa.len++;
nfa.states[nfa_fin_state].c = NFA_FIN;
nfa.states[nfa_fin_state].val.fin = 0;
parser.stack[parser.stack_len].end = nfa_fin_state;
parser.stack[parser.stack_len].start = MAX_U32;
parser.stack[parser.stack_len].seq_first = MAX_U32;
parser.stack[parser.stack_len].seq_last = MAX_U32;
parser.stack[parser.stack_len].unary_first = MAX_U32;
parser.stack[parser.stack_len].unary_last = MAX_U32;
parser.stack[parser.stack_len].eqc = MAX_U32;
/* no need to set eqc_* if eqc == MAX_U32 */
++parser.stack_len;
}
while (1) {
u32 acc_val;
u32 c = lexer->c;
switch (c) {
case '\n': c = READER_EOF; break;
case READER_EOF: return PUTERR("error: unexpected EOF"), 1;
case READER_ERROR: return PUTERRNO("read"), 1;
}
regex_lex_state = regex_dfa_delta(regex_lex_state, c);
acc_val = regex_dfa_acc(regex_lex_state);
if (REGEX_DFA_ERROR_STATE(regex_lex_state)) {
PUTERR("error: regex_lex");
return 1;
}
if (acc_val) {
struct regex_token tok;
if (acc_val - 1 < REGEX_TOKEN_TYPE_COUNT) {
tok.type = acc_val - 1;
tok.c = 0; /* tok.c may be whatever */
assert(tok.type != REGEX_TOK_C);
} else {
tok.type = REGEX_TOK_C;
tok.c = (acc_val - 1) - REGEX_TOKEN_TYPE_COUNT;
}
regex_lex_state = REGEX_DFA_START_STATE;
if (regex_nfaparse(&parser, tok)) {
(void)writer_puterr_prefix(stde(), __FILE__, __LINE__);
(void)writer_puts(stde(), "error: ");
(void)writer_puts(stde(),
regex_nfaparse_status2string(parser.error));
(void)writer_putc(stde(), '\n');
(void)writer_flush(stde());
return 1;
}
}
if (lexer->c == '\n') {
lexer_next(lexer);
break;
}
lexer_next(lexer);
}
++lineno;
if (parser.stack_len) {
PUTERR("error: regex_parse unexpected end of regex");
return 1;
}
nfa.start_state = parser.start;
if (0&&DEBUG) {
pr_str(p, "nfalen = ");
pr_u32(p, nfa.len);
pr_str(p, "\n");
}
if (0&&DEBUG) {
p->err = p->err || nfa_print_dot(&nfa, p->writer, nfa2dfa_dat_epsstack,
nfa2dfa_dat_visited);
}
{
dfa.mat = dfa_mat;
dfa.width = ARRAY_COUNT(c2s);
dfa.tot_elems = ARRAY_COUNT(dfa_mat);
dfa.len = 0;
dfa.acc = dfa_acc;
dfa.acc_cap = ARRAY_COUNT(dfa_acc);
STRUCT_ZERO(nfa2dfa_dat);
nfa2dfa_dat.nfa = &nfa;
nfa2dfa_dat.dfa = &dfa;
nfa2dfa_dat.eqc = &eqc;
nfa2dfa_dat.epsstack = nfa2dfa_dat_epsstack;
nfa2dfa_dat.visited = nfa2dfa_dat_visited;
nfa2dfa_dat.setoff = nfa2dfa_dat_setoff;
nfa2dfa_dat.hashtable = nfa2dfa_dat_hashtable;
nfa2dfa_dat.stack = nfa2dfa_dat_stack;
nfa2dfa_dat.stack_cap = ARRAY_COUNT(nfa2dfa_dat_stack);
nfa2dfa_dat.setdata = nfa2dfa_dat_setdata;
nfa2dfa_dat.setdata_cap = ARRAY_COUNT(nfa2dfa_dat_setdata);
}
switch (nfa2dfa(&nfa2dfa_dat)) {
case NFA2DFA_STATUS_OK:
break;
case NFA2DFA_STATUS_OUT_OF_SETDATA_MEM:
return PUTERR("error: nfa2dfa: out of setdata mem"), 1;
case NFA2DFA_STATUS_OUT_OF_STACK_MEM:
return PUTERR("error: nfa2dfa: out of stack mem"), 1;
case NFA2DFA_STATUS_OUT_OF_DFA_MEM:
return PUTERR("error: nfa2dfa: out of dfa mem"), 1;
case NFA2DFA_STATUS_BROKEN:
return assert(0), PUTERR("error: nfa2dfa: broken"), 1;
}
if (1) {
STATIC_ASSERT_LOCAL(ARRAY_COUNT(nfa2dfa_dat_epsstack)
>= ARRAY_COUNT(c2s), can_reuse_epsstack_in_c2s);
dfa_c2s(&dfa, nfa2dfa_dat_epsstack, c2s);
} else {
u32 i;
for (i = 0; i < ARRAY_COUNT(c2s); ++i) {
c2s[i] = i;
}
}
if (1&&DEBUG) {
pr_str(p, "dfalen = ");
pr_u32(p, dfa.len);
pr_str(p, "\n");
pr_str(p, "dfawidth = ");
pr_u32(p, dfa.width);
pr_str(p, "\n");
}
if (1&&DEBUG) {
p->err = p->err || dfa_print_dot(&dfa, p->writer, NULL);
}
if (1) {
dfa_minimize(&dfa, nfa2dfa_dat_setoff, nfa2dfa_dat_stack,
nfa2dfa_dat_hashtable);
/* do c2s again */
if (0) {
u32 i;
u32 *extra_c2s = nfa2dfa_dat_setoff;
STATIC_ASSERT_LOCAL(ARRAY_COUNT(nfa2dfa_dat_epsstack)
>= ARRAY_COUNT(c2s),
can_reuse_epsstack_in_min_c2s);
STATIC_ASSERT_LOCAL(ARRAY_COUNT(nfa2dfa_dat_setoff)
>= ARRAY_COUNT(c2s),
can_reuse_setoff_in_min_c2s);
dfa_c2s(&dfa, nfa2dfa_dat_epsstack, extra_c2s);
for (i = 0; i < ARRAY_COUNT(c2s); ++i) {
c2s[i] = extra_c2s[c2s[i]];
}
}
if (1&&DEBUG) {
pr_str(p, "mindfalen = ");
pr_u32(p, dfa.len);
pr_str(p, "\n");
pr_str(p, "mindfawidth = ");
pr_u32(p, dfa.width);
pr_str(p, "\n");
}
if (1&&DEBUG) {
p->err = p->err || dfa_print_dot(&dfa, p->writer, NULL);
}
}
lexer_skip_space(lexer);
switch (lexer->c) {
case '%': lexer_next(lexer); break;
case READER_EOF: return PUTERR("error: unexpected EOF"), 1;
case READER_ERROR: return PUTERRNO("read"), 1;
default: return PUTERR("error: expected '%'"), 1;
}
while (lexer->c != '\n') {
if (lexer->c == READER_EOF) {
return PUTERR("error: unexpected EOF"), 1;
} else if (lexer->c == READER_ERROR) {
return PUTERRNO("read"), 1;
} else if (!lexer_is_space(lexer->c)) {
return PUTERR("error: expected whitespace or EOL"), 1;
}
lexer_next(lexer);
}
lexer_next(lexer);
++lineno;
{ /* strings expected to be accepted */
while (lexer->c != '%') {
u32 state = dfa.start_state;
while (lexer->c != '\n') {
if (lexer->c == READER_EOF) {
return PUTERR("error: unexpected EOF"), 1;
} else if (lexer->c == READER_ERROR) {
return PUTERRNO("read"), 1;
}
state = dfa_delta(&dfa, state, c2s[lexer->c]);
lexer_next(lexer);
}
if (!(state < dfa.len && dfa.acc[state])) {
err = 1;
pr_str(p, "fail: line ");
pr_u32(p, lineno);
pr_str(p, ": expected ACC got REJ\n");
}
++lineno;
lexer_next(lexer);
}
lexer_next(lexer);
}
while (lexer->c != '\n') {
if (lexer->c == READER_EOF) {
return PUTERR("error: unexpected EOF"), 1;
} else if (lexer->c == READER_ERROR) {
return PUTERRNO("read"), 1;
} else if (!lexer_is_space(lexer->c)) {
return PUTERR("error: expected whitespace or EOL"), 1;
}
lexer_next(lexer);
}
lexer_next(lexer);
++lineno;
{ /* strings expected to be rejected */
while (lexer->c != READER_EOF) {
u32 state = dfa.start_state;
while (lexer->c != '\n' && lexer->c != READER_EOF) {
if (lexer->c == READER_ERROR) {
return PUTERRNO("read"), 1;
}
if (lexer->c == '%') {
return PUTERR("error: % in pattern"), 1;
}
state = dfa_delta(&dfa, state, c2s[lexer->c]);
lexer_next(lexer);
}
if (state < dfa.len && dfa.acc[state]) {
err = 1;
pr_str(p, "fail: line ");
pr_u32(p, lineno);
pr_str(p, ": expected REJ got ACC\n");
}
if (lexer->c != READER_EOF) {
lexer_next(lexer);
++lineno;
}
}
}
return err;
}