mirrors/z3
3
0
Fork 0
mirror of https://github.com/Z3Prover/z3 synced 2025-04-23 09:05:31 +00:00
Code Activity

add range / loop handling for re. Fix regression reading mixed numerals reported by Trentin

Browse source 
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
This commit is contained in:
Nikolaj Bjorner 2016-01-13 00:49:31 -08:00
parent 9a6fe93e6c
commit 9909c056f0
8 changed files with 200 additions and 58 deletions
Download patch file Download diff file Expand all files Collapse all files

77
src/ast/rewriter/seq_rewriter.cpp
View file

@ -24,18 +24,35 @@ Notes:
#include"uint_set.h"
#include"automaton.h"
#include"well_sorted.h"
#include"var_subst.h"
expr_ref sym_expr::accept(expr* e) {
ast_manager& m = m_t.get_manager();
expr_ref result(m);
if (m_is_pred) {
var_subst subst(m);
subst(m_t, 1, &e, result);
}
else {
result = m.mk_eq(e, m_t);
}
return result;
}
std::ostream& sym_expr::display(std::ostream& out) const {
return out << m_t;
}
struct display_expr1 {
ast_manager& m;
display_expr1(ast_manager& m): m(m) {}
std::ostream& display(std::ostream& out, expr* e) const {
return out << mk_pp(e, m);
std::ostream& display(std::ostream& out, sym_expr* e) const {
return e->display(out);
}
};
re2automaton::re2automaton(ast_manager& m): m(m), u(m) {}
re2automaton::re2automaton(ast_manager& m): m(m), u(m), bv(m) {}
eautomaton* re2automaton::operator()(expr* e) {
eautomaton* r = re2aut(e);
@ -53,6 +70,7 @@ eautomaton* re2automaton::re2aut(expr* e) {
expr* e1, *e2;
scoped_ptr<eautomaton> a, b;
unsigned lo, hi;
zstring s1, s2;
if (u.re.is_to_re(e, e1)) {
return seq2aut(e1);
}
@ -76,12 +94,27 @@ eautomaton* re2automaton::re2aut(expr* e) {
a = eautomaton::mk_opt(*a);
return a.detach();
}
else if (u.re.is_range(e)) {
// TBD
else if (u.re.is_range(e, e1, e2)) {
if (u.str.is_string(e1, s1) && u.str.is_string(e2, s2) &&
s1.length() == 1 && s2.length() == 1) {
unsigned start = s1[0];
unsigned stop = s2[0];
unsigned nb = s1.num_bits();
sort_ref s(bv.mk_sort(nb), m);
expr_ref v(m.mk_var(0, s), m);
expr_ref _start(bv.mk_numeral(start, nb), m);
expr_ref _stop(bv.mk_numeral(stop, nb), m);
expr_ref cond(m.mk_and(bv.mk_ule(_start, v), bv.mk_ule(v, _stop)), m);
a = alloc(eautomaton, sm, sym_expr::mk_pred(cond));
return a.detach();
}
else {
TRACE("seq", tout << "Range expression is not handled: " << mk_pp(e, m) << "\n";);
}
}
else if (u.re.is_loop(e, e1, lo, hi) && (a = re2aut(e1))) {
scoped_ptr<eautomaton> eps = eautomaton::mk_epsilon(m);
b = eautomaton::mk_epsilon(m);
scoped_ptr<eautomaton> eps = eautomaton::mk_epsilon(sm);
b = eautomaton::mk_epsilon(sm);
while (hi > lo) {
scoped_ptr<eautomaton> c = eautomaton::mk_concat(*a, *b);
b = eautomaton::mk_union(*eps, *c);
@ -94,10 +127,12 @@ eautomaton* re2automaton::re2aut(expr* e) {
return b.detach();
}
#if 0
else if (u.re.is_intersect(e, e1, e2)) {
}
else if (u.re.is_empty(e)) {
return alloc(eautomaton, m);
}
else if (u.re.is_full(e)) {
}
else if (u.re.is_intersect(e, e1, e2)) {
}
#endif
@ -114,10 +149,10 @@ eautomaton* re2automaton::seq2aut(expr* e) {
return eautomaton::mk_concat(*a, *b);
}
else if (u.str.is_unit(e, e1)) {
return alloc(eautomaton, m, e1);
return alloc(eautomaton, sm, sym_expr::mk_char(m, e1));
}
else if (u.str.is_empty(e)) {
return eautomaton::mk_epsilon(m);
return eautomaton::mk_epsilon(sm);
}
else if (u.str.is_string(e, s)) {
unsigned init = 0;
@ -126,9 +161,9 @@ eautomaton* re2automaton::seq2aut(expr* e) {
final.push_back(s.length());
for (unsigned k = 0; k < s.length(); ++k) {
// reference count?
mvs.push_back(eautomaton::move(m, k, k+1, u.str.mk_char(s, k)));
mvs.push_back(eautomaton::move(sm, k, k+1, sym_expr::mk_char(m, u.str.mk_char(s, k))));
}
return alloc(eautomaton, m, init, final, mvs);
return alloc(eautomaton, sm, init, final, mvs);
}
return 0;
}
@ -679,11 +714,11 @@ bool seq_rewriter::is_sequence(eautomaton& aut, expr_ref_vector& seq) {
return false;
}
visited.insert(state);
expr* t = mvs[0].t();
if (!t) {
sym_expr* t = mvs[0].t();
if (!t || !t->is_char()) {
return false;
}
seq.push_back(m_util.str.mk_unit(t));
seq.push_back(m_util.str.mk_unit(t->get_char()));
state = mvs[0].dst();
mvs.reset();
aut.get_moves_from(state, mvs, true);
@ -727,7 +762,7 @@ bool seq_rewriter::is_sequence(expr* e, expr_ref_vector& seq) {
br_status seq_rewriter::mk_str_in_regexp(expr* a, expr* b, expr_ref& result) {
scoped_ptr<eautomaton> aut;
expr_ref_vector seq(m());
if (!(aut = re2automaton(m())(b))) {
if (!(aut = m_re2aut(b))) {
return BR_FAILED;
}
@ -769,8 +804,8 @@ br_status seq_rewriter::mk_str_in_regexp(expr* a, expr* b, expr_ref& result) {
aut->get_moves_from(state, mvs, false);
for (unsigned j = 0; j < mvs.size(); ++j) {
eautomaton::move const& mv = mvs[j];
if (m().is_value(mv.t()) && m().is_value(ch)) {
if (mv.t() == ch) {
if (mv.t()->is_char() && m().is_value(mv.t()->get_char()) && m().is_value(ch)) {
if (mv.t()->get_char() == ch) {
add_next(next, mv.dst(), acc);
}
else {
@ -778,7 +813,7 @@ br_status seq_rewriter::mk_str_in_regexp(expr* a, expr* b, expr_ref& result) {
}
}
else {
cond = m().mk_eq(mv.t(), ch);
cond = mv.t()->accept(ch);
if (!m().is_true(acc)) cond = m().mk_and(acc, cond);
add_next(next, mv.dst(), cond);
}

33
src/ast/rewriter/seq_rewriter.h
View file

@ -26,11 +26,37 @@ Notes:
#include"lbool.h"
#include"automaton.h"
class sym_expr {
bool m_is_pred;
expr_ref m_t;
unsigned m_ref;
sym_expr(bool is_pred, expr_ref& t) : m_is_pred(is_pred), m_t(t), m_ref(0) {}
public:
expr_ref accept(expr* e);
static sym_expr* mk_char(expr_ref& t) { return alloc(sym_expr, false, t); }
static sym_expr* mk_char(ast_manager& m, expr* t) { expr_ref tr(t, m); return alloc(sym_expr, false, tr); }
static sym_expr* mk_pred(expr_ref& t) { return alloc(sym_expr, true, t); }
void inc_ref() { ++m_ref; }
void dec_ref() { --m_ref; if (m_ref == 0) dealloc(this); }
std::ostream& display(std::ostream& out) const;
bool is_char() const { return !m_is_pred; }
bool is_pred() const { return m_is_pred; }
expr* get_char() const { SASSERT(is_char()); return m_t; }
typedef automaton<expr, ast_manager> eautomaton;
};
class sym_expr_manager {
public:
void inc_ref(sym_expr* s) { if (s) s->inc_ref(); }
void dec_ref(sym_expr* s) { if (s) s->dec_ref(); }
};
typedef automaton<sym_expr, sym_expr_manager> eautomaton;
class re2automaton {
ast_manager& m;
seq_util u;
sym_expr_manager sm;
seq_util u;
bv_util bv;
eautomaton* re2aut(expr* e);
eautomaton* seq2aut(expr* e);
public:
@ -44,6 +70,7 @@ class re2automaton {
class seq_rewriter {
seq_util m_util;
arith_util m_autil;
re2automaton m_re2aut;
expr_ref_vector m_es, m_lhs, m_rhs;
br_status mk_seq_concat(expr* a, expr* b, expr_ref& result);
@ -80,7 +107,7 @@ class seq_rewriter {
public:
seq_rewriter(ast_manager & m, params_ref const & p = params_ref()):
m_util(m), m_autil(m), m_es(m), m_lhs(m), m_rhs(m) {
m_util(m), m_autil(m), m_re2aut(m), m_es(m), m_lhs(m), m_rhs(m) {
}
ast_manager & m() const { return m_util.get_manager(); }
family_id get_fid() const { return m_util.get_family_id(); }

47
src/ast/seq_decl_plugin.cpp
View file

@ -22,12 +22,46 @@ Revision History:
#include "ast_pp.h"
#include <sstream>
static bool is_hex_digit(char ch, unsigned& d) {
if ('0' <= ch && ch <= '9') {
d = ch - '0';
return true;
}
if ('A' <= ch && ch <= 'F') {
d = 10 + ch - 'A';
return true;
}
return false;
}
static bool is_escape_char(char const *& s, unsigned& result) {
unsigned d1, d2;
if (*s == '\\' && *(s + 1) == 'x' &&
is_hex_digit(*(s + 2), d1) && is_hex_digit(*(s + 3), d2)) {
result = d1*16 + d2;
s += 4;
return true;
}
if (*s == '\\' && *(s + 1) == '\\') {
result = '\\';
s += 2;
return true;
}
return false;
}
zstring::zstring(encoding enc): m_encoding(enc) {}
zstring::zstring(char const* s, encoding enc): m_encoding(enc) {
while (*s) {
m_buffer.push_back(*s);
++s;
unsigned ch;
if (is_escape_char(s, ch)) {
m_buffer.push_back(ch);
}
else {
m_buffer.push_back(*s);
++s;
}
}
}
@ -80,9 +114,10 @@ zstring zstring::replace(zstring const& src, zstring const& dst) const {
return result;
}
static const char esc_table[32][3] =
{ "\\0", "^A", "^B", "^C", "^D", "^E", "^F", "\\a", "\\b", "\\t", "\\n", "\\v", "\\f", "\\r", "^N",
"^O", "^P", "^Q", "^R", "^S", "^T", "^U", "^V","^W","^X","^Y","^Z","\\e","^\\","^]","^^","^_"};
static const char esc_table[32][6] =
{ "\\x00", "\\x01", "\\x02", "\\x03", "\\x04", "\\x05", "\\x06", "\\x07", "\\x08", "\\x09", "\\n", "\\v", "\\f", "\\r", "\\x0E", "\\x0F",
"\\x10", "\\x11", "\\x12", "\\x13", "\\x14", "\\x15", "\\x16", "\\x17", "\\x18", "\\x19", "\\x1A", "\\x1B", "\\x1C", "\\x1D", "\\x1E", "\\x1F"
};
std::string zstring::encode() const {
SASSERT(m_encoding == ascii);
@ -91,7 +126,7 @@ std::string zstring::encode() const {
unsigned char ch = m_buffer[i];
if (0 <= ch && ch < 32) {
strm << esc_table[ch];
}
}
else if (ch == 127) {
strm << "^?";
}

9
src/ast/seq_decl_plugin.h
View file

@ -98,6 +98,7 @@ public:
zstring& operator=(zstring const& other);
zstring replace(zstring const& src, zstring const& dst) const;
unsigned num_bits() const { return (m_encoding==ascii)?8:16; }
encoding get_encoding() const { return m_encoding; }
std::string encode() const;
unsigned length() const { return m_buffer.size(); }
unsigned operator[](unsigned i) const { return m_buffer[i]; }
@ -298,17 +299,19 @@ public:
bool is_to_re(expr const* n) const { return is_app_of(n, m_fid, OP_SEQ_TO_RE); }
bool is_concat(expr const* n) const { return is_app_of(n, m_fid, OP_RE_CONCAT); }
bool is_union(expr const* n) const { return is_app_of(n, m_fid, OP_RE_UNION); }
bool is_inter(expr const* n) const { return is_app_of(n, m_fid, OP_RE_INTERSECT); }
bool is_intersection(expr const* n) const { return is_app_of(n, m_fid, OP_RE_INTERSECT); }
bool is_star(expr const* n) const { return is_app_of(n, m_fid, OP_RE_STAR); }
bool is_plus(expr const* n) const { return is_app_of(n, m_fid, OP_RE_PLUS); }
bool is_opt(expr const* n) const { return is_app_of(n, m_fid, OP_RE_OPTION); }
bool is_range(expr const* n) const { return is_app_of(n, m_fid, OP_RE_RANGE); }
bool is_loop(expr const* n) const { return is_app_of(n, m_fid, OP_RE_LOOP); }
bool is_empty(expr const* n) const { return is_app_of(n, m_fid, OP_RE_EMPTY_SET); }
bool is_full(expr const* n) const { return is_app_of(n, m_fid, OP_RE_FULL_SET); }
MATCH_UNARY(is_to_re);
MATCH_BINARY(is_concat);
MATCH_BINARY(is_union);
MATCH_BINARY(is_inter);
MATCH_BINARY(is_intersection);
MATCH_BINARY(is_range);
MATCH_UNARY(is_star);
MATCH_UNARY(is_plus);
MATCH_UNARY(is_opt);