Browse Source

rename codegen_state to parser_state and move it

later commits will rewrite the parser and separate the codegen
bits into their own API
master
Daniel Kolesa 4 weeks ago
parent
commit
ccb0c09d59
12 changed files with 1531 additions and 1548 deletions
  1. +2
    -2
      src/cs_error.cc
  2. +1
    -1388
      src/cs_gen.cc
  3. +0
    -142
      src/cs_gen.hh
  4. +0
    -1
      src/cs_ident.cc
  5. +1381
    -2
      src/cs_parser.cc
  6. +140
    -1
      src/cs_parser.hh
  7. +1
    -2
      src/cs_state.cc
  8. +0
    -2
      src/cs_thread.cc
  9. +2
    -2
      src/cs_thread.hh
  10. +1
    -2
      src/cs_val.cc
  11. +3
    -3
      src/cs_vm.cc
  12. +0
    -1
      src/cs_vm.hh

+ 2
- 2
src/cs_error.cc View File

@ -2,7 +2,7 @@
#include <algorithm>
#include "cs_gen.hh"
#include "cs_parser.hh"
#include "cs_thread.hh"
namespace cubescript {
@ -49,7 +49,7 @@ LIBCUBESCRIPT_EXPORT char *error::request_buf(
) {
auto &ts = state_p{cs}.ts();
charbuf &cb = ts.errbuf;
codegen_state *gs = ts.cstate;
parser_state *gs = ts.cstate;
cb.clear();
std::size_t sz = 0;
if (gs) {


+ 1
- 1388
src/cs_gen.cc
File diff suppressed because it is too large
View File


+ 0
- 142
src/cs_gen.hh View File

@ -3,150 +3,8 @@
#include "cubescript/cubescript.hh"
#include <cstdlib>
#include <array>
#include <vector>
#include <type_traits>
#include "cs_std.hh"
#include "cs_bcode.hh"
#include "cs_ident.hh"
#include "cs_thread.hh"
namespace cubescript {
struct codegen_state {
thread_state &ts;
codegen_state *prevps;
bool parsing = true;
valbuf<uint32_t> code;
char const *source, *send;
std::size_t current_line;
std::string_view src_name;
codegen_state() = delete;
codegen_state(thread_state &tsr):
ts{tsr}, prevps{tsr.cstate}, code{tsr.istate},
source{}, send{}, current_line{1}, src_name{}
{
tsr.cstate = this;
}
~codegen_state() {
done();
}
void done() {
if (!parsing) {
return;
}
ts.cstate = prevps;
parsing = false;
}
std::string_view get_str();
charbuf get_str_dup();
std::string_view get_word();
void gen_str(std::string_view word = std::string_view{}) {
if (word.size() <= 3) {
std::uint32_t op = BC_INST_VAL_INT | BC_RET_STRING;
for (size_t i = 0; i < word.size(); ++i) {
op |= std::uint32_t(
static_cast<unsigned char>(word[i])
) << ((i + 1) * 8);
}
code.push_back(op);
return;
}
code.push_back(
BC_INST_VAL | BC_RET_STRING | std::uint32_t(word.size() << 8)
);
auto it = reinterpret_cast<std::uint32_t const *>(word.data());
code.append(it, it + (word.size() / sizeof(std::uint32_t)));
std::size_t esz = word.size() % sizeof(std::uint32_t);
char c[sizeof(std::uint32_t)] = {0};
std::memcpy(c, word.data() + word.size() - esz, esz);
std::uint32_t u;
std::memcpy(&u, c, sizeof(u));
code.push_back(u);
}
void gen_null() {
code.push_back(BC_INST_VAL_INT | BC_RET_NULL);
}
void gen_int(integer_type i = 0) {
if (i >= -0x800000 && i <= 0x7FFFFF) {
code.push_back(BC_INST_VAL_INT | BC_RET_INT | (i << 8));
} else {
std::uint32_t u[bc_store_size<integer_type>] = {0};
std::memcpy(u, &i, sizeof(i));
code.push_back(BC_INST_VAL | BC_RET_INT);
code.append(u, u + bc_store_size<integer_type>);
}
}
void gen_int(std::string_view word);
void gen_float(float_type f = 0.0f) {
if (integer_type(f) == f && f >= -0x800000 && f <= 0x7FFFFF) {
code.push_back(BC_INST_VAL_INT | BC_RET_FLOAT | (integer_type(f) << 8));
} else {
std::uint32_t u[bc_store_size<float_type>] = {0};
std::memcpy(u, &f, sizeof(f));
code.push_back(BC_INST_VAL | BC_RET_FLOAT);
code.append(u, u + bc_store_size<float_type>);
}
}
void gen_float(std::string_view word);
void gen_ident(ident &id) {
code.push_back(BC_INST_IDENT | (id.get_index() << 8));
}
void gen_ident() {
gen_ident(*ts.istate->id_dummy);
}
void gen_ident(std::string_view word) {
gen_ident(ts.istate->new_ident(*ts.pstate, word, IDENT_FLAG_UNKNOWN));
}
void gen_value(
int wordtype, std::string_view word = std::string_view(),
int line = 0
);
void gen_main(std::string_view s, int ret_type = VAL_ANY);
void next_char() {
if (source == send) {
return;
}
if (*source == '\n') {
++current_line;
}
++source;
}
char current(size_t ahead = 0) {
if (std::size_t(send - source) <= ahead) {
return '\0';
}
return source[ahead];
}
std::string_view read_macro_name();
char skip_until(std::string_view chars);
char skip_until(char cf);
void skip_comments();
};
} /* namespace cubescript */
#endif /* LIBCUBESCRIPT_GEN_HH */

+ 0
- 1
src/cs_ident.cc View File

@ -1,7 +1,6 @@
#include "cs_ident.hh"
#include "cs_bcode.hh"
#include "cs_gen.hh"
#include "cs_thread.hh"
namespace cubescript {


+ 1381
- 2
src/cs_parser.cc
File diff suppressed because it is too large
View File


+ 140
- 1
src/cs_parser.hh View File

@ -1,9 +1,16 @@
#ifndef LIBCUBESCRIPT_PARSER_HH
#define LIBCUBESCRIPT_PARSER_HH
#include <cstdlib>
#include <string_view>
#include <type_traits>
#include <cubescript/cubescript.hh>
#include <string_view>
#include "cs_std.hh"
#include "cs_bcode.hh"
#include "cs_ident.hh"
#include "cs_thread.hh"
namespace cubescript {
@ -12,6 +19,138 @@ float_type parse_float(std::string_view input, std::string_view *end = nullptr);
bool is_valid_name(std::string_view input);
struct parser_state {
thread_state &ts;
parser_state *prevps;
bool parsing = true;
valbuf<uint32_t> code;
char const *source, *send;
std::size_t current_line;
std::string_view src_name;
parser_state() = delete;
parser_state(thread_state &tsr):
ts{tsr}, prevps{tsr.cstate}, code{tsr.istate},
source{}, send{}, current_line{1}, src_name{}
{
tsr.cstate = this;
}
~parser_state() {
done();
}
void done() {
if (!parsing) {
return;
}
ts.cstate = prevps;
parsing = false;
}
std::string_view get_str();
charbuf get_str_dup();
std::string_view get_word();
void gen_str(std::string_view word = std::string_view{}) {
if (word.size() <= 3) {
std::uint32_t op = BC_INST_VAL_INT | BC_RET_STRING;
for (size_t i = 0; i < word.size(); ++i) {
op |= std::uint32_t(
static_cast<unsigned char>(word[i])
) << ((i + 1) * 8);
}
code.push_back(op);
return;
}
code.push_back(
BC_INST_VAL | BC_RET_STRING | std::uint32_t(word.size() << 8)
);
auto it = reinterpret_cast<std::uint32_t const *>(word.data());
code.append(it, it + (word.size() / sizeof(std::uint32_t)));
std::size_t esz = word.size() % sizeof(std::uint32_t);
char c[sizeof(std::uint32_t)] = {0};
std::memcpy(c, word.data() + word.size() - esz, esz);
std::uint32_t u;
std::memcpy(&u, c, sizeof(u));
code.push_back(u);
}
void gen_null() {
code.push_back(BC_INST_VAL_INT | BC_RET_NULL);
}
void gen_int(integer_type i = 0) {
if (i >= -0x800000 && i <= 0x7FFFFF) {
code.push_back(BC_INST_VAL_INT | BC_RET_INT | (i << 8));
} else {
std::uint32_t u[bc_store_size<integer_type>] = {0};
std::memcpy(u, &i, sizeof(i));
code.push_back(BC_INST_VAL | BC_RET_INT);
code.append(u, u + bc_store_size<integer_type>);
}
}
void gen_int(std::string_view word);
void gen_float(float_type f = 0.0f) {
if (integer_type(f) == f && f >= -0x800000 && f <= 0x7FFFFF) {
code.push_back(BC_INST_VAL_INT | BC_RET_FLOAT | (integer_type(f) << 8));
} else {
std::uint32_t u[bc_store_size<float_type>] = {0};
std::memcpy(u, &f, sizeof(f));
code.push_back(BC_INST_VAL | BC_RET_FLOAT);
code.append(u, u + bc_store_size<float_type>);
}
}
void gen_float(std::string_view word);
void gen_ident(ident &id) {
code.push_back(BC_INST_IDENT | (id.get_index() << 8));
}
void gen_ident() {
gen_ident(*ts.istate->id_dummy);
}
void gen_ident(std::string_view word) {
gen_ident(ts.istate->new_ident(*ts.pstate, word, IDENT_FLAG_UNKNOWN));
}
void gen_value(
int wordtype, std::string_view word = std::string_view(),
int line = 0
);
void gen_main(std::string_view s, int ret_type = VAL_ANY);
void next_char() {
if (source == send) {
return;
}
if (*source == '\n') {
++current_line;
}
++source;
}
char current(size_t ahead = 0) {
if (std::size_t(send - source) <= ahead) {
return '\0';
}
return source[ahead];
}
std::string_view read_macro_name();
char skip_until(std::string_view chars);
char skip_until(char cf);
void skip_comments();
};
} /* namespace cubescript */
#endif

+ 1
- 2
src/cs_state.cc View File

@ -6,7 +6,6 @@
#include "cs_state.hh"
#include "cs_thread.hh"
#include "cs_strman.hh"
#include "cs_gen.hh"
#include "cs_vm.hh" // break/continue, call_with_args
#include "cs_parser.hh"
@ -656,7 +655,7 @@ static any_value do_run(
thread_state &ts, std::string_view file, std::string_view code
) {
any_value ret{*ts.pstate};
codegen_state gs{ts};
parser_state gs{ts};
gs.src_name = file;
gs.code.reserve(64);
gs.gen_main(code, VAL_ANY);


+ 0
- 2
src/cs_thread.cc View File

@ -1,7 +1,5 @@
#include "cs_thread.hh"
#include "cs_gen.hh"
namespace cubescript {
thread_state::thread_state(internal_state *cs):


+ 2
- 2
src/cs_thread.hh View File

@ -11,7 +11,7 @@
namespace cubescript {
struct codegen_state;
struct parser_state;
struct thread_state {
using astack_allocator = std_allocator<std::pair<int const, alias_stack>>;
@ -22,7 +22,7 @@ struct thread_state {
/* the public state interface */
state *pstate{};
/* current codegen state for diagnostics */
codegen_state *cstate{};
parser_state *cstate{};
/* VM stack */
valbuf<any_value> vmstack;
/* ident stack */


+ 1
- 2
src/cs_val.cc View File

@ -1,5 +1,4 @@
#include <cubescript/cubescript.hh>
#include "cs_gen.hh"
#include "cs_std.hh"
#include "cs_parser.hh"
#include "cs_state.hh"
@ -270,7 +269,7 @@ bcode_ref any_value::force_code(state &cs) {
default:
break;
}
codegen_state gs{state_p{cs}.ts()};
parser_state gs{state_p{cs}.ts()};
gs.code.reserve(64);
gs.gen_main(get_string());
gs.done();


+ 3
- 3
src/cs_vm.cc View File

@ -239,7 +239,7 @@ bool exec_alias(
ident_link aliaslink = {a, ts.callstack, uargs};
ts.callstack = &aliaslink;
if (!aast.node->code) {
codegen_state gs{ts};
parser_state gs{ts};
gs.code.reserve(64);
gs.gen_main(aast.node->val_s.get_string());
/* i wish i could steal the memory somehow */
@ -687,7 +687,7 @@ std::uint32_t *vm_exec(
case BC_INST_COMPILE: {
any_value &arg = args.back();
codegen_state gs{ts};
parser_state gs{ts};
switch (arg.get_type()) {
case value_type::INTEGER:
gs.code.reserve(8);
@ -733,7 +733,7 @@ std::uint32_t *vm_exec(
case value_type::STRING: {
std::string_view s = arg.get_string();
if (!s.empty()) {
codegen_state gs{ts};
parser_state gs{ts};
gs.code.reserve(64);
gs.gen_main(s);
gs.done();


+ 0
- 1
src/cs_vm.hh View File

@ -5,7 +5,6 @@
#include "cs_std.hh"
#include "cs_ident.hh"
#include "cs_gen.hh"
#include "cs_thread.hh"
namespace cubescript {


Loading…
Cancel
Save