djsundog
/
circuitpython
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Abstract out back-end stream functionality from lexer.

This commit is contained in:
Damien 2013-10-20 14:41:27 +01:00
parent 68f59a96e3
commit a5185f4bc8
9 changed files with 186 additions and 153 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
py/emitcpy.c
View File

@ -211,7 +211,6 @@ static void emit_cpy_load_const_verbatim_strn(emit_t *emit, const char *str, int
}
static void emit_cpy_load_const_verbatim_quoted_str(emit_t *emit, qstr qstr, bool bytes) {
// TODO strings should be escaped before we get here
if (emit->pass == PASS_3) {
const char *str = qstr_str(qstr);
int len = strlen(str);
@ -237,13 +236,8 @@ static void emit_cpy_load_const_verbatim_quoted_str(emit_t *emit, qstr qstr, boo
for (int i = 0; i < len; i++) {
if (str[i] == '\n') {
printf("\\n");
} else if (str[i] == '\\' && str[i + 1] == '\'') {
i += 1;
if (quote_single) {
printf("\\'");
} else {
printf("'");
}
} else if (str[i] == '\\') {
printf("\\\\");
} else if (str[i] == '\'' && quote_single) {
printf("\\'");
} else {

212
py/lexer.c
View File

@ -9,48 +9,43 @@
#include "lexer.h"
#define TAB_SIZE (8)
#define CHR_EOF (-1)
struct _py_lexer_t {
const char *name; // (file) name of source
bool free; // free source when done with it
const char *name; // name of source
void *stream_data; // data for stream
py_lexer_stream_next_char_t stream_next_char; // stream callback to get next char
py_lexer_stream_free_t stream_free; // stream callback to free
const char *src_beg; // beginning of source
const char *src_cur; // current location in source; points to chr0
const char *src_end; // end (exclusive) of source
unichar chr0, chr1, chr2; // current characters from source
unichar chr0, chr1, chr2; // current cached characters from source
uint line; // source line
uint column; // source column
uint cont_line; // continued line
int emit_dent;
int nested_bracket_level;
int emit_dent; // non-zero when there are INDENT/DEDENT tokens to emit
int nested_bracket_level; // >0 when there are nested brackets over multiple lines
uint alloc_indent_level;
uint num_indent_level;
uint16_t *indent_level;
vstr_t vstr;
py_token_t tok_cur;
py_token_t tok_next;
};
static bool py_token_is_str(const py_token_t *tok, const char *str) {
bool str_strn_equal(const char *str, const char *strn, int len) {
uint i = 0;
const char *tstr = tok->str;
while (i < tok->len && *tstr == *str) {
while (i < len && *str == *strn) {
++i;
++tstr;
++str;
++strn;
}
return i == tok->len && *str == 0;
return i == len && *str == 0;
}
void py_token_show(const py_token_t *tok) {
printf("(%s:%d:%d) kind:%d cont_line:%d str:%p len:%d", tok->src_name, tok->src_line, tok->src_column, tok->kind, tok->cont_line, tok->str, tok->len);
printf("(%s:%d:%d) kind:%d str:%p len:%d", tok->src_name, tok->src_line, tok->src_column, tok->kind, tok->str, tok->len);
if (tok->str != NULL && tok->len > 0) {
const char *i = tok->str;
const char *j = i + tok->len;
@ -77,8 +72,10 @@ bool py_token_show_error(const py_token_t *tok, const char *msg) {
return false;
}
#define CUR_CHAR(lex) ((lex)->chr0)
static bool is_end(py_lexer_t *lex) {
return lex->chr0 == CHR_EOF;
return lex->chr0 == PY_LEXER_CHAR_EOF;
}
static bool is_physical_newline(py_lexer_t *lex) {
@ -142,7 +139,7 @@ static bool is_tail_of_identifier(py_lexer_t *lex) {
}
static void next_char(py_lexer_t *lex) {
if (lex->chr0 == CHR_EOF) {
if (lex->chr0 == PY_LEXER_CHAR_EOF) {
return;
}
@ -152,12 +149,10 @@ static void next_char(py_lexer_t *lex) {
// LF is a new line
++lex->line;
lex->column = 1;
lex->cont_line = lex->line;
} else if (lex->chr0 == '\r') {
// CR is a new line
++lex->line;
lex->column = 1;
lex->cont_line = lex->line;
if (lex->chr1 == '\n') {
// CR LF is a single new line
advance = 2;
@ -173,15 +168,11 @@ static void next_char(py_lexer_t *lex) {
for (; advance > 0; advance--) {
lex->chr0 = lex->chr1;
lex->chr1 = lex->chr2;
lex->src_cur++;
if (lex->src_cur + 2 < lex->src_end) {
lex->chr2 = lex->src_cur[2];
} else {
lex->chr2 = lex->stream_next_char(lex->stream_data);
if (lex->chr2 == PY_LEXER_CHAR_EOF) {
// EOF
if (lex->chr1 != CHR_EOF && lex->chr1 != '\n' && lex->chr1 != '\r') {
if (lex->chr1 != PY_LEXER_CHAR_EOF && lex->chr1 != '\n' && lex->chr1 != '\r') {
lex->chr2 = '\n'; // insert newline at end of file
} else {
lex->chr2 = CHR_EOF;
}
}
}
@ -286,9 +277,9 @@ static const char *tok_kw[] = {
NULL,
};
static void py_lexer_next_token_into(py_lexer_t *lex, py_token_t *tok) {
static void py_lexer_next_token_into(py_lexer_t *lex, py_token_t *tok, bool first_token) {
// skip white space and comments
bool had_physical_newline = false;
while (!is_end(lex)) {
if (is_physical_newline(lex)) {
had_physical_newline = true;
@ -315,15 +306,22 @@ static void py_lexer_next_token_into(py_lexer_t *lex, py_token_t *tok) {
}
}
// set token source information
tok->src_name = lex->name;
tok->src_line = lex->line;
tok->src_column = lex->column;
tok->kind = PY_TOKEN_INVALID;
tok->cont_line = lex->cont_line;
tok->str = lex->src_cur;
tok->len = 0;
if (lex->emit_dent < 0) {
// start new token text
vstr_reset(&lex->vstr);
if (first_token && lex->line == 1 && lex->column != 1) {
// check that the first token is in the first column
// if first token is not on first line, we get a physical newline and
// this check is done as part of normal indent/dedent checking below
// (done to get equivalence with CPython)
tok->kind = PY_TOKEN_INDENT;
} else if (lex->emit_dent < 0) {
tok->kind = PY_TOKEN_DEDENT;
lex->emit_dent += 1;
@ -414,19 +412,42 @@ static void py_lexer_next_token_into(py_lexer_t *lex, py_token_t *tok) {
num_quotes = 1;
}
// set start of token
tok->str = lex->src_cur;
// parse the literal
// TODO proper escaping
int n_closing = 0;
while (!is_end(lex) && (num_quotes > 1 || !is_char(lex, '\n')) && n_closing < num_quotes) {
if (is_char(lex, quote_char)) {
n_closing += 1;
vstr_add_char(&lex->vstr, CUR_CHAR(lex));
} else {
n_closing = 0;
if (!is_raw && is_char(lex, '\\')) {
next_char(lex);
unichar c = CUR_CHAR(lex);
switch (c) {
case PY_LEXER_CHAR_EOF: break; // TODO a proper error message?
case '\n': c = PY_LEXER_CHAR_EOF; break; // TODO check this works correctly (we are supposed to ignore it
case '\\': break;
case '\'': break;
case '"': break;
case 'a': c = 0x07; break;
case 'b': c = 0x08; break;
case 't': c = 0x09; break;
case 'n': c = 0x0a; break;
case 'v': c = 0x0b; break;
case 'f': c = 0x0c; break;
case 'r': c = 0x0d; break;
// TODO \ooo octal
case 'x': // TODO \xhh
case 'N': // TODO \N{name} only in strings
case 'u': // TODO \uxxxx only in strings
case 'U': // TODO \Uxxxxxxxx only in strings
default: break; // TODO error message
}
if (c != PY_LEXER_CHAR_EOF) {
vstr_add_char(&lex->vstr, c);
}
} else {
vstr_add_char(&lex->vstr, CUR_CHAR(lex));
}
}
next_char(lex);
@ -437,33 +458,40 @@ static void py_lexer_next_token_into(py_lexer_t *lex, py_token_t *tok) {
tok->kind = PY_TOKEN_LONELY_STRING_OPEN;
}
// set token string (byte) length
tok->len = lex->src_cur - tok->str - n_closing;
// we set the length, return now so it's not set incorrectly below
return;
// cut off the end quotes from the token text
vstr_cut_tail(&lex->vstr, n_closing);
} else if (is_head_of_identifier(lex)) {
tok->kind = PY_TOKEN_NAME;
// get first char
vstr_add_char(&lex->vstr, CUR_CHAR(lex));
next_char(lex);
// get tail chars
while (!is_end(lex) && is_tail_of_identifier(lex)) {
vstr_add_char(&lex->vstr, CUR_CHAR(lex));
next_char(lex);
}
} else if (is_digit(lex) || (is_char(lex, '.') && is_following_digit(lex))) {
tok->kind = PY_TOKEN_NUMBER;
// get first char
vstr_add_char(&lex->vstr, CUR_CHAR(lex));
next_char(lex);
// get tail chars
while (!is_end(lex)) {
if (is_char_or(lex, 'e', 'E')) {
vstr_add_char(&lex->vstr, 'e');
next_char(lex);
if (is_char(lex, '+') || is_char(lex, '-')) {
vstr_add_char(&lex->vstr, CUR_CHAR(lex));
next_char(lex);
}
} else if (is_letter(lex) || is_digit(lex) || is_char_or(lex, '_', '.')) {
vstr_add_char(&lex->vstr, CUR_CHAR(lex));
next_char(lex);
} else {
break;
@ -546,13 +574,14 @@ static void py_lexer_next_token_into(py_lexer_t *lex, py_token_t *tok) {
}
}
// set token string (byte) length
tok->len = lex->src_cur - tok->str;
// point token text to vstr buffer
tok->str = vstr_str(&lex->vstr);
tok->len = vstr_len(&lex->vstr);
// check for keywords (must be done after setting token string length)
// check for keywords
if (tok->kind == PY_TOKEN_NAME) {
for (int i = 0; tok_kw[i] != NULL; i++) {
if (py_token_is_str(tok, tok_kw[i])) {
if (str_strn_equal(tok_kw[i], tok->str, tok->len)) {
tok->kind = PY_TOKEN_KW_FALSE + i;
break;
}
@ -560,83 +589,58 @@ static void py_lexer_next_token_into(py_lexer_t *lex, py_token_t *tok) {
}
}
py_lexer_t *py_lexer_from_str_len(const char *src_name, const char *str, uint len, bool free_str) {
py_lexer_t *lex;
py_lexer_t *py_lexer_new(const char *src_name, void *stream_data, py_lexer_stream_next_char_t stream_next_char, py_lexer_stream_free_t stream_free) {
py_lexer_t *lex = m_new(py_lexer_t, 1);
lex = m_new(py_lexer_t, 1);
//lex->name = g_strdup(src_name); // TODO
lex->name = src_name;
lex->free = free_str;
lex->src_beg = str;
lex->src_cur = str;
lex->src_end = str + len;
lex->name = src_name; // TODO do we need to strdup this?
lex->stream_data = stream_data;
lex->stream_next_char = stream_next_char;
lex->stream_free = stream_free;
lex->line = 1;
lex->column = 1;
lex->cont_line = lex->line;
lex->emit_dent = 0;
lex->nested_bracket_level = 0;
lex->alloc_indent_level = 16;
lex->num_indent_level = 1;
lex->indent_level = m_new(uint16_t, lex->alloc_indent_level);
lex->indent_level[0] = 0;
vstr_init(&lex->vstr);
// preload characters
// TODO unicode
if (len == 0) {
lex->chr0 = '\n'; // insert newline at end of file
lex->chr1 = CHR_EOF;
lex->chr2 = CHR_EOF;
} else if (len == 1) {
lex->chr0 = str[0];
lex->chr0 = stream_next_char(stream_data);
lex->chr1 = stream_next_char(stream_data);
lex->chr2 = stream_next_char(stream_data);
// if input stream is 0, 1 or 2 characters long and doesn't end in a newline, then insert a newline at the end
if (lex->chr0 == PY_LEXER_CHAR_EOF) {
lex->chr0 = '\n';
} else if (lex->chr1 == PY_LEXER_CHAR_EOF) {
if (lex->chr0 != '\n' && lex->chr0 != '\r') {
lex->chr1 = '\n'; // insert newline at end of file
} else {
lex->chr1 = CHR_EOF;
lex->chr1 = '\n';
}
lex->chr2 = CHR_EOF;
} else if (len == 2) {
lex->chr0 = str[0];
lex->chr1 = str[1];
} else if (lex->chr2 == PY_LEXER_CHAR_EOF) {
if (lex->chr1 != '\n' && lex->chr1 != '\r') {
lex->chr2 = '\n'; // insert newline at end of file
} else {
lex->chr2 = CHR_EOF;
lex->chr2 = '\n';
}
} else {
lex->chr0 = str[0];
lex->chr1 = str[1];
lex->chr2 = str[2];
}
py_lexer_next_token_into(lex, &lex->tok_cur);
// check that the first token is in the first column
// (done to get equivalence with CPython)
if (lex->tok_cur.src_line == 1 && lex->tok_cur.src_column != 1) {
lex->tok_next = lex->tok_cur;
lex->tok_cur.kind = PY_TOKEN_INDENT;
} else {
py_lexer_next_token_into(lex, &lex->tok_next);
}
// preload first token
py_lexer_next_token_into(lex, &lex->tok_cur, true);
return lex;
}
void py_lexer_free(py_lexer_t *lex) {
if (lex == NULL) {
return;
if (lex) {
if (lex->stream_free) {
lex->stream_free(lex->stream_data);
}
m_free(lex);
}
//m_free(lex->name);
if (lex->free) {
m_free((char*)lex->src_beg);
}
m_free(lex);
}
void py_lexer_to_next(py_lexer_t *lex) {
lex->tok_cur = lex->tok_next;
py_lexer_next_token_into(lex, &lex->tok_next);
py_lexer_next_token_into(lex, &lex->tok_cur, false);
}
const py_token_t *py_lexer_cur(const py_lexer_t *lex) {
@ -652,14 +656,6 @@ bool py_lexer_is_str(py_lexer_t *lex, const char *str) {
return py_token_is_str(&lex->tok_cur, str);
}
bool py_lexer_is_next_kind(py_lexer_t *lex, py_token_kind_t kind) {
return lex->tok_next.kind == kind;
}
bool py_lexer_is_next_str(py_lexer_t *lex, const char *str) {
return py_token_is_str(&lex->tok_next, str);
}
bool py_lexer_opt_kind(py_lexer_t *lex, py_token_kind_t kind) {
if (py_lexer_is_kind(lex, kind)) {
py_lexer_to_next(lex);

21
py/lexer.h
View File

@ -108,32 +108,35 @@ typedef enum _py_token_kind_t {
} py_token_kind_t;
typedef struct _py_token_t {
const char *src_name; // (file) name of source
uint src_line; // actual source line
uint src_column; // actual source column
const char *src_name; // name of source
uint src_line; // source line
uint src_column; // source column
py_token_kind_t kind; // kind of token
uint cont_line; // token belongs to this line in a continued line
const char *str; // string of token
const char *str; // string of token (valid only while this token is current token)
uint len; // (byte) length of string of token
} py_token_t;
// the next-char function must return the next character in the stream
// it must return PY_LEXER_CHAR_EOF if end of stream
// it can be called again after returning PY_LEXER_CHAR_EOF, and in that case must return PY_LEXER_CHAR_EOF
#define PY_LEXER_CHAR_EOF (-1)
typedef unichar (*py_lexer_stream_next_char_t)(void*);
typedef void (*py_lexer_stream_free_t)(void*);
typedef struct _py_lexer_t py_lexer_t;
void py_token_show(const py_token_t *tok);
void py_token_show_error_prefix(const py_token_t *tok);
bool py_token_show_error(const py_token_t *tok, const char *msg);
py_lexer_t *py_lexer_from_file(const char *filename);
py_lexer_t *py_lexer_from_str_len(const char *src_name, const char *str, uint len, bool free_str);
py_lexer_t *py_lexer_new(const char *src_name, void *stream_data, py_lexer_stream_next_char_t stream_next_char, py_lexer_stream_free_t stream_free);
void py_lexer_free(py_lexer_t *lex);
void py_lexer_to_next(py_lexer_t *lex);
const py_token_t *py_lexer_cur(const py_lexer_t *lex);
bool py_lexer_is_kind(py_lexer_t *lex, py_token_kind_t kind);
/* unused
bool py_lexer_is_str(py_lexer_t *lex, const char *str);
bool py_lexer_is_next_kind(py_lexer_t *lex, py_token_kind_t kind);
bool py_lexer_is_next_str(py_lexer_t *lex, const char *str);
bool py_lexer_opt_kind(py_lexer_t *lex, py_token_kind_t kind);
bool py_lexer_opt_str(py_lexer_t *lex, const char *str);
*/

23
py/lexerfile.c
View File

@ -1,23 +0,0 @@
#include <stdint.h>
#include <stdio.h>
#include <unistd.h>
#include <fcntl.h>
#include "misc.h"
#include "lexer.h"
py_lexer_t *py_lexer_from_file(const char *filename) {
// TODO abstract away file functionality
int fd = open(filename, O_RDONLY);
if (fd < 0) {
printf("cannot open file %s\n", filename);
return NULL;
}
uint size = lseek(fd, 0, SEEK_END);
lseek(fd, 0, SEEK_SET);
char *data = m_new(char, size);
read(fd, data, size);
close(fd);
return py_lexer_from_str_len(filename, data, size, true);
}

3
unix/Makefile
View File

@ -7,14 +7,15 @@ LDFLAGS =
SRC_C = \
main.c \
lexerunix.c \
PY_O = \
nlrx64.o \
malloc.o \
qstr.o \
vstr.o \
misc.o \
lexer.o \
lexerfile.o \
parse.o \
scope.o \
compile.o \

55
unix/lexerunix.c Normal file
View File

@ -0,0 +1,55 @@
#include <stdint.h>
#include <stdio.h>
#include <unistd.h>
#include <fcntl.h>
#include "misc.h"
#include "lexer.h"
typedef struct _str_buf_t {
bool free; // free src_beg when done
const char *src_beg; // beginning of source
const char *src_cur; // current location in source
const char *src_end; // end (exclusive) of source
} str_buf_t;
unichar str_buf_next_char(str_buf_t *sb) {
if (sb->src_cur < sb->src_end) {
return *sb->src_cur++;
} else {
return PY_LEXER_CHAR_EOF;
}
}
void str_buf_free(str_buf_t *sb) {
if (sb) {
if (sb->free) {
m_free((char*)sb->src_beg);
}
m_free(sb);
}
}
py_lexer_t *py_lexer_new_from_str_len(const char *src_name, const char *str, uint len, bool free_str) {
str_buf_t *sb = m_new(str_buf_t, 1);
sb->free = free_str;
sb->src_beg = str;
sb->src_cur = str;
sb->src_end = str + len;
return py_lexer_new(src_name, sb, (py_lexer_stream_next_char_t)str_buf_next_char, (py_lexer_stream_free_t)str_buf_free);
}
py_lexer_t *py_lexer_new_from_file(const char *filename) {
int fd = open(filename, O_RDONLY);
if (fd < 0) {
printf("cannot open file %s\n", filename);
return NULL;
}
uint size = lseek(fd, 0, SEEK_END);
lseek(fd, 0, SEEK_SET);
char *data = m_new(char, size);
read(fd, data, size);
close(fd);
return py_lexer_new_from_str_len(filename, data, size, true);
}

2
unix/lexerunix.h Normal file
View File

@ -0,0 +1,2 @@
py_lexer_t *py_lexer_new_from_str_len(const char *src_name, const char *str, uint len, bool free_str);
py_lexer_t *py_lexer_new_from_file(const char *filename);

11
unix/main.c
View File

@ -6,6 +6,7 @@
#include "misc.h"
#include "mpyconfig.h"
#include "lexer.h"
#include "lexerunix.h"
#include "parse.h"
#include "compile.h"
#include "runtime.h"
@ -67,7 +68,7 @@ void do_repl() {
line = line3;
}
}
py_lexer_t *lex = py_lexer_from_str_len("<stdin>", line, strlen(line), false);
py_lexer_t *lex = py_lexer_new_from_str_len("<stdin>", line, strlen(line), false);
py_parse_node_t pn = py_parse(lex, PY_PARSE_SINGLE_INPUT);
if (pn != PY_PARSE_NODE_NULL) {
//py_parse_node_show(pn, 0);
@ -91,7 +92,7 @@ void do_repl() {
}
void do_file(const char *file) {
py_lexer_t *lex = py_lexer_from_file(file);
py_lexer_t *lex = py_lexer_new_from_file(file);
//const char *pysrc = "def f():\n x=x+1\n print(42)\n";
//py_lexer_t *lex = py_lexer_from_str_len("<>", pysrc, strlen(pysrc), false);
if (lex == NULL) {
@ -119,7 +120,11 @@ void do_file(const char *file) {
py_lexer_free(lex);
#if !MICROPY_EMIT_CPYTHON
#if MICROPY_EMIT_CPYTHON
if (!comp_ok) {
printf("compile error\n");
}
#else
if (1 && comp_ok) {
// execute it
py_obj_t module_fun = rt_make_function_from_id(1);

2
unix/mpyconfig.h
View File

@ -1,7 +1,7 @@
// options to control how Micro Python is built
#define MICROPY_ENABLE_FLOAT (1)
#define MICROPY_EMIT_CPYTHON (0)
#define MICROPY_EMIT_CPYTHON (1)
#define MICROPY_EMIT_X64 (1)
#define MICROPY_EMIT_THUMB (0)
#define MICROPY_EMIT_INLINE_THUMB (0)