neil
/
interpret
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

More strict: separated we've just seen a newline from general text.

This commit is contained in:
Neil 2022-02-15 13:05:15 -08:00
parent 4fdb66076a
commit fb814b7223
3 changed files with 73 additions and 81 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/interpret.c
View File

@ -15,9 +15,9 @@ static void int_to_string(const int *const n,
#define ARRAY_NAME int #define ARRAY_NAME int
#define ARRAY_TYPE int #define ARRAY_TYPE int
#define ARRAY_EXPECT_TRAIT #define ARRAY_EXPECT_TRAIT
#include "../src/array.h" #include "array.h"
#define ARRAY_TO_STRING &int_to_string #define ARRAY_TO_STRING &int_to_string
#include "../src/array.h" #include "array.h"
static int int_cmp(const int *const a, const int *const b) static int int_cmp(const int *const a, const int *const b)
{ return (*b < *a) - (*a < *b); } { return (*b < *a) - (*a < *b); }
static int void_int_cmp(const void *const a, const void *const b) static int void_int_cmp(const void *const a, const void *const b)

9
src/lex.h
View File

@ -4,10 +4,11 @@ int lex_looks_like_year(const char *, int *);
int lex_looks_like_month(const char *); int lex_looks_like_month(const char *);
int lex_looks_like_day(const char *); int lex_looks_like_day(const char *);
#define LEX_SYMBOL X(END), X(ERROR), X(TEXT), X(BANG), X(BRACKET), X(WHITE), \ #define LEX_SYMBOL \
X(MAP), \ /* Results. */ X(END), X(SYNTAX), X(ILLEGAL), X(NOT_FOUND), \
/* Commands */ X(SOURCE), X(DEFAULT), X(SIGNIFICANT), X(SCORE), \ /* Text */ X(PARAGRAPH), X(TEXT), \
/* Arguments */ X(ARG_KEYWORD), X(ARG_DATE), X(ARG_NATURAL), X(ARG_FREEFORM) /* Directive. */ X(SOURCE), X(DEFAULT), X(SIGNIFICANT), X(SCORE), X(MAP), \
/* Arguments. */ X(ARG_KEYWORD), X(ARG_DATE), X(ARG_NATURAL), X(ARG_FREEFORM)
#define X(n) n #define X(n) n
struct lex { struct lex {

141
src/lex.re_c.c
View File

@ -20,6 +20,7 @@ re2c:define:YYCTYPE = char;
int lex_looks_like_year(const char *const a, int *const year) { int lex_looks_like_year(const char *const a, int *const year) {
const char *YYCURSOR = a, *YYMARKER = a, *s0; const char *YYCURSOR = a, *YYMARKER = a, *s0;
/*!stags:re2c format = 'const char *@@;\n'; */ /*!stags:re2c format = 'const char *@@;\n'; */
(void)yyt1;
assert(a && year); assert(a && year);
/*!re2c /*!re2c
@s0 ("-"? [1-9][0-9]* | "0") "\x00" { @s0 ("-"? [1-9][0-9]* | "0") "\x00" {
@ -81,7 +82,7 @@ static struct scan {
#define X(n) EXPECT_ ## n #define X(n) EXPECT_ ## n
/* "[something: expect; there; to; be; args]", in this case, expect would /* "[something: expect; there; to; be; args]", in this case, expect would
be a stack of `size = 5` `EXPECT_KEYWORD`. */ be a stack of `size = 5` `EXPECT_KEYWORD`. */
struct { unsigned size; enum { EXPECT } expect[16]; } command; struct { unsigned size; enum { EXPECT } expect[16]; } edict;
#undef X #undef X
} scan; /* Terrible, gah. */ } scan; /* Terrible, gah. */
@ -92,45 +93,30 @@ void lex_reset(const char *const buffer) {
scan.line = 1; scan.line = 1;
} }
/*!re2c
re2c:flags:tags = 1;
re2c:define:YYCURSOR = scan.cursor;
re2c:define:YYMARKER = scan.marker;
re2c:define:YYCTXMARKER = scan.ctx_marker;
re2c:define:YYCONDTYPE = 'condition';
re2c:define:YYGETCONDITION = 'scan.condition';
re2c:define:YYGETCONDITION:naked = 1;
re2c:define:YYSETCONDITION = 'scan.condition = @@;';
re2c:define:YYSETCONDITION:naked = 1;
// Eof is marked by null when preparing files for lexing.
// Mutually exclusive; only !, [, are not covered.
end = "\x00";
newline = "\n" | "\r" "\n"?;
ws = [ \t\v\f];
glyph = [^ \t\n\r\v\f![\x00];
glyphs = glyph+;
// inside the block
natural = [1-9][0-9]*;
decimal = "-"? ([1-9][0-9]* | [0])? "." [0-9]+ | [1-9][0-9]* | [0];
id = [a-zA-Z_][a-zA-Z_\-0-9]{0,63};
date = "-"? natural "-" [0-1][0-9] "-" [0-1][0-9];
*/
/** I don't think `re2c` supports branching on variable conditions. /** I don't think `re2c` supports branching on variable conditions.
It does now? */ It does now? */
static void expect_pop(void) { static void expect_pop(void) {
printf("<expect_pop>"); printf("<expect_pop>");
if(!scan.command.size) { printf("allfinished\n"); scan.condition = yyccommand_end; return; } if(!scan.edict.size) { printf("allfinished\n"); scan.condition = yycedict_end; return; }
switch(scan.command.expect[--scan.command.size]) { switch(scan.edict.expect[--scan.edict.size]) {
case EXPECT_KEYWORD: printf("keyword\n");scan.condition = yyccommand_keyword; break; case EXPECT_KEYWORD: printf("keyword\n");scan.condition = yycedict_keyword; break;
case EXPECT_DATE: printf("date\n");scan.condition = yyccommand_date; break; case EXPECT_DATE: printf("date\n");scan.condition = yycedict_date; break;
case EXPECT_FREEFORM: printf("freeform\n");scan.condition = yyccommand_freeform; break; case EXPECT_FREEFORM: printf("freeform\n");scan.condition = yycedict_freeform; break;
} }
} }
int lex_next(struct lex *const x) { int lex_next(struct lex *const x) {
/*!re2c
re2c:flags:tags = 1;
re2c:define:YYCURSOR = scan.cursor;
re2c:define:YYMARKER = scan.marker;
re2c:define:YYCTXMARKER = scan.ctx_marker;
re2c:define:YYCONDTYPE = 'condition';
re2c:define:YYGETCONDITION = 'scan.condition';
re2c:define:YYGETCONDITION:naked = 1;
re2c:define:YYSETCONDITION = 'scan.condition = @@;';
re2c:define:YYSETCONDITION:naked = 1;
*/
const char *s0, *s1; const char *s0, *s1;
const size_t prev_line = scan.line; const size_t prev_line = scan.line;
/*!stags:re2c format = 'const char *@@;\n'; */ /*!stags:re2c format = 'const char *@@;\n'; */
@ -142,24 +128,27 @@ int lex_next(struct lex *const x) {
x->new_paragraph = 0; x->new_paragraph = 0;
scan: scan:
/*!re2c /*!re2c
<text> end { printf("end\n");return x->symbol = END, 0; } end = "\x00";
<image, command, command_keyword, command_date, command_end> end illegal = [\x01-\x08\x0a-\x1f\x7f]; // unix-style control characters
{ return x->symbol = ERROR, 0; } newline = "\n";
ws = [ \t];
glyph = [^\x00-\x1f \x7f];
<*> illegal { return x->symbol = ILLEGAL, 0; }
<line> end { return x->symbol = END, 0; }
<text, text, image, edict, edict_keyword, edict_date, edict_freeform, edict_end>
end { return x->symbol = ILLEGAL, 0; }
<line> newline { x->line = ++scan.line; return x->symbol = PARAGRAPH, 1; }
<line> "![" :=> image
<line> "[" :=> edict
<line> "" / glyph :=> text
<line> * { return x->symbol = SYNTAX, 1; }
<text> newline <text> newline => line { x->line = ++scan.line; goto scan; }
{ x->line = ++scan.line; <text> ws+ { goto scan; }
x->new_paragraph = prev_line + 2 >= scan.line;
x->ws_before = 1;
goto scan; }
<text> ws+ { x->ws_before = 1; goto scan; }
<text> @s0 glyph+ @s1 { x->s0 = s0, x->s1 = s1; <text> @s0 glyph+ @s1 { x->s0 = s0, x->s1 = s1;
return x->symbol = TEXT, 1; } return x->symbol = TEXT, 1; }
<text> @s0 "!" @s1 { x->s0 = s0, x->s1 = s1;
return x->symbol = BANG, 1; } decimal = "-"? ([1-9][0-9]* | [0])? "." [0-9]+ | [1-9][0-9]* | [0];
<text> "\\" @s0 "[" @s1 { x->s0 = s0, x->s1 = s1;
return x->symbol = BRACKET, 1; }
<text> "![" :=> image
<text> "[" :=> command
<image> ws* "osm" ws* "](geo:" @s0 decimal "," @s1 decimal ")" => text { <image> ws* "osm" ws* "](geo:" @s0 decimal "," @s1 decimal ")" => text {
x->symbol = MAP, x->s0 = s0, x->s1 = s1; x->symbol = MAP, x->s0 = s0, x->s1 = s1;
@ -168,58 +157,60 @@ scan:
} }
<image> * { printf("image(broken)\n");return 0; } <image> * { printf("image(broken)\n");return 0; }
natural = [1-9][0-9]*;
id = [a-zA-Z_][a-zA-Z_\-0-9]{0,63};
date = "-"? natural "-" [0-1][0-9] "-" [0-1][0-9];
// source // source
<command> "source" <edict> "source"
{ if(scan.is_ws_expected || scan.command.size) { if(scan.is_ws_expected || scan.edict.size)
return x->symbol = ERROR, 0; return x->symbol = SYNTAX, 0;
scan.is_ws_expected = 1, scan.is_source = 1; scan.is_ws_expected = 1, scan.is_source = 1;
scan.command.size = 2; scan.edict.size = 2;
scan.command.expect[1] = EXPECT_KEYWORD; scan.edict.expect[1] = EXPECT_KEYWORD;
scan.command.expect[0] = EXPECT_FREEFORM; scan.edict.expect[0] = EXPECT_FREEFORM;
return x->symbol = SOURCE, 1; } return x->symbol = SOURCE, 1; }
<command> "default" <edict> "default"
{ if(scan.is_ws_expected || !scan.is_source) { if(scan.is_ws_expected || !scan.is_source)
return x->symbol = ERROR, 0; return x->symbol = SYNTAX, 0;
scan.is_ws_expected = 1, scan.is_source = 0; scan.is_ws_expected = 1, scan.is_source = 0;
return x->symbol = DEFAULT, 1; } return x->symbol = DEFAULT, 1; }
// score // score
<command> "significant" <edict> "significant"
{ if(scan.is_ws_expected || scan.command.size) { if(scan.is_ws_expected || scan.edict.size)
return x->symbol = ERROR, 0; return x->symbol = SYNTAX, 0;
scan.is_ws_expected = 1; scan.is_ws_expected = 1;
scan.command.size = 2; scan.edict.size = 2;
scan.command.expect[1] = EXPECT_FREEFORM; scan.edict.expect[1] = EXPECT_FREEFORM;
scan.command.expect[0] = EXPECT_DATE; scan.edict.expect[0] = EXPECT_DATE;
return x->symbol = SIGNIFICANT, 1; } return x->symbol = SIGNIFICANT, 1; }
<command> @s0 natural @s1 <edict> @s0 natural @s1
{ if(scan.is_ws_expected || scan.command.size) { if(scan.is_ws_expected || scan.edict.size)
return x->symbol = ERROR, 0; return x->symbol = SYNTAX, 0;
scan.is_ws_expected = 1; scan.is_ws_expected = 1;
x->s0 = s0, x->s1 = s1; x->s0 = s0, x->s1 = s1;
return x->symbol = SCORE, 1; } return x->symbol = SCORE, 1; }
// general command stuff <edict> ws+ { scan.is_ws_expected = 0; goto scan; }
<command> ws+ { scan.is_ws_expected = 0; goto scan; } <edict> ":"
<command> ":" { if(!scan.edict.size) return x->symbol = SYNTAX, 0;
{ if(!scan.command.size) return x->symbol = ERROR, 0;
scan.is_ws_expected = 0, scan.is_source = 0; scan.is_ws_expected = 0, scan.is_source = 0;
expect_pop(); goto scan; } expect_pop(); goto scan; }
<command_keyword> ws* @s0 id @s1 ws* ";"? / "]"? <edict_keyword> ws* @s0 id @s1 ws* ";"? / "]"?
{ x->s0 = s0, x->s1 = s1; expect_pop(); { x->s0 = s0, x->s1 = s1; expect_pop();
return x->symbol = ARG_KEYWORD, 1; } return x->symbol = ARG_KEYWORD, 1; }
<command_date> ws* @s0 date @s1 ws* ";"? / "]"? <edict_date> ws* @s0 date @s1 ws* ";"? / "]"?
{ x->s0 = s0, x->s1 = s1; expect_pop(); { x->s0 = s0, x->s1 = s1; expect_pop();
return x->symbol = ARG_DATE, 1; } return x->symbol = ARG_DATE, 1; }
<command_freeform> ws* @s0 <edict_freeform> ws* @s0
[^ \t\n\r\v\f;[\]\x00][^\t\n\r\v\f;[\]\x00]*[^ \t\n\r\v\f;[\]\x00]* [^ \t\n\r\v\f;[\]\x00][^\t\n\r\v\f;[\]\x00]*[^ \t\n\r\v\f;[\]\x00]*
@s1 ws* ";"? / "]"? @s1 ws* ";"? / "]"?
{ x->s0 = s0, x->s1 = s1; expect_pop(); { x->s0 = s0, x->s1 = s1; expect_pop();
return x->symbol = ARG_FREEFORM, 1; } return x->symbol = ARG_FREEFORM, 1; }
<command, command_end> "]" => text <edict, edict_end> "]" => text
{ if(scan.command.size) return 0; { if(scan.edict.size) return 0;
goto scan; } goto scan; }
<command, command_keyword, command_date, command_freeform, command_end> * <edict, edict_keyword, edict_date, edict_freeform, edict_end> *
{ return x->symbol = ERROR, 0; } { return x->symbol = SYNTAX, 0; }
*/ */
} }