abdelrahman/helloJSON
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Handle objects, arrays and strings

Browse Source
This commit is contained in:
Abdelrahman Said 2023-06-17 15:27:16 +01:00
parent d7791e3794
commit 43630526dc
4 changed files with 243 additions and 537 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

518
compile_commands.json
View File

@ -17,42 +17,6 @@
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/main.c", "file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/main.c",
"output": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/main" "output": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/main"
}, },
{
"arguments": [
"/usr/bin/clang",
"-c",
"-g",
"-Wall",
"-Werror",
"-pedantic",
"-Iinclude",
"-Iinclude/lexer",
"-o",
"main",
"src/lexer/array_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/array_states.c",
"output": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/main"
},
{
"arguments": [
"/usr/bin/clang",
"-c",
"-g",
"-Wall",
"-Werror",
"-pedantic",
"-Iinclude",
"-Iinclude/lexer",
"-o",
"main",
"src/lexer/keyword_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/keyword_states.c",
"output": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/main"
},
{ {
"arguments": [ "arguments": [
"/usr/bin/clang", "/usr/bin/clang",
@ -71,78 +35,6 @@
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/lexer_states.c", "file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/lexer_states.c",
"output": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/main" "output": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/main"
}, },
{
"arguments": [
"/usr/bin/clang",
"-c",
"-g",
"-Wall",
"-Werror",
"-pedantic",
"-Iinclude",
"-Iinclude/lexer",
"-o",
"main",
"src/lexer/num_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/num_states.c",
"output": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/main"
},
{
"arguments": [
"/usr/bin/clang",
"-c",
"-g",
"-Wall",
"-Werror",
"-pedantic",
"-Iinclude",
"-Iinclude/lexer",
"-o",
"main",
"src/lexer/obj_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/obj_states.c",
"output": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/main"
},
{
"arguments": [
"/usr/bin/clang",
"-c",
"-g",
"-Wall",
"-Werror",
"-pedantic",
"-Iinclude",
"-Iinclude/lexer",
"-o",
"main",
"src/lexer/string_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/string_states.c",
"output": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/main"
},
{
"arguments": [
"/usr/bin/clang",
"-c",
"-g",
"-Wall",
"-Werror",
"-pedantic",
"-Iinclude",
"-Iinclude/lexer",
"-o",
"main",
"src/lexer/value_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/value_states.c",
"output": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/main"
},
{ {
"arguments": [ "arguments": [
"/usr/bin/clang-16", "/usr/bin/clang-16",
@ -203,12 +95,12 @@
"-x", "-x",
"c", "c",
"-o", "-o",
"/tmp/main-3a4e7b.o", "/tmp/main-759cce.o",
"src/main.c" "src/main.c"
], ],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json", "directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/main.c", "file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/main.c",
"output": "/tmp/main-3a4e7b.o" "output": "/tmp/main-759cce.o"
}, },
{ {
"arguments": [ "arguments": [
@ -270,413 +162,11 @@
"-x", "-x",
"c", "c",
"-o", "-o",
"/tmp/array_states-ba3bba.o", "/tmp/lexer_states-c32084.o",
"src/lexer/array_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/array_states.c",
"output": "/tmp/array_states-ba3bba.o"
},
{
"arguments": [
"/usr/bin/clang-16",
"-cc1",
"-triple",
"x86_64-redhat-linux-gnu",
"-emit-obj",
"-mrelax-all",
"-disable-free",
"-clear-ast-before-backend",
"-disable-llvm-verifier",
"-discard-value-names",
"-main-file-name",
"-mrelocation-model",
"static",
"-mframe-pointer=all",
"-fmath-errno",
"-ffp-contract=on",
"-fno-rounding-math",
"-mconstructor-aliases",
"-funwind-tables=2",
"-target-cpu",
"x86-64",
"-tune-cpu",
"generic",
"-mllvm",
"-treat-scalable-fixed-error-as-warning",
"-debug-info-kind=constructor",
"-dwarf-version=4",
"-debugger-tuning=gdb",
"-fcoverage-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-resource-dir",
"/usr/lib64/clang/16",
"-I",
"include",
"-I",
"include/lexer",
"-internal-isystem",
"/usr/lib64/clang/16/include",
"-internal-isystem",
"/usr/local/include",
"-internal-isystem",
"/usr/bin/../lib/gcc/x86_64-redhat-linux/13/../../../../x86_64-redhat-linux/include",
"-internal-externc-isystem",
"/include",
"-internal-externc-isystem",
"/usr/include",
"-Wall",
"-Werror",
"-pedantic",
"-fdebug-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-ferror-limit",
"19",
"-fgnuc-version=4.2.1",
"-fcolor-diagnostics",
"-faddrsig",
"-D__GCC_HAVE_DWARF2_CFI_ASM=1",
"-x",
"c",
"-o",
"/tmp/keyword_states-0752de.o",
"src/lexer/keyword_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/keyword_states.c",
"output": "/tmp/keyword_states-0752de.o"
},
{
"arguments": [
"/usr/bin/clang-16",
"-cc1",
"-triple",
"x86_64-redhat-linux-gnu",
"-emit-obj",
"-mrelax-all",
"-disable-free",
"-clear-ast-before-backend",
"-disable-llvm-verifier",
"-discard-value-names",
"-main-file-name",
"-mrelocation-model",
"static",
"-mframe-pointer=all",
"-fmath-errno",
"-ffp-contract=on",
"-fno-rounding-math",
"-mconstructor-aliases",
"-funwind-tables=2",
"-target-cpu",
"x86-64",
"-tune-cpu",
"generic",
"-mllvm",
"-treat-scalable-fixed-error-as-warning",
"-debug-info-kind=constructor",
"-dwarf-version=4",
"-debugger-tuning=gdb",
"-fcoverage-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-resource-dir",
"/usr/lib64/clang/16",
"-I",
"include",
"-I",
"include/lexer",
"-internal-isystem",
"/usr/lib64/clang/16/include",
"-internal-isystem",
"/usr/local/include",
"-internal-isystem",
"/usr/bin/../lib/gcc/x86_64-redhat-linux/13/../../../../x86_64-redhat-linux/include",
"-internal-externc-isystem",
"/include",
"-internal-externc-isystem",
"/usr/include",
"-Wall",
"-Werror",
"-pedantic",
"-fdebug-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-ferror-limit",
"19",
"-fgnuc-version=4.2.1",
"-fcolor-diagnostics",
"-faddrsig",
"-D__GCC_HAVE_DWARF2_CFI_ASM=1",
"-x",
"c",
"-o",
"/tmp/lexer_states-b5dca0.o",
"src/lexer/lexer_states.c" "src/lexer/lexer_states.c"
], ],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json", "directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/lexer_states.c", "file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/lexer_states.c",
"output": "/tmp/lexer_states-b5dca0.o" "output": "/tmp/lexer_states-c32084.o"
},
{
"arguments": [
"/usr/bin/clang-16",
"-cc1",
"-triple",
"x86_64-redhat-linux-gnu",
"-emit-obj",
"-mrelax-all",
"-disable-free",
"-clear-ast-before-backend",
"-disable-llvm-verifier",
"-discard-value-names",
"-main-file-name",
"-mrelocation-model",
"static",
"-mframe-pointer=all",
"-fmath-errno",
"-ffp-contract=on",
"-fno-rounding-math",
"-mconstructor-aliases",
"-funwind-tables=2",
"-target-cpu",
"x86-64",
"-tune-cpu",
"generic",
"-mllvm",
"-treat-scalable-fixed-error-as-warning",
"-debug-info-kind=constructor",
"-dwarf-version=4",
"-debugger-tuning=gdb",
"-fcoverage-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-resource-dir",
"/usr/lib64/clang/16",
"-I",
"include",
"-I",
"include/lexer",
"-internal-isystem",
"/usr/lib64/clang/16/include",
"-internal-isystem",
"/usr/local/include",
"-internal-isystem",
"/usr/bin/../lib/gcc/x86_64-redhat-linux/13/../../../../x86_64-redhat-linux/include",
"-internal-externc-isystem",
"/include",
"-internal-externc-isystem",
"/usr/include",
"-Wall",
"-Werror",
"-pedantic",
"-fdebug-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-ferror-limit",
"19",
"-fgnuc-version=4.2.1",
"-fcolor-diagnostics",
"-faddrsig",
"-D__GCC_HAVE_DWARF2_CFI_ASM=1",
"-x",
"c",
"-o",
"/tmp/num_states-f8ad26.o",
"src/lexer/num_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/num_states.c",
"output": "/tmp/num_states-f8ad26.o"
},
{
"arguments": [
"/usr/bin/clang-16",
"-cc1",
"-triple",
"x86_64-redhat-linux-gnu",
"-emit-obj",
"-mrelax-all",
"-disable-free",
"-clear-ast-before-backend",
"-disable-llvm-verifier",
"-discard-value-names",
"-main-file-name",
"-mrelocation-model",
"static",
"-mframe-pointer=all",
"-fmath-errno",
"-ffp-contract=on",
"-fno-rounding-math",
"-mconstructor-aliases",
"-funwind-tables=2",
"-target-cpu",
"x86-64",
"-tune-cpu",
"generic",
"-mllvm",
"-treat-scalable-fixed-error-as-warning",
"-debug-info-kind=constructor",
"-dwarf-version=4",
"-debugger-tuning=gdb",
"-fcoverage-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-resource-dir",
"/usr/lib64/clang/16",
"-I",
"include",
"-I",
"include/lexer",
"-internal-isystem",
"/usr/lib64/clang/16/include",
"-internal-isystem",
"/usr/local/include",
"-internal-isystem",
"/usr/bin/../lib/gcc/x86_64-redhat-linux/13/../../../../x86_64-redhat-linux/include",
"-internal-externc-isystem",
"/include",
"-internal-externc-isystem",
"/usr/include",
"-Wall",
"-Werror",
"-pedantic",
"-fdebug-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-ferror-limit",
"19",
"-fgnuc-version=4.2.1",
"-fcolor-diagnostics",
"-faddrsig",
"-D__GCC_HAVE_DWARF2_CFI_ASM=1",
"-x",
"c",
"-o",
"/tmp/obj_states-9ef381.o",
"src/lexer/obj_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/obj_states.c",
"output": "/tmp/obj_states-9ef381.o"
},
{
"arguments": [
"/usr/bin/clang-16",
"-cc1",
"-triple",
"x86_64-redhat-linux-gnu",
"-emit-obj",
"-mrelax-all",
"-disable-free",
"-clear-ast-before-backend",
"-disable-llvm-verifier",
"-discard-value-names",
"-main-file-name",
"-mrelocation-model",
"static",
"-mframe-pointer=all",
"-fmath-errno",
"-ffp-contract=on",
"-fno-rounding-math",
"-mconstructor-aliases",
"-funwind-tables=2",
"-target-cpu",
"x86-64",
"-tune-cpu",
"generic",
"-mllvm",
"-treat-scalable-fixed-error-as-warning",
"-debug-info-kind=constructor",
"-dwarf-version=4",
"-debugger-tuning=gdb",
"-fcoverage-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-resource-dir",
"/usr/lib64/clang/16",
"-I",
"include",
"-I",
"include/lexer",
"-internal-isystem",
"/usr/lib64/clang/16/include",
"-internal-isystem",
"/usr/local/include",
"-internal-isystem",
"/usr/bin/../lib/gcc/x86_64-redhat-linux/13/../../../../x86_64-redhat-linux/include",
"-internal-externc-isystem",
"/include",
"-internal-externc-isystem",
"/usr/include",
"-Wall",
"-Werror",
"-pedantic",
"-fdebug-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-ferror-limit",
"19",
"-fgnuc-version=4.2.1",
"-fcolor-diagnostics",
"-faddrsig",
"-D__GCC_HAVE_DWARF2_CFI_ASM=1",
"-x",
"c",
"-o",
"/tmp/string_states-c65204.o",
"src/lexer/string_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/string_states.c",
"output": "/tmp/string_states-c65204.o"
},
{
"arguments": [
"/usr/bin/clang-16",
"-cc1",
"-triple",
"x86_64-redhat-linux-gnu",
"-emit-obj",
"-mrelax-all",
"-disable-free",
"-clear-ast-before-backend",
"-disable-llvm-verifier",
"-discard-value-names",
"-main-file-name",
"-mrelocation-model",
"static",
"-mframe-pointer=all",
"-fmath-errno",
"-ffp-contract=on",
"-fno-rounding-math",
"-mconstructor-aliases",
"-funwind-tables=2",
"-target-cpu",
"x86-64",
"-tune-cpu",
"generic",
"-mllvm",
"-treat-scalable-fixed-error-as-warning",
"-debug-info-kind=constructor",
"-dwarf-version=4",
"-debugger-tuning=gdb",
"-fcoverage-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-resource-dir",
"/usr/lib64/clang/16",
"-I",
"include",
"-I",
"include/lexer",
"-internal-isystem",
"/usr/lib64/clang/16/include",
"-internal-isystem",
"/usr/local/include",
"-internal-isystem",
"/usr/bin/../lib/gcc/x86_64-redhat-linux/13/../../../../x86_64-redhat-linux/include",
"-internal-externc-isystem",
"/include",
"-internal-externc-isystem",
"/usr/include",
"-Wall",
"-Werror",
"-pedantic",
"-fdebug-compilation-dir=/home/abdelrahman/dev_work/say_it_in_json",
"-ferror-limit",
"19",
"-fgnuc-version=4.2.1",
"-fcolor-diagnostics",
"-faddrsig",
"-D__GCC_HAVE_DWARF2_CFI_ASM=1",
"-x",
"c",
"-o",
"/tmp/value_states-016fa1.o",
"src/lexer/value_states.c"
],
"directory": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json",
"file": "/mnt/3A5CDF785CDF2CFF/Users/abdoo/dev/say_it_in_json/src/lexer/value_states.c",
"output": "/tmp/value_states-016fa1.o"
} }
] ]

7
include/lexer/lexer_states.h
View File

@ -1,6 +1,13 @@
#ifndef LEXER_STATES_H #ifndef LEXER_STATES_H
#define LEXER_STATES_H #define LEXER_STATES_H
#include <stdbool.h>
#define VALID_JSON true
#define INVALID_JSON false
typedef struct lexer lexer_t; typedef struct lexer lexer_t;
bool validate_json(char *json);
#endif // !LEXER_STATES_H #endif // !LEXER_STATES_H

224
src/lexer/lexer_states.c
View File

@ -1,21 +1,29 @@
#include "lexer_states.h" #include "lexer_states.h"
#include "aliases.h" #include "aliases.h"
#include <ctype.h> #include <ctype.h>
#include <stdbool.h>
#include <stdio.h>
#define MAX_KEYWORD_LENGTH 5
#define MAX_STACK_CAPACITY 1024
// clang-format off // clang-format off
typedef enum { typedef enum {
// GENERAL STATES // GENERAL STATES
LEXER_STATE_START, LEXER_STATE_START,
LEXER_STATE_SUCCESS,
LEXER_STATE_ERROR, LEXER_STATE_ERROR,
LEXER_STATE_VALUE, LEXER_STATE_VALUE,
LEXER_STATE_KEYWORD, LEXER_STATE_KEYWORD,
// COLLECTION STATES // COLLECTION STATES
LEXER_STATE_OBJECT_START,
LEXER_STATE_OBJECT, LEXER_STATE_OBJECT,
LEXER_STATE_OBJECT_END,
LEXER_STATE_ARRAY_START,
LEXER_STATE_ARRAY, LEXER_STATE_ARRAY,
LEXER_STATE_ARRAY_END,
LEXER_STATE_LAST_COLLECTION,
// OBJECT STATES // OBJECT STATES
LEXER_STATE_KEY_START, LEXER_STATE_KEY,
LEXER_STATE_KEY_END,
// NUMBER STATES // NUMBER STATES
LEXER_STATE_NUMBER_START, LEXER_STATE_NUMBER_START,
LEXER_STATE_DECIMAL, LEXER_STATE_DECIMAL,
@ -27,6 +35,7 @@ typedef enum {
LEXER_STATE_NUMBER_END, LEXER_STATE_NUMBER_END,
// STRING STATES // STRING STATES
LEXER_STATE_STRING, LEXER_STATE_STRING,
LEXER_STATE_STRING_END,
LEXER_STATE_ESCAPE_SEQUENCE, LEXER_STATE_ESCAPE_SEQUENCE,
LEXER_STATE_UNICODE_HEX, LEXER_STATE_UNICODE_HEX,
@ -34,41 +43,106 @@ typedef enum {
} lexer_state_t; } lexer_state_t;
// clang-format on // clang-format on
typedef struct {
lexer_state_t stack[MAX_STACK_CAPACITY];
u64 size;
} state_stack_t;
struct lexer { struct lexer {
lexer_state_t current; lexer_state_t current;
state_stack_t stack;
u64 line; u64 line;
u64 column; u64 column;
char current_keyword[MAX_KEYWORD_LENGTH + 1];
}; };
void stack_push(state_stack_t *stack, lexer_state_t value);
lexer_state_t stack_pop(state_stack_t *stack);
void lexer_state_machine(lexer_t *lexer, char input); void lexer_state_machine(lexer_t *lexer, char input);
lexer_state_t handle_lexer_start(char input); lexer_state_t handle_lexer_start(char input);
lexer_state_t handle_object(char input); lexer_state_t handle_last_collection(char input);
lexer_state_t handle_character(char input); lexer_state_t handle_collection_end(lexer_t *lexer, char input);
lexer_state_t handle_object(lexer_t *lexer, char input);
lexer_state_t handle_array(lexer_t *lexer, char input);
lexer_state_t handle_key(lexer_t *lexer);
lexer_state_t handle_value(lexer_t *lexer, char input);
lexer_state_t handle_string(char input);
lexer_state_t handle_string_end(lexer_t *lexer, char input);
lexer_state_t handle_escape_sequence(char input); lexer_state_t handle_escape_sequence(char input);
bool validate_json(char *json) {
lexer_t lexer = {0};
lexer.current = LEXER_STATE_START;
for (char *c = json; *c != '\0'; ++c) {
// printf("\nINPUT=>%s\n", c);
// printf("STACK SIZE: %zu\n", lexer.stack.size);
lexer_state_machine(&lexer, *c);
if (lexer.current == LEXER_STATE_ERROR) {
return INVALID_JSON;
}
}
return lexer.current == LEXER_STATE_LAST_COLLECTION || lexer.stack.size == 0;
}
void stack_push(state_stack_t *stack, lexer_state_t state) {
if (stack->size + 1 >= MAX_STACK_CAPACITY) {
return;
}
stack->stack[(stack->size)++] = state;
}
lexer_state_t stack_pop(state_stack_t *stack) {
if (stack->size == 0) {
return LEXER_STATE_ERROR;
}
lexer_state_t state = stack->stack[--(stack->size)];
return state;
}
void lexer_state_machine(lexer_t *lexer, char input) { void lexer_state_machine(lexer_t *lexer, char input) {
switch (lexer->current) { switch (lexer->current) {
case LEXER_STATE_START: case LEXER_STATE_START:
lexer->current = handle_lexer_start(input); lexer->current = handle_lexer_start(input);
break; break;
case LEXER_STATE_SUCCESS:
break;
case LEXER_STATE_ERROR:
break;
case LEXER_STATE_VALUE: case LEXER_STATE_VALUE:
lexer->current = handle_value(lexer, input);
break; break;
case LEXER_STATE_KEYWORD: case LEXER_STATE_KEYWORD:
break; break;
case LEXER_STATE_OBJECT_START:
stack_push(&(lexer->stack), LEXER_STATE_OBJECT);
// break is left intentionally here to utilise the fallthrough behaviour of
// the switch statement
case LEXER_STATE_OBJECT: case LEXER_STATE_OBJECT:
lexer->current = handle_object(input); lexer->current = handle_object(lexer, input);
break; break;
case LEXER_STATE_ARRAY_START:
stack_push(&(lexer->stack), LEXER_STATE_ARRAY);
// break is left intentionally here to utilise the fallthrough behaviour of
// the switch statement
case LEXER_STATE_ARRAY: case LEXER_STATE_ARRAY:
lexer->current = LEXER_STATE_VALUE; lexer->current = handle_array(lexer, input);
break; break;
case LEXER_STATE_KEY_START: case LEXER_STATE_OBJECT_END:
lexer->current = LEXER_STATE_STRING; case LEXER_STATE_ARRAY_END:
if (lexer->stack.size > 1) {
stack_pop(&(lexer->stack));
lexer->current = lexer->stack.stack[lexer->stack.size - 1];
} else {
lexer->current = LEXER_STATE_LAST_COLLECTION;
}
break; break;
case LEXER_STATE_KEY_END: case LEXER_STATE_KEY:
lexer->current = handle_key(lexer);
break; break;
case LEXER_STATE_NUMBER_START: case LEXER_STATE_NUMBER_START:
break; break;
@ -87,15 +161,23 @@ void lexer_state_machine(lexer_t *lexer, char input) {
case LEXER_STATE_NUMBER_END: case LEXER_STATE_NUMBER_END:
break; break;
case LEXER_STATE_STRING: case LEXER_STATE_STRING:
lexer->current = handle_character(input); lexer->current = handle_string(input);
break;
case LEXER_STATE_STRING_END:
lexer->current = handle_string_end(lexer, input);
break; break;
case LEXER_STATE_ESCAPE_SEQUENCE: case LEXER_STATE_ESCAPE_SEQUENCE:
lexer->current = handle_escape_sequence(input); lexer->current = handle_escape_sequence(input);
break; break;
case LEXER_STATE_UNICODE_HEX: case LEXER_STATE_UNICODE_HEX:
break; break;
default: case LEXER_STATE_LAST_COLLECTION:
lexer->current = handle_last_collection(input);
break;
case LEXER_STATE_ERROR:
case COUNT_LEXER_STATES:
lexer->current = LEXER_STATE_ERROR; lexer->current = LEXER_STATE_ERROR;
break;
} }
} }
@ -106,35 +188,133 @@ lexer_state_t handle_lexer_start(char input) {
switch (input) { switch (input) {
case '{': case '{':
return LEXER_STATE_OBJECT; return LEXER_STATE_OBJECT_START;
case '[': case '[':
return LEXER_STATE_ARRAY; return LEXER_STATE_ARRAY_START;
} }
return LEXER_STATE_ERROR; return LEXER_STATE_ERROR;
} }
lexer_state_t handle_object(char input) { lexer_state_t handle_last_collection(char input) {
if (isspace(input)) {
return LEXER_STATE_LAST_COLLECTION;
}
return LEXER_STATE_ERROR;
}
lexer_state_t handle_collection_end(lexer_t *lexer, char input) {
// No need to ignore space as this is only called when input is } or ]
lexer->current = lexer->stack.stack[lexer->stack.size - 1];
bool object_end = lexer->current == LEXER_STATE_OBJECT && input == '}';
if (object_end) {
return LEXER_STATE_OBJECT_END;
}
bool array_end = lexer->current == LEXER_STATE_ARRAY && input == ']';
if (array_end) {
return LEXER_STATE_ARRAY_END;
}
return LEXER_STATE_ERROR;
}
lexer_state_t handle_object(lexer_t *lexer, char input) {
if (isspace(input)) { if (isspace(input)) {
return LEXER_STATE_OBJECT; return LEXER_STATE_OBJECT;
} else if (input == '"') { } else if (input == '"') {
return LEXER_STATE_KEY_START; stack_push(&(lexer->stack), LEXER_STATE_KEY);
return LEXER_STATE_KEY;
} else if (input == '}') {
return handle_collection_end(lexer, input);
} }
return LEXER_STATE_ERROR; return LEXER_STATE_ERROR;
} }
lexer_state_t handle_character(char input) { lexer_state_t handle_array(lexer_t *lexer, char input) {
if (isspace(input)) {
return LEXER_STATE_ARRAY;
} else if (input == '"') {
stack_push(&(lexer->stack), LEXER_STATE_VALUE);
return LEXER_STATE_STRING;
} else if (input == ']') {
return handle_collection_end(lexer, input);
}
return LEXER_STATE_VALUE;
}
lexer_state_t handle_key(lexer_t *lexer) { return LEXER_STATE_STRING; }
lexer_state_t handle_value(lexer_t *lexer, char input) {
if (isspace(input)) {
return LEXER_STATE_VALUE;
}
switch (input) {
case '"':
stack_push(&(lexer->stack), LEXER_STATE_VALUE);
return LEXER_STATE_STRING;
case '0':
return LEXER_STATE_DECIMAL;
case '{':
return LEXER_STATE_OBJECT_START;
case '[':
return LEXER_STATE_ARRAY_START;
case 't':
case 'f':
case 'n':
return LEXER_STATE_KEYWORD;
}
return LEXER_STATE_ERROR;
}
lexer_state_t handle_string(char input) {
switch (input) { switch (input) {
case '\\': case '\\':
return LEXER_STATE_ESCAPE_SEQUENCE; return LEXER_STATE_ESCAPE_SEQUENCE;
case '"': case '"':
return LEXER_STATE_STRING; // WHAT TO RETURN HERE? return LEXER_STATE_STRING_END;
} }
return LEXER_STATE_STRING; return LEXER_STATE_STRING;
} }
lexer_state_t handle_string_end(lexer_t *lexer, char input) {
if (isspace(input)) {
return LEXER_STATE_STRING_END;
}
lexer->current = stack_pop(&(lexer->stack));
bool key_end = lexer->current == LEXER_STATE_KEY && input == ':';
if (key_end) {
return LEXER_STATE_VALUE;
}
bool value_end = lexer->current == LEXER_STATE_VALUE && input == ',';
if (value_end) {
return lexer->stack.stack[lexer->stack.size - 1];
}
bool collection_end = input == '}' || input == ']';
return collection_end ? handle_collection_end(lexer, input)
: LEXER_STATE_ERROR;
}
lexer_state_t handle_escape_sequence(char input) { lexer_state_t handle_escape_sequence(char input) {
switch (input) { switch (input) {
case '"': case '"':

31
src/main.c
View File

@ -1,3 +1,32 @@
#include "aliases.h"
#include "lexer_states.h"
#include <stdio.h>
#include <stdlib.h> #include <stdlib.h>
#include <string.h>
int main(int argc, char *argv[]) { return EXIT_SUCCESS; } int main(int argc, char *argv[]) {
if (argc < 2) {
printf("NO FILE PROVIDED\n");
return EXIT_FAILURE;
}
FILE *fp = fopen(argv[1], "r");
fseek(fp, 0, SEEK_END);
u64 length = ftell(fp);
fseek(fp, 0, SEEK_SET);
char json[length + 1];
memset(json, 0, length + 1);
fread(json, sizeof(char), length, fp);
fclose(fp);
printf("\n%s\n", validate_json(json) ? "VALID" : "INVALID");
return EXIT_SUCCESS;
}