lbrycrd/lib/univalue_read.cpp
MarcoFalke 619bb05037 Squashed 'src/univalue/' changes from 16a1f7f6e..fe805ea74
fe805ea74 Declare single-argument (non-converting) constructors "explicit"
8a2d6f1e3 Merge pull request #41 from jgarzik/get-obj-map
ba341a20d Add getObjMap() helper method.  Also, constify checkObject().
ceb119413 Handle .pushKV() and .checkObject() edge cases.
107db9829 Add ::push_back(double) method for feature parity.
d41530031 Move one-line implementation of UniValue::read() to header.
52e85b35b Move exception-throwing get_* methods into separate implementation module.
dac529675 README.md: update code quotes
3e31dcffb README.md: close code quote
d09b8429d Update README.md
f1b86edb4 Convert README to markdown style.
1dfe464ef Import UniValue class unit tests from bitcoin project.
0d3e74dd1 operator[] takes size_t index parameter (versus unsigned int)
640158fa2 Private findKey() method becomes size_t clean, and returns bool on failure.
709913585 Merge pull request #36 from ryanofsky/pr/end-str
a31231b51 Version 1.0.3
4fd5444d1 Reject unterminated strings
81eba332b Merge pull request #26 from isle2983/pushBackHelpers
36405413e Merge PR #32 from branch 'nul-not-special' of git://github.com/ryanofsky/univalue into merge
89bb07322 Merge pull request #31 from ryanofsky/raw-literals
511008c36 Merge pull request #30 from ryanofsky/test-driver
77974f3a9 Merge pull request #34 from paveljanik/20161116_Wshadow_codepoint
a38fcd355 Do not shadow member variable codepoint.
fd32d1ab8 Don't require nul-terminated string inputs
0bb1439d0 Support parsing raw literals in UniValue
28876d045 Merge pull request #29 from btcdrak/exportspace
839ccd71f Add test driver for JSONTestSuite
26ef3fff1 Remove trailing whitespace from JSON export
cfa0384d6 Convenience wrappers for push_back-ing integer types
REVERT: 16a1f7f6e Merge #3: Pull upstream
REVERT: daf1285af Merge pull request #2 from jgarzik/master
REVERT: f32df99e9 Merge branch '2016_04_unicode' into bitcoin
REVERT: 280b191cb Merge remote-tracking branch 'jgarzik/master' into bitcoin
REVERT: 2740c4f71 Merge branch '2015_11_escape_plan' into bitcoin

git-subtree-dir: src/univalue
git-subtree-split: fe805ea74f8919382720b09a905a14e81311b3ad
2017-09-29 14:27:20 +02:00

451 lines
12 KiB
C++

// Copyright 2014 BitPay Inc.
// Distributed under the MIT software license, see the accompanying
// file COPYING or http://www.opensource.org/licenses/mit-license.php.
#include <string.h>
#include <vector>
#include <stdio.h>
#include "univalue.h"
#include "univalue_utffilter.h"
using namespace std;
static bool json_isdigit(int ch)
{
return ((ch >= '0') && (ch <= '9'));
}
// convert hexadecimal string to unsigned integer
static const char *hatoui(const char *first, const char *last,
unsigned int& out)
{
unsigned int result = 0;
for (; first != last; ++first)
{
int digit;
if (json_isdigit(*first))
digit = *first - '0';
else if (*first >= 'a' && *first <= 'f')
digit = *first - 'a' + 10;
else if (*first >= 'A' && *first <= 'F')
digit = *first - 'A' + 10;
else
break;
result = 16 * result + digit;
}
out = result;
return first;
}
enum jtokentype getJsonToken(string& tokenVal, unsigned int& consumed,
const char *raw, const char *end)
{
tokenVal.clear();
consumed = 0;
const char *rawStart = raw;
while (raw < end && (json_isspace(*raw))) // skip whitespace
raw++;
if (raw >= end)
return JTOK_NONE;
switch (*raw) {
case '{':
raw++;
consumed = (raw - rawStart);
return JTOK_OBJ_OPEN;
case '}':
raw++;
consumed = (raw - rawStart);
return JTOK_OBJ_CLOSE;
case '[':
raw++;
consumed = (raw - rawStart);
return JTOK_ARR_OPEN;
case ']':
raw++;
consumed = (raw - rawStart);
return JTOK_ARR_CLOSE;
case ':':
raw++;
consumed = (raw - rawStart);
return JTOK_COLON;
case ',':
raw++;
consumed = (raw - rawStart);
return JTOK_COMMA;
case 'n':
case 't':
case 'f':
if (!strncmp(raw, "null", 4)) {
raw += 4;
consumed = (raw - rawStart);
return JTOK_KW_NULL;
} else if (!strncmp(raw, "true", 4)) {
raw += 4;
consumed = (raw - rawStart);
return JTOK_KW_TRUE;
} else if (!strncmp(raw, "false", 5)) {
raw += 5;
consumed = (raw - rawStart);
return JTOK_KW_FALSE;
} else
return JTOK_ERR;
case '-':
case '0':
case '1':
case '2':
case '3':
case '4':
case '5':
case '6':
case '7':
case '8':
case '9': {
// part 1: int
string numStr;
const char *first = raw;
const char *firstDigit = first;
if (!json_isdigit(*firstDigit))
firstDigit++;
if ((*firstDigit == '0') && json_isdigit(firstDigit[1]))
return JTOK_ERR;
numStr += *raw; // copy first char
raw++;
if ((*first == '-') && (raw < end) && (!json_isdigit(*raw)))
return JTOK_ERR;
while (raw < end && json_isdigit(*raw)) { // copy digits
numStr += *raw;
raw++;
}
// part 2: frac
if (raw < end && *raw == '.') {
numStr += *raw; // copy .
raw++;
if (raw >= end || !json_isdigit(*raw))
return JTOK_ERR;
while (raw < end && json_isdigit(*raw)) { // copy digits
numStr += *raw;
raw++;
}
}
// part 3: exp
if (raw < end && (*raw == 'e' || *raw == 'E')) {
numStr += *raw; // copy E
raw++;
if (raw < end && (*raw == '-' || *raw == '+')) { // copy +/-
numStr += *raw;
raw++;
}
if (raw >= end || !json_isdigit(*raw))
return JTOK_ERR;
while (raw < end && json_isdigit(*raw)) { // copy digits
numStr += *raw;
raw++;
}
}
tokenVal = numStr;
consumed = (raw - rawStart);
return JTOK_NUMBER;
}
case '"': {
raw++; // skip "
string valStr;
JSONUTF8StringFilter writer(valStr);
while (true) {
if (raw >= end || (unsigned char)*raw < 0x20)
return JTOK_ERR;
else if (*raw == '\\') {
raw++; // skip backslash
if (raw >= end)
return JTOK_ERR;
switch (*raw) {
case '"': writer.push_back('\"'); break;
case '\\': writer.push_back('\\'); break;
case '/': writer.push_back('/'); break;
case 'b': writer.push_back('\b'); break;
case 'f': writer.push_back('\f'); break;
case 'n': writer.push_back('\n'); break;
case 'r': writer.push_back('\r'); break;
case 't': writer.push_back('\t'); break;
case 'u': {
unsigned int codepoint;
if (raw + 1 + 4 >= end ||
hatoui(raw + 1, raw + 1 + 4, codepoint) !=
raw + 1 + 4)
return JTOK_ERR;
writer.push_back_u(codepoint);
raw += 4;
break;
}
default:
return JTOK_ERR;
}
raw++; // skip esc'd char
}
else if (*raw == '"') {
raw++; // skip "
break; // stop scanning
}
else {
writer.push_back(*raw);
raw++;
}
}
if (!writer.finalize())
return JTOK_ERR;
tokenVal = valStr;
consumed = (raw - rawStart);
return JTOK_STRING;
}
default:
return JTOK_ERR;
}
}
enum expect_bits {
EXP_OBJ_NAME = (1U << 0),
EXP_COLON = (1U << 1),
EXP_ARR_VALUE = (1U << 2),
EXP_VALUE = (1U << 3),
EXP_NOT_VALUE = (1U << 4),
};
#define expect(bit) (expectMask & (EXP_##bit))
#define setExpect(bit) (expectMask |= EXP_##bit)
#define clearExpect(bit) (expectMask &= ~EXP_##bit)
bool UniValue::read(const char *raw, size_t size)
{
clear();
uint32_t expectMask = 0;
vector<UniValue*> stack;
string tokenVal;
unsigned int consumed;
enum jtokentype tok = JTOK_NONE;
enum jtokentype last_tok = JTOK_NONE;
const char* end = raw + size;
do {
last_tok = tok;
tok = getJsonToken(tokenVal, consumed, raw, end);
if (tok == JTOK_NONE || tok == JTOK_ERR)
return false;
raw += consumed;
bool isValueOpen = jsonTokenIsValue(tok) ||
tok == JTOK_OBJ_OPEN || tok == JTOK_ARR_OPEN;
if (expect(VALUE)) {
if (!isValueOpen)
return false;
clearExpect(VALUE);
} else if (expect(ARR_VALUE)) {
bool isArrValue = isValueOpen || (tok == JTOK_ARR_CLOSE);
if (!isArrValue)
return false;
clearExpect(ARR_VALUE);
} else if (expect(OBJ_NAME)) {
bool isObjName = (tok == JTOK_OBJ_CLOSE || tok == JTOK_STRING);
if (!isObjName)
return false;
} else if (expect(COLON)) {
if (tok != JTOK_COLON)
return false;
clearExpect(COLON);
} else if (!expect(COLON) && (tok == JTOK_COLON)) {
return false;
}
if (expect(NOT_VALUE)) {
if (isValueOpen)
return false;
clearExpect(NOT_VALUE);
}
switch (tok) {
case JTOK_OBJ_OPEN:
case JTOK_ARR_OPEN: {
VType utyp = (tok == JTOK_OBJ_OPEN ? VOBJ : VARR);
if (!stack.size()) {
if (utyp == VOBJ)
setObject();
else
setArray();
stack.push_back(this);
} else {
UniValue tmpVal(utyp);
UniValue *top = stack.back();
top->values.push_back(tmpVal);
UniValue *newTop = &(top->values.back());
stack.push_back(newTop);
}
if (utyp == VOBJ)
setExpect(OBJ_NAME);
else
setExpect(ARR_VALUE);
break;
}
case JTOK_OBJ_CLOSE:
case JTOK_ARR_CLOSE: {
if (!stack.size() || (last_tok == JTOK_COMMA))
return false;
VType utyp = (tok == JTOK_OBJ_CLOSE ? VOBJ : VARR);
UniValue *top = stack.back();
if (utyp != top->getType())
return false;
stack.pop_back();
clearExpect(OBJ_NAME);
setExpect(NOT_VALUE);
break;
}
case JTOK_COLON: {
if (!stack.size())
return false;
UniValue *top = stack.back();
if (top->getType() != VOBJ)
return false;
setExpect(VALUE);
break;
}
case JTOK_COMMA: {
if (!stack.size() ||
(last_tok == JTOK_COMMA) || (last_tok == JTOK_ARR_OPEN))
return false;
UniValue *top = stack.back();
if (top->getType() == VOBJ)
setExpect(OBJ_NAME);
else
setExpect(ARR_VALUE);
break;
}
case JTOK_KW_NULL:
case JTOK_KW_TRUE:
case JTOK_KW_FALSE: {
UniValue tmpVal;
switch (tok) {
case JTOK_KW_NULL:
// do nothing more
break;
case JTOK_KW_TRUE:
tmpVal.setBool(true);
break;
case JTOK_KW_FALSE:
tmpVal.setBool(false);
break;
default: /* impossible */ break;
}
if (!stack.size()) {
*this = tmpVal;
break;
}
UniValue *top = stack.back();
top->values.push_back(tmpVal);
setExpect(NOT_VALUE);
break;
}
case JTOK_NUMBER: {
UniValue tmpVal(VNUM, tokenVal);
if (!stack.size()) {
*this = tmpVal;
break;
}
UniValue *top = stack.back();
top->values.push_back(tmpVal);
setExpect(NOT_VALUE);
break;
}
case JTOK_STRING: {
if (expect(OBJ_NAME)) {
UniValue *top = stack.back();
top->keys.push_back(tokenVal);
clearExpect(OBJ_NAME);
setExpect(COLON);
} else {
UniValue tmpVal(VSTR, tokenVal);
if (!stack.size()) {
*this = tmpVal;
break;
}
UniValue *top = stack.back();
top->values.push_back(tmpVal);
}
setExpect(NOT_VALUE);
break;
}
default:
return false;
}
} while (!stack.empty ());
/* Check that nothing follows the initial construct (parsed above). */
tok = getJsonToken(tokenVal, consumed, raw, end);
if (tok != JTOK_NONE)
return false;
return true;
}