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
This commit is contained in:
MarcoFalke
2017-09-29 14:27:20 +02:00
parent 2ca7faab42
commit 619bb05037
23 changed files with 784 additions and 230 deletions

View File

@@ -4,75 +4,12 @@
// file COPYING or http://www.opensource.org/licenses/mit-license.php.
#include <stdint.h>
#include <errno.h>
#include <iomanip>
#include <limits>
#include <sstream>
#include <stdexcept>
#include <stdlib.h>
#include <string.h>
#include "univalue.h"
namespace
{
static bool ParsePrechecks(const std::string& str)
{
if (str.empty()) // No empty string allowed
return false;
if (str.size() >= 1 && (json_isspace(str[0]) || json_isspace(str[str.size()-1]))) // No padding allowed
return false;
if (str.size() != strlen(str.c_str())) // No embedded NUL characters allowed
return false;
return true;
}
bool ParseInt32(const std::string& str, int32_t *out)
{
if (!ParsePrechecks(str))
return false;
char *endp = NULL;
errno = 0; // strtol will not set errno if valid
long int n = strtol(str.c_str(), &endp, 10);
if(out) *out = (int32_t)n;
// Note that strtol returns a *long int*, so even if strtol doesn't report a over/underflow
// we still have to check that the returned value is within the range of an *int32_t*. On 64-bit
// platforms the size of these types may be different.
return endp && *endp == 0 && !errno &&
n >= std::numeric_limits<int32_t>::min() &&
n <= std::numeric_limits<int32_t>::max();
}
bool ParseInt64(const std::string& str, int64_t *out)
{
if (!ParsePrechecks(str))
return false;
char *endp = NULL;
errno = 0; // strtoll will not set errno if valid
long long int n = strtoll(str.c_str(), &endp, 10);
if(out) *out = (int64_t)n;
// Note that strtoll returns a *long long int*, so even if strtol doesn't report a over/underflow
// we still have to check that the returned value is within the range of an *int64_t*.
return endp && *endp == 0 && !errno &&
n >= std::numeric_limits<int64_t>::min() &&
n <= std::numeric_limits<int64_t>::max();
}
bool ParseDouble(const std::string& str, double *out)
{
if (!ParsePrechecks(str))
return false;
if (str.size() >= 2 && str[0] == '0' && str[1] == 'x') // No hexadecimal floats allowed
return false;
std::istringstream text(str);
text.imbue(std::locale::classic());
double result;
text >> result;
if(out) *out = result;
return text.eof() && !text.fail();
}
}
using namespace std;
const UniValue NullUniValue;
@@ -104,7 +41,7 @@ static bool validNumStr(const string& s)
{
string tokenVal;
unsigned int consumed;
enum jtokentype tt = getJsonToken(tokenVal, consumed, s.c_str());
enum jtokentype tt = getJsonToken(tokenVal, consumed, s.data(), s.data() + s.size());
return (tt == JTOK_NUMBER);
}
@@ -189,13 +126,22 @@ bool UniValue::push_backV(const std::vector<UniValue>& vec)
return true;
}
void UniValue::__pushKV(const std::string& key, const UniValue& val_)
{
keys.push_back(key);
values.push_back(val_);
}
bool UniValue::pushKV(const std::string& key, const UniValue& val_)
{
if (typ != VOBJ)
return false;
keys.push_back(key);
values.push_back(val_);
size_t idx;
if (findKey(key, idx))
values[idx] = val_;
else
__pushKV(key, val_);
return true;
}
@@ -204,30 +150,43 @@ bool UniValue::pushKVs(const UniValue& obj)
if (typ != VOBJ || obj.typ != VOBJ)
return false;
for (unsigned int i = 0; i < obj.keys.size(); i++) {
keys.push_back(obj.keys[i]);
values.push_back(obj.values.at(i));
}
for (size_t i = 0; i < obj.keys.size(); i++)
__pushKV(obj.keys[i], obj.values.at(i));
return true;
}
int UniValue::findKey(const std::string& key) const
void UniValue::getObjMap(std::map<std::string,UniValue>& kv) const
{
for (unsigned int i = 0; i < keys.size(); i++) {
if (keys[i] == key)
return (int) i;
}
if (typ != VOBJ)
return;
return -1;
kv.clear();
for (size_t i = 0; i < keys.size(); i++)
kv[keys[i]] = values[i];
}
bool UniValue::checkObject(const std::map<std::string,UniValue::VType>& t)
bool UniValue::findKey(const std::string& key, size_t& retIdx) const
{
for (size_t i = 0; i < keys.size(); i++) {
if (keys[i] == key) {
retIdx = i;
return true;
}
}
return false;
}
bool UniValue::checkObject(const std::map<std::string,UniValue::VType>& t) const
{
if (typ != VOBJ)
return false;
for (std::map<std::string,UniValue::VType>::const_iterator it = t.begin();
it != t.end(); ++it) {
int idx = findKey(it->first);
if (idx < 0)
size_t idx = 0;
if (!findKey(it->first, idx))
return false;
if (values.at(idx).getType() != it->second)
@@ -242,14 +201,14 @@ const UniValue& UniValue::operator[](const std::string& key) const
if (typ != VOBJ)
return NullUniValue;
int index = findKey(key);
if (index < 0)
size_t index = 0;
if (!findKey(key, index))
return NullUniValue;
return values.at(index);
}
const UniValue& UniValue::operator[](unsigned int index) const
const UniValue& UniValue::operator[](size_t index) const
{
if (typ != VOBJ && typ != VARR)
return NullUniValue;
@@ -283,75 +242,3 @@ const UniValue& find_value(const UniValue& obj, const std::string& name)
return NullUniValue;
}
const std::vector<std::string>& UniValue::getKeys() const
{
if (typ != VOBJ)
throw std::runtime_error("JSON value is not an object as expected");
return keys;
}
const std::vector<UniValue>& UniValue::getValues() const
{
if (typ != VOBJ && typ != VARR)
throw std::runtime_error("JSON value is not an object or array as expected");
return values;
}
bool UniValue::get_bool() const
{
if (typ != VBOOL)
throw std::runtime_error("JSON value is not a boolean as expected");
return getBool();
}
const std::string& UniValue::get_str() const
{
if (typ != VSTR)
throw std::runtime_error("JSON value is not a string as expected");
return getValStr();
}
int UniValue::get_int() const
{
if (typ != VNUM)
throw std::runtime_error("JSON value is not an integer as expected");
int32_t retval;
if (!ParseInt32(getValStr(), &retval))
throw std::runtime_error("JSON integer out of range");
return retval;
}
int64_t UniValue::get_int64() const
{
if (typ != VNUM)
throw std::runtime_error("JSON value is not an integer as expected");
int64_t retval;
if (!ParseInt64(getValStr(), &retval))
throw std::runtime_error("JSON integer out of range");
return retval;
}
double UniValue::get_real() const
{
if (typ != VNUM)
throw std::runtime_error("JSON value is not a number as expected");
double retval;
if (!ParseDouble(getValStr(), &retval))
throw std::runtime_error("JSON double out of range");
return retval;
}
const UniValue& UniValue::get_obj() const
{
if (typ != VOBJ)
throw std::runtime_error("JSON value is not an object as expected");
return *this;
}
const UniValue& UniValue::get_array() const
{
if (typ != VARR)
throw std::runtime_error("JSON value is not an array as expected");
return *this;
}

147
lib/univalue_get.cpp Normal file
View File

@@ -0,0 +1,147 @@
// Copyright 2014 BitPay Inc.
// Copyright 2015 Bitcoin Core Developers
// Distributed under the MIT software license, see the accompanying
// file COPYING or http://www.opensource.org/licenses/mit-license.php.
#include <stdint.h>
#include <errno.h>
#include <string.h>
#include <stdlib.h>
#include <stdexcept>
#include <vector>
#include <limits>
#include <string>
#include "univalue.h"
namespace
{
static bool ParsePrechecks(const std::string& str)
{
if (str.empty()) // No empty string allowed
return false;
if (str.size() >= 1 && (json_isspace(str[0]) || json_isspace(str[str.size()-1]))) // No padding allowed
return false;
if (str.size() != strlen(str.c_str())) // No embedded NUL characters allowed
return false;
return true;
}
bool ParseInt32(const std::string& str, int32_t *out)
{
if (!ParsePrechecks(str))
return false;
char *endp = NULL;
errno = 0; // strtol will not set errno if valid
long int n = strtol(str.c_str(), &endp, 10);
if(out) *out = (int32_t)n;
// Note that strtol returns a *long int*, so even if strtol doesn't report a over/underflow
// we still have to check that the returned value is within the range of an *int32_t*. On 64-bit
// platforms the size of these types may be different.
return endp && *endp == 0 && !errno &&
n >= std::numeric_limits<int32_t>::min() &&
n <= std::numeric_limits<int32_t>::max();
}
bool ParseInt64(const std::string& str, int64_t *out)
{
if (!ParsePrechecks(str))
return false;
char *endp = NULL;
errno = 0; // strtoll will not set errno if valid
long long int n = strtoll(str.c_str(), &endp, 10);
if(out) *out = (int64_t)n;
// Note that strtoll returns a *long long int*, so even if strtol doesn't report a over/underflow
// we still have to check that the returned value is within the range of an *int64_t*.
return endp && *endp == 0 && !errno &&
n >= std::numeric_limits<int64_t>::min() &&
n <= std::numeric_limits<int64_t>::max();
}
bool ParseDouble(const std::string& str, double *out)
{
if (!ParsePrechecks(str))
return false;
if (str.size() >= 2 && str[0] == '0' && str[1] == 'x') // No hexadecimal floats allowed
return false;
std::istringstream text(str);
text.imbue(std::locale::classic());
double result;
text >> result;
if(out) *out = result;
return text.eof() && !text.fail();
}
}
const std::vector<std::string>& UniValue::getKeys() const
{
if (typ != VOBJ)
throw std::runtime_error("JSON value is not an object as expected");
return keys;
}
const std::vector<UniValue>& UniValue::getValues() const
{
if (typ != VOBJ && typ != VARR)
throw std::runtime_error("JSON value is not an object or array as expected");
return values;
}
bool UniValue::get_bool() const
{
if (typ != VBOOL)
throw std::runtime_error("JSON value is not a boolean as expected");
return getBool();
}
const std::string& UniValue::get_str() const
{
if (typ != VSTR)
throw std::runtime_error("JSON value is not a string as expected");
return getValStr();
}
int UniValue::get_int() const
{
if (typ != VNUM)
throw std::runtime_error("JSON value is not an integer as expected");
int32_t retval;
if (!ParseInt32(getValStr(), &retval))
throw std::runtime_error("JSON integer out of range");
return retval;
}
int64_t UniValue::get_int64() const
{
if (typ != VNUM)
throw std::runtime_error("JSON value is not an integer as expected");
int64_t retval;
if (!ParseInt64(getValStr(), &retval))
throw std::runtime_error("JSON integer out of range");
return retval;
}
double UniValue::get_real() const
{
if (typ != VNUM)
throw std::runtime_error("JSON value is not a number as expected");
double retval;
if (!ParseDouble(getValStr(), &retval))
throw std::runtime_error("JSON double out of range");
return retval;
}
const UniValue& UniValue::get_obj() const
{
if (typ != VOBJ)
throw std::runtime_error("JSON value is not an object as expected");
return *this;
}
const UniValue& UniValue::get_array() const
{
if (typ != VARR)
throw std::runtime_error("JSON value is not an array as expected");
return *this;
}

View File

@@ -43,21 +43,21 @@ static const char *hatoui(const char *first, const char *last,
}
enum jtokentype getJsonToken(string& tokenVal, unsigned int& consumed,
const char *raw)
const char *raw, const char *end)
{
tokenVal.clear();
consumed = 0;
const char *rawStart = raw;
while ((*raw) && (json_isspace(*raw))) // skip whitespace
while (raw < end && (json_isspace(*raw))) // skip whitespace
raw++;
switch (*raw) {
case 0:
if (raw >= end)
return JTOK_NONE;
switch (*raw) {
case '{':
raw++;
consumed = (raw - rawStart);
@@ -127,40 +127,40 @@ enum jtokentype getJsonToken(string& tokenVal, unsigned int& consumed,
numStr += *raw; // copy first char
raw++;
if ((*first == '-') && (!json_isdigit(*raw)))
if ((*first == '-') && (raw < end) && (!json_isdigit(*raw)))
return JTOK_ERR;
while ((*raw) && json_isdigit(*raw)) { // copy digits
while (raw < end && json_isdigit(*raw)) { // copy digits
numStr += *raw;
raw++;
}
// part 2: frac
if (*raw == '.') {
if (raw < end && *raw == '.') {
numStr += *raw; // copy .
raw++;
if (!json_isdigit(*raw))
if (raw >= end || !json_isdigit(*raw))
return JTOK_ERR;
while ((*raw) && json_isdigit(*raw)) { // copy digits
while (raw < end && json_isdigit(*raw)) { // copy digits
numStr += *raw;
raw++;
}
}
// part 3: exp
if (*raw == 'e' || *raw == 'E') {
if (raw < end && (*raw == 'e' || *raw == 'E')) {
numStr += *raw; // copy E
raw++;
if (*raw == '-' || *raw == '+') { // copy +/-
if (raw < end && (*raw == '-' || *raw == '+')) { // copy +/-
numStr += *raw;
raw++;
}
if (!json_isdigit(*raw))
if (raw >= end || !json_isdigit(*raw))
return JTOK_ERR;
while ((*raw) && json_isdigit(*raw)) { // copy digits
while (raw < end && json_isdigit(*raw)) { // copy digits
numStr += *raw;
raw++;
}
@@ -177,13 +177,16 @@ enum jtokentype getJsonToken(string& tokenVal, unsigned int& consumed,
string valStr;
JSONUTF8StringFilter writer(valStr);
while (*raw) {
if ((unsigned char)*raw < 0x20)
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;
@@ -196,7 +199,8 @@ enum jtokentype getJsonToken(string& tokenVal, unsigned int& consumed,
case 'u': {
unsigned int codepoint;
if (hatoui(raw + 1, raw + 1 + 4, codepoint) !=
if (raw + 1 + 4 >= end ||
hatoui(raw + 1, raw + 1 + 4, codepoint) !=
raw + 1 + 4)
return JTOK_ERR;
writer.push_back_u(codepoint);
@@ -246,7 +250,7 @@ enum expect_bits {
#define setExpect(bit) (expectMask |= EXP_##bit)
#define clearExpect(bit) (expectMask &= ~EXP_##bit)
bool UniValue::read(const char *raw)
bool UniValue::read(const char *raw, size_t size)
{
clear();
@@ -257,10 +261,11 @@ bool UniValue::read(const char *raw)
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);
tok = getJsonToken(tokenVal, consumed, raw, end);
if (tok == JTOK_NONE || tok == JTOK_ERR)
return false;
raw += consumed;
@@ -371,9 +376,6 @@ bool UniValue::read(const char *raw)
case JTOK_KW_NULL:
case JTOK_KW_TRUE:
case JTOK_KW_FALSE: {
if (!stack.size())
return false;
UniValue tmpVal;
switch (tok) {
case JTOK_KW_NULL:
@@ -388,6 +390,11 @@ bool UniValue::read(const char *raw)
default: /* impossible */ break;
}
if (!stack.size()) {
*this = tmpVal;
break;
}
UniValue *top = stack.back();
top->values.push_back(tmpVal);
@@ -396,10 +403,12 @@ bool UniValue::read(const char *raw)
}
case JTOK_NUMBER: {
if (!stack.size())
return false;
UniValue tmpVal(VNUM, tokenVal);
if (!stack.size()) {
*this = tmpVal;
break;
}
UniValue *top = stack.back();
top->values.push_back(tmpVal);
@@ -408,17 +417,18 @@ bool UniValue::read(const char *raw)
}
case JTOK_STRING: {
if (!stack.size())
return false;
UniValue *top = stack.back();
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);
}
@@ -432,7 +442,7 @@ bool UniValue::read(const char *raw)
} while (!stack.empty ());
/* Check that nothing follows the initial construct (parsed above). */
tok = getJsonToken(tokenVal, consumed, raw);
tok = getJsonToken(tokenVal, consumed, raw, end);
if (tok != JTOK_NONE)
return false;

View File

@@ -13,7 +13,7 @@
class JSONUTF8StringFilter
{
public:
JSONUTF8StringFilter(std::string &s):
explicit JSONUTF8StringFilter(std::string &s):
str(s), is_valid(true), codepoint(0), state(0), surpair(0)
{
}
@@ -46,19 +46,19 @@ public:
}
}
// Write codepoint directly, possibly collating surrogate pairs
void push_back_u(unsigned int codepoint)
void push_back_u(unsigned int codepoint_)
{
if (state) // Only accept full codepoints in open state
is_valid = false;
if (codepoint >= 0xD800 && codepoint < 0xDC00) { // First half of surrogate pair
if (codepoint_ >= 0xD800 && codepoint_ < 0xDC00) { // First half of surrogate pair
if (surpair) // Two subsequent surrogate pair openers - fail
is_valid = false;
else
surpair = codepoint;
} else if (codepoint >= 0xDC00 && codepoint < 0xE000) { // Second half of surrogate pair
surpair = codepoint_;
} else if (codepoint_ >= 0xDC00 && codepoint_ < 0xE000) { // Second half of surrogate pair
if (surpair) { // Open surrogate pair, expect second half
// Compute code point from UTF-16 surrogate pair
append_codepoint(0x10000 | ((surpair - 0xD800)<<10) | (codepoint - 0xDC00));
append_codepoint(0x10000 | ((surpair - 0xD800)<<10) | (codepoint_ - 0xDC00));
surpair = 0;
} else // Second half doesn't follow a first half - fail
is_valid = false;
@@ -66,7 +66,7 @@ public:
if (surpair) // First half of surrogate pair not followed by second - fail
is_valid = false;
else
append_codepoint(codepoint);
append_codepoint(codepoint_);
}
}
// Check that we're in a state where the string can be ended
@@ -96,22 +96,22 @@ private:
// Two subsequent \u.... may have to be replaced with one actual codepoint.
unsigned int surpair; // First half of open UTF-16 surrogate pair, or 0
void append_codepoint(unsigned int codepoint)
void append_codepoint(unsigned int codepoint_)
{
if (codepoint <= 0x7f)
str.push_back((char)codepoint);
else if (codepoint <= 0x7FF) {
str.push_back((char)(0xC0 | (codepoint >> 6)));
str.push_back((char)(0x80 | (codepoint & 0x3F)));
} else if (codepoint <= 0xFFFF) {
str.push_back((char)(0xE0 | (codepoint >> 12)));
str.push_back((char)(0x80 | ((codepoint >> 6) & 0x3F)));
str.push_back((char)(0x80 | (codepoint & 0x3F)));
} else if (codepoint <= 0x1FFFFF) {
str.push_back((char)(0xF0 | (codepoint >> 18)));
str.push_back((char)(0x80 | ((codepoint >> 12) & 0x3F)));
str.push_back((char)(0x80 | ((codepoint >> 6) & 0x3F)));
str.push_back((char)(0x80 | (codepoint & 0x3F)));
if (codepoint_ <= 0x7f)
str.push_back((char)codepoint_);
else if (codepoint_ <= 0x7FF) {
str.push_back((char)(0xC0 | (codepoint_ >> 6)));
str.push_back((char)(0x80 | (codepoint_ & 0x3F)));
} else if (codepoint_ <= 0xFFFF) {
str.push_back((char)(0xE0 | (codepoint_ >> 12)));
str.push_back((char)(0x80 | ((codepoint_ >> 6) & 0x3F)));
str.push_back((char)(0x80 | (codepoint_ & 0x3F)));
} else if (codepoint_ <= 0x1FFFFF) {
str.push_back((char)(0xF0 | (codepoint_ >> 18)));
str.push_back((char)(0x80 | ((codepoint_ >> 12) & 0x3F)));
str.push_back((char)(0x80 | ((codepoint_ >> 6) & 0x3F)));
str.push_back((char)(0x80 | (codepoint_ & 0x3F)));
}
}
};

View File

@@ -79,8 +79,6 @@ void UniValue::writeArray(unsigned int prettyIndent, unsigned int indentLevel, s
s += values[i].write(prettyIndent, indentLevel + 1);
if (i != (values.size() - 1)) {
s += ",";
if (prettyIndent)
s += " ";
}
if (prettyIndent)
s += "\n";