lukechilds
/
ethminer
mirror of https://github.com/lukechilds/ethminer.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
Browse Source

Remove serpent.

cl-refactor
Paweł Bylica 9 years ago
parent
561d0fa0d5
commit
58800610f4
37 changed files with 2 additions and 4423 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Unified View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 7
      CMakeLists.txt
  2. 9
      libserpent/All.h
  3. 22
      libserpent/CMakeLists.txt
  4. 112
      libserpent/bignum.cpp
  5. 41
      libserpent/bignum.h
  6. 510
      libserpent/compiler.cpp
  7. 40
      libserpent/compiler.h
  8. 35
      libserpent/funcs.cpp
  9. 35
      libserpent/funcs.h
  10. 203
      libserpent/functions.cpp
  11. 39
      libserpent/functions.h
  12. 70
      libserpent/lllparser.cpp
  13. 13
      libserpent/lllparser.h
  14. 154
      libserpent/opcodes.cpp
  15. 45
      libserpent/opcodes.h
  16. 98
      libserpent/optimize.cpp
  17. 19
      libserpent/optimize.h
  18. 437
      libserpent/parser.cpp
  19. 13
      libserpent/parser.h
  20. 327
      libserpent/preprocess.cpp
  21. 50
      libserpent/preprocess.h
  22. 905
      libserpent/rewriter.cpp
  23. 16
      libserpent/rewriter.h
  24. 212
      libserpent/rewriteutils.cpp
  25. 76
      libserpent/rewriteutils.h
  26. 115
      libserpent/tokenize.cpp
  27. 16
      libserpent/tokenize.h
  28. 333
      libserpent/util.cpp
  29. 137
      libserpent/util.h
  30. 3
      libweb3jsonrpc/WebThreeStubServerBase.cpp
  31. 12
      pullSerpent.sh
  32. 13
      pysol/MANIFEST.in
  33. 0
      pysol/README.md
  34. 115
      pysol/pysolidity.cpp
  35. 41
      pysol/setup.py
  36. 19
      sc/CMakeLists.txt
  37. 129
      sc/cmdline.cpp

7
CMakeLists.txt
View File

@ -302,7 +302,6 @@ eth_format_option(ETHASHCL)
eth_format_option(ETHASHCUDA) eth_format_option(ETHASHCUDA)
eth_format_option(JSCONSOLE) eth_format_option(JSCONSOLE)
eth_format_option(OLYMPIC) eth_format_option(OLYMPIC)
eth_format_option(SERPENT)
eth_format_option(ETHSTRATUM) eth_format_option(ETHSTRATUM)
if (JSCONSOLE) if (JSCONSOLE)
@ -346,7 +345,6 @@ message("-- OLYMPIC Default to the Olympic network ${OLYMPIC}
message("------------------------------------------------------------- components") message("------------------------------------------------------------- components")
message("-- MINER Build miner ${MINER}") message("-- MINER Build miner ${MINER}")
message("-- TOOLS Build basic tools ${TOOLS}") message("-- TOOLS Build basic tools ${TOOLS}")
message("-- SERPENT Build Serpent language components ${SERPENT}")
message("-- GUI Build GUI components ${GUI}") message("-- GUI Build GUI components ${GUI}")
message("-- TESTS Build tests ${TESTS}") message("-- TESTS Build tests ${TESTS}")
message("-- ETHASHCL Build OpenCL components ${ETHASHCL}") message("-- ETHASHCL Build OpenCL components ${ETHASHCL}")
@ -389,11 +387,6 @@ if (GENERAL)
add_subdirectory(liblll) add_subdirectory(liblll)
endif () endif ()
if (SERPENT)
add_subdirectory(libserpent)
add_subdirectory(sc)
endif ()
if (TOOLS) if (TOOLS)
add_subdirectory(lllc) add_subdirectory(lllc)
endif () endif ()

9
libserpent/All.h
View File

@ -1,9 +0,0 @@
#pragma once
#include "compiler.h"
#include "funcs.h"
#include "lllparser.h"
#include "parser.h"
#include "rewriter.h"
#include "tokenize.h"
#include "util.h"

22
libserpent/CMakeLists.txt
View File

@ -1,22 +0,0 @@
cmake_policy(SET CMP0015 NEW)
set(CMAKE_AUTOMOC OFF)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DSTATICLIB")
aux_source_directory(. SRC_LIST)
include_directories(BEFORE ..)
set(EXECUTABLE serpent)
file(GLOB HEADERS "*.h")
add_library(${EXECUTABLE} ${SRC_LIST} ${HEADERS})
target_link_libraries(${EXECUTABLE} lll)
target_link_libraries(${EXECUTABLE} evmcore)
target_link_libraries(${EXECUTABLE} devcore)
install( TARGETS ${EXECUTABLE} RUNTIME DESTINATION bin ARCHIVE DESTINATION lib LIBRARY DESTINATION lib )
install( FILES ${HEADERS} DESTINATION include/${EXECUTABLE} )

112
libserpent/bignum.cpp
View File

@ -1,112 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "bignum.h"
//Integer to string conversion
std::string unsignedToDecimal(unsigned branch) {
if (branch < 10) return nums.substr(branch, 1);
else return unsignedToDecimal(branch / 10) + nums.substr(branch % 10,1);
}
//Add two strings representing decimal values
std::string decimalAdd(std::string a, std::string b) {
std::string o = a;
while (b.length() < a.length()) b = "0" + b;
while (o.length() < b.length()) o = "0" + o;
bool carry = false;
for (int i = o.length() - 1; i >= 0; i--) {
o[i] = o[i] + b[i] - '0';
if (carry) o[i]++;
if (o[i] > '9') {
o[i] -= 10;
carry = true;
}
else carry = false;
}
if (carry) o = "1" + o;
return o;
}
//Helper function for decimalMul
std::string decimalDigitMul(std::string a, int dig) {
if (dig == 0) return "0";
else return decimalAdd(a, decimalDigitMul(a, dig - 1));
}
//Multiply two strings representing decimal values
std::string decimalMul(std::string a, std::string b) {
std::string o = "0";
for (unsigned i = 0; i < b.length(); i++) {
std::string n = decimalDigitMul(a, b[i] - '0');
if (n != "0") {
for (unsigned j = i + 1; j < b.length(); j++) n += "0";
}
o = decimalAdd(o, n);
}
return o;
}
//Modexp
std::string decimalModExp(std::string b, std::string e, std::string m) {
if (e == "0") return "1";
else if (e == "1") return b;
else if (decimalMod(e, "2") == "0") {
std::string o = decimalModExp(b, decimalDiv(e, "2"), m);
return decimalMod(decimalMul(o, o), m);
}
else {
std::string o = decimalModExp(b, decimalDiv(e, "2"), m);
return decimalMod(decimalMul(decimalMul(o, o), b), m);
}
}
//Is a greater than b? Flag allows equality
bool decimalGt(std::string a, std::string b, bool eqAllowed) {
if (a == b) return eqAllowed;
return (a.length() > b.length()) || (a.length() >= b.length() && a > b);
}
//Subtract the two strings representing decimal values
std::string decimalSub(std::string a, std::string b) {
if (b == "0") return a;
if (b == a) return "0";
while (b.length() < a.length()) b = "0" + b;
std::string c = b;
for (unsigned i = 0; i < c.length(); i++) c[i] = '0' + ('9' - c[i]);
std::string o = decimalAdd(decimalAdd(a, c).substr(1), "1");
while (o.size() > 1 && o[0] == '0') o = o.substr(1);
return o;
}
//Divide the two strings representing decimal values
std::string decimalDiv(std::string a, std::string b) {
std::string c = b;
if (decimalGt(c, a)) return "0";
int zeroes = -1;
while (decimalGt(a, c, true)) {
zeroes += 1;
c = c + "0";
}
c = c.substr(0, c.size() - 1);
std::string quot = "0";
while (decimalGt(a, c, true)) {
a = decimalSub(a, c);
quot = decimalAdd(quot, "1");
}
for (int i = 0; i < zeroes; i++) quot += "0";
return decimalAdd(quot, decimalDiv(a, b));
}
//Modulo the two strings representing decimal values
std::string decimalMod(std::string a, std::string b) {
return decimalSub(a, decimalMul(decimalDiv(a, b), b));
}
//String to int conversion
unsigned decimalToUnsigned(std::string a) {
if (a.size() == 0) return 0;
else return (a[a.size() - 1] - '0')
+ decimalToUnsigned(a.substr(0,a.size()-1)) * 10;
}

41
libserpent/bignum.h
View File

@ -1,41 +0,0 @@
#ifndef ETHSERP_BIGNUM
#define ETHSERP_BIGNUM
const std::string nums = "0123456789";
const std::string tt256 =
"115792089237316195423570985008687907853269984665640564039457584007913129639936"
;
const std::string tt256m1 =
"115792089237316195423570985008687907853269984665640564039457584007913129639935"
;
const std::string tt255 =
"57896044618658097711785492504343953926634992332820282019728792003956564819968";
const std::string tt176 =
"95780971304118053647396689196894323976171195136475136";
std::string unsignedToDecimal(unsigned branch);
std::string decimalAdd(std::string a, std::string b);
std::string decimalMul(std::string a, std::string b);
std::string decimalSub(std::string a, std::string b);
std::string decimalDiv(std::string a, std::string b);
std::string decimalMod(std::string a, std::string b);
std::string decimalModExp(std::string b, std::string e, std::string m);
bool decimalGt(std::string a, std::string b, bool eqAllowed=false);
unsigned decimalToUnsigned(std::string a);
#define utd unsignedToDecimal
#define dtu decimalToUnsigned
#endif

510
libserpent/compiler.cpp
View File

@ -1,510 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
#include "bignum.h"
#include "opcodes.h"
// Auxiliary data that is gathered while compiling
struct programAux {
std::map<std::string, std::string> vars;
int nextVarMem;
bool allocUsed;
bool calldataUsed;
int step;
int labelLength;
};
// Auxiliary data that gets passed down vertically
// but not back up
struct programVerticalAux {
int height;
std::string innerScopeName;
std::map<std::string, int> dupvars;
std::map<std::string, int> funvars;
std::vector<mss> scopes;
};
// Compilation result
struct programData {
programAux aux;
Node code;
int outs;
};
programAux Aux() {
programAux o;
o.allocUsed = false;
o.calldataUsed = false;
o.step = 0;
o.nextVarMem = 32;
return o;
}
programVerticalAux verticalAux() {
programVerticalAux o;
o.height = 0;
o.dupvars = std::map<std::string, int>();
o.funvars = std::map<std::string, int>();
o.scopes = std::vector<mss>();
return o;
}
programData pd(programAux aux = Aux(), Node code=token("_"), int outs=0) {
programData o;
o.aux = aux;
o.code = code;
o.outs = outs;
return o;
}
Node multiToken(Node nodes[], int len, Metadata met) {
std::vector<Node> out;
for (int i = 0; i < len; i++) {
out.push_back(nodes[i]);
}
return astnode("_", out, met);
}
Node finalize(programData c);
Node popwrap(Node node) {
Node nodelist[] = {
node,
token("POP", node.metadata)
};
return multiToken(nodelist, 2, node.metadata);
}
// Turns LLL tree into tree of code fragments
programData opcodeify(Node node,
programAux aux=Aux(),
programVerticalAux vaux=verticalAux()) {
std::string symb = "_"+mkUniqueToken();
Metadata m = node.metadata;
// Numbers
if (node.type == TOKEN) {
return pd(aux, nodeToNumeric(node), 1);
}
else if (node.val == "ref" || node.val == "get" || node.val == "set") {
std::string varname = node.args[0].val;
// Determine reference to variable
if (!aux.vars.count(node.args[0].val)) {
aux.vars[node.args[0].val] = utd(aux.nextVarMem);
aux.nextVarMem += 32;
}
Node varNode = tkn(aux.vars[varname], m);
//std::cerr << varname << " " << printSimple(varNode) << "\n";
// Set variable
if (node.val == "set") {
programData sub = opcodeify(node.args[1], aux, vaux);
if (!sub.outs)
err("Value to set variable must have nonzero arity!", m);
// What if we are setting a stack variable?
if (vaux.dupvars.count(node.args[0].val)) {
int h = vaux.height - vaux.dupvars[node.args[0].val];
if (h > 16) err("Too deep for stack variable (max 16)", m);
Node nodelist[] = {
sub.code,
token("SWAP"+unsignedToDecimal(h), m),
token("POP", m)
};
return pd(sub.aux, multiToken(nodelist, 3, m), 0);
}
// Setting a memory variable
else {
Node nodelist[] = {
sub.code,
varNode,
token("MSTORE", m),
};
return pd(sub.aux, multiToken(nodelist, 3, m), 0);
}
}
// Get variable
else if (node.val == "get") {
// Getting a stack variable
if (vaux.dupvars.count(node.args[0].val)) {
int h = vaux.height - vaux.dupvars[node.args[0].val];
if (h > 16) err("Too deep for stack variable (max 16)", m);
return pd(aux, token("DUP"+unsignedToDecimal(h)), 1);
}
// Getting a memory variable
else {
Node nodelist[] =
{ varNode, token("MLOAD", m) };
return pd(aux, multiToken(nodelist, 2, m), 1);
}
}
// Refer variable
else if (node.val == "ref") {
if (vaux.dupvars.count(node.args[0].val))
err("Cannot ref stack variable!", m);
return pd(aux, varNode, 1);
}
}
// Comments do nothing
else if (node.val == "comment") {
return pd(aux, astnode("_", m), 0);
}
// Custom operation sequence
// eg. (ops bytez id msize swap1 msize add 0 swap1 mstore) == alloc
if (node.val == "ops") {
std::vector<Node> subs2;
int depth = 0;
for (unsigned i = 0; i < node.args.size(); i++) {
std::string op = upperCase(node.args[i].val);
if (node.args[i].type == ASTNODE || opinputs(op) == -1) {
programVerticalAux vaux2 = vaux;
vaux2.height = vaux.height - i - 1 + node.args.size();
programData sub = opcodeify(node.args[i], aux, vaux2);
aux = sub.aux;
depth += sub.outs;
subs2.push_back(sub.code);
}
else {
subs2.push_back(token(op, m));
depth += opoutputs(op) - opinputs(op);
}
}
if (depth < 0 || depth > 1) err("Stack depth mismatch", m);
return pd(aux, astnode("_", subs2, m), 0);
}
// Code blocks
if (node.val == "lll" && node.args.size() == 2) {
if (node.args[1].val != "0") aux.allocUsed = true;
std::vector<Node> o;
o.push_back(finalize(opcodeify(node.args[0])));
programData sub = opcodeify(node.args[1], aux, vaux);
Node code = astnode("____CODE", o, m);
Node nodelist[] = {
token("$begincode"+symb+".endcode"+symb, m), token("DUP1", m),
token("$begincode"+symb, m), sub.code, token("CODECOPY", m),
token("$endcode"+symb, m), token("JUMP", m),
token("~begincode"+symb, m), code,
token("~endcode"+symb, m), token("JUMPDEST", m)
};
return pd(sub.aux, multiToken(nodelist, 11, m), 1);
}
// Stack variables
if (node.val == "with") {
programData initial = opcodeify(node.args[1], aux, vaux);
programVerticalAux vaux2 = vaux;
vaux2.dupvars[node.args[0].val] = vaux.height;
vaux2.height += 1;
if (!initial.outs)
err("Initial variable value must have nonzero arity!", m);
programData sub = opcodeify(node.args[2], initial.aux, vaux2);
Node nodelist[] = {
initial.code,
sub.code
};
programData o = pd(sub.aux, multiToken(nodelist, 2, m), sub.outs);
if (sub.outs)
o.code.args.push_back(token("SWAP1", m));
o.code.args.push_back(token("POP", m));
return o;
}
// Seq of multiple statements
if (node.val == "seq") {
std::vector<Node> children;
int lastOut = 0;
for (unsigned i = 0; i < node.args.size(); i++) {
programData sub = opcodeify(node.args[i], aux, vaux);
aux = sub.aux;
if (sub.outs == 1) {
if (i < node.args.size() - 1) sub.code = popwrap(sub.code);
else lastOut = 1;
}
children.push_back(sub.code);
}
return pd(aux, astnode("_", children, m), lastOut);
}
// 2-part conditional (if gets rewritten to unless in rewrites)
else if (node.val == "unless" && node.args.size() == 2) {
programData cond = opcodeify(node.args[0], aux, vaux);
programData action = opcodeify(node.args[1], cond.aux, vaux);
aux = action.aux;
if (!cond.outs) err("Condition of if/unless statement has arity 0", m);
if (action.outs) action.code = popwrap(action.code);
Node nodelist[] = {
cond.code,
token("$endif"+symb, m), token("JUMPI", m),
action.code,
token("~endif"+symb, m), token("JUMPDEST", m)
};
return pd(aux, multiToken(nodelist, 6, m), 0);
}
// 3-part conditional
else if (node.val == "if" && node.args.size() == 3) {
programData ifd = opcodeify(node.args[0], aux, vaux);
programData thend = opcodeify(node.args[1], ifd.aux, vaux);
programData elsed = opcodeify(node.args[2], thend.aux, vaux);
aux = elsed.aux;
if (!ifd.outs)
err("Condition of if/unless statement has arity 0", m);
// Handle cases where one conditional outputs something
// and the other does not
int outs = (thend.outs && elsed.outs) ? 1 : 0;
if (thend.outs > outs) thend.code = popwrap(thend.code);
if (elsed.outs > outs) elsed.code = popwrap(elsed.code);
Node nodelist[] = {
ifd.code,
token("ISZERO", m),
token("$else"+symb, m), token("JUMPI", m),
thend.code,
token("$endif"+symb, m), token("JUMP", m),
token("~else"+symb, m), token("JUMPDEST", m),
elsed.code,
token("~endif"+symb, m), token("JUMPDEST", m)
};
return pd(aux, multiToken(nodelist, 12, m), outs);
}
// While (rewritten to this in rewrites)
else if (node.val == "until") {
programData cond = opcodeify(node.args[0], aux, vaux);
programData action = opcodeify(node.args[1], cond.aux, vaux);
aux = action.aux;
if (!cond.outs)
err("Condition of while/until loop has arity 0", m);
if (action.outs) action.code = popwrap(action.code);
Node nodelist[] = {
token("~beg"+symb, m), token("JUMPDEST", m),
cond.code,
token("$end"+symb, m), token("JUMPI", m),
action.code,
token("$beg"+symb, m), token("JUMP", m),
token("~end"+symb, m), token("JUMPDEST", m),
};
return pd(aux, multiToken(nodelist, 10, m));
}
// Memory allocations
else if (node.val == "alloc") {
programData bytez = opcodeify(node.args[0], aux, vaux);
aux = bytez.aux;
if (!bytez.outs)
err("Alloc input has arity 0", m);
aux.allocUsed = true;
Node nodelist[] = {
bytez.code,
token("MSIZE", m), token("SWAP1", m), token("MSIZE", m),
token("ADD", m),
token("0", m), token("SWAP1", m), token("MSTORE", m)
};
return pd(aux, multiToken(nodelist, 8, m), 1);
}
// All other functions/operators
else {
std::vector<Node> subs2;
int depth = opinputs(upperCase(node.val));
if (depth == -1)
err("Not a function or opcode: "+node.val, m);
if ((int)node.args.size() != depth)
err("Invalid arity for "+node.val, m);
for (int i = node.args.size() - 1; i >= 0; i--) {
programVerticalAux vaux2 = vaux;
vaux2.height = vaux.height - i - 1 + node.args.size();
programData sub = opcodeify(node.args[i], aux, vaux2);
aux = sub.aux;
if (!sub.outs)
err("Input "+unsignedToDecimal(i)+" has arity 0", sub.code.metadata);
subs2.push_back(sub.code);
}
subs2.push_back(token(upperCase(node.val), m));
int outdepth = opoutputs(upperCase(node.val));
return pd(aux, astnode("_", subs2, m), outdepth);
}
}
// Adds necessary wrappers to a program
Node finalize(programData c) {
std::vector<Node> bottom;
Metadata m = c.code.metadata;
// If we are using both alloc and variables, we need to pre-zfill
// some memory
if ((c.aux.allocUsed || c.aux.calldataUsed) && c.aux.vars.size() > 0) {
Node nodelist[] = {
token("0", m),
token(unsignedToDecimal(c.aux.nextVarMem - 1)),
token("MSTORE8", m)
};
bottom.push_back(multiToken(nodelist, 3, m));
}
// The actual code
bottom.push_back(c.code);
return astnode("_", bottom, m);
}
//LLL -> code fragment tree
Node buildFragmentTree(Node node) {
return finalize(opcodeify(node));
}
// Builds a dictionary mapping labels to variable names
void buildDict(Node program, programAux &aux, int labelLength) {
Metadata m = program.metadata;
// Token
if (program.type == TOKEN) {
if (isNumberLike(program)) {
aux.step += 1 + toByteArr(program.val, m).size();
}
else if (program.val[0] == '~') {
aux.vars[program.val.substr(1)] = unsignedToDecimal(aux.step);
}
else if (program.val[0] == '$') {
aux.step += labelLength + 1;
}
else aux.step += 1;
}
// A sub-program (ie. LLL)
else if (program.val == "____CODE") {
int step = aux.step;
aux.step = 0;
for (unsigned i = 0; i < program.args.size(); i++) {
buildDict(program.args[i], aux, labelLength);
}
aux.step += step;
}
// Normal sub-block
else {
for (unsigned i = 0; i < program.args.size(); i++) {
buildDict(program.args[i], aux, labelLength);
}
}
}
// Applies that dictionary
void substDict(Node program, programAux aux, int labelLength, std::vector<Node> &out) {
Metadata m = program.metadata;
std::vector<Node> inner;
if (program.type == TOKEN) {
if (program.val[0] == '$') {
std::string tokStr = "PUSH"+unsignedToDecimal(labelLength);
out.push_back(token(tokStr, m));
int dotLoc = program.val.find('.');
if (dotLoc == -1) {
std::string val = aux.vars[program.val.substr(1)];
inner = toByteArr(val, m, labelLength);
}
else {
std::string start = aux.vars[program.val.substr(1, dotLoc-1)],
end = aux.vars[program.val.substr(dotLoc + 1)],
dist = decimalSub(end, start);
inner = toByteArr(dist, m, labelLength);
}
for (unsigned i = 0; i < inner.size(); i++) out.push_back(inner[i]);
}
else if (program.val[0] == '~') { }
else if (isNumberLike(program)) {
inner = toByteArr(program.val, m);
out.push_back(token("PUSH"+unsignedToDecimal(inner.size())));
for (unsigned i = 0; i < inner.size(); i++) out.push_back(inner[i]);
}
else out.push_back(program);
}
else {
for (unsigned i = 0; i < program.args.size(); i++) {
substDict(program.args[i], aux, labelLength, out);
}
}
}
// Compiled fragtree -> compiled fragtree without labels
std::vector<Node> dereference(Node program) {
int sz = treeSize(program) * 4;
int labelLength = 1;
while (sz >= 256) { labelLength += 1; sz /= 256; }
programAux aux = Aux();
buildDict(program, aux, labelLength);
std::vector<Node> o;
substDict(program, aux, labelLength, o);
return o;
}
// Opcodes -> bin
std::string serialize(std::vector<Node> codons) {
std::string o;
for (unsigned i = 0; i < codons.size(); i++) {
int v;
if (isNumberLike(codons[i])) {
v = decimalToUnsigned(codons[i].val);
}
else if (codons[i].val.substr(0,4) == "PUSH") {
v = 95 + decimalToUnsigned(codons[i].val.substr(4));
}
else {
v = opcode(codons[i].val);
}
o += (char)v;
}
return o;
}
// Bin -> opcodes
std::vector<Node> deserialize(std::string ser) {
std::vector<Node> o;
int backCount = 0;
for (unsigned i = 0; i < ser.length(); i++) {
unsigned char v = (unsigned char)ser[i];
std::string oper = op((int)v);
if (oper != "" && backCount <= 0) o.push_back(token(oper));
else if (v >= 96 && v < 128 && backCount <= 0) {
o.push_back(token("PUSH"+unsignedToDecimal(v - 95)));
}
else o.push_back(token(unsignedToDecimal(v)));
if (v >= 96 && v < 128 && backCount <= 0) {
backCount = v - 95;
}
else backCount--;
}
return o;
}
// Fragtree -> bin
std::string assemble(Node fragTree) {
return serialize(dereference(fragTree));
}
// Fragtree -> tokens
std::vector<Node> prettyAssemble(Node fragTree) {
return dereference(fragTree);
}
// LLL -> bin
std::string compileLLL(Node program) {
return assemble(buildFragmentTree(program));
}
// LLL -> tokens
std::vector<Node> prettyCompileLLL(Node program) {
return prettyAssemble(buildFragmentTree(program));
}
// Converts a list of integer values to binary transaction data
std::string encodeDatalist(std::vector<std::string> vals) {
std::string o;
for (unsigned i = 0; i < vals.size(); i++) {
std::vector<Node> n = toByteArr(strToNumeric(vals[i]), Metadata(), 32);
for (unsigned j = 0; j < n.size(); j++) {
int v = decimalToUnsigned(n[j].val);
o += (char)v;
}
}
return o;
}
// Converts binary transaction data into a list of integer values
std::vector<std::string> decodeDatalist(std::string ser) {
std::vector<std::string> out;
for (unsigned i = 0; i < ser.length(); i+= 32) {
std::string o = "0";
for (unsigned j = i; j < i + 32; j++) {
int vj = (int)(unsigned char)ser[j];
o = decimalAdd(decimalMul(o, "256"), unsignedToDecimal(vj));
}
out.push_back(o);
}
return out;
}

40
libserpent/compiler.h
View File

@ -1,40 +0,0 @@
#ifndef ETHSERP_COMPILER
#define ETHSERP_COMPILER
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
// Compiled fragtree -> compiled fragtree without labels
std::vector<Node> dereference(Node program);
// LLL -> fragtree
Node buildFragmentTree(Node program);
// opcodes -> bin
std::string serialize(std::vector<Node> codons);
// Fragtree -> bin
std::string assemble(Node fragTree);
// Fragtree -> opcodes
std::vector<Node> prettyAssemble(Node fragTree);
// LLL -> bin
std::string compileLLL(Node program);
// LLL -> opcodes
std::vector<Node> prettyCompileLLL(Node program);
// bin -> opcodes
std::vector<Node> deserialize(std::string ser);
// Converts a list of integer values to binary transaction data
std::string encodeDatalist(std::vector<std::string> vals);
// Converts binary transaction data into a list of integer values
std::vector<std::string> decodeDatalist(std::string ser);
#endif

35
libserpent/funcs.cpp
View File

@ -1,35 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include "funcs.h"
#include "bignum.h"
#include "util.h"
#include "parser.h"
#include "lllparser.h"
#include "compiler.h"
#include "rewriter.h"
#include "tokenize.h"
Node compileToLLL(std::string input) {
return rewrite(parseSerpent(input));
}
Node compileChunkToLLL(std::string input) {
return rewriteChunk(parseSerpent(input));
}
std::string compile(std::string input) {
return compileLLL(compileToLLL(input));
}
std::vector<Node> prettyCompile(std::string input) {
return prettyCompileLLL(compileToLLL(input));
}
std::string compileChunk(std::string input) {
return compileLLL(compileChunkToLLL(input));
}
std::vector<Node> prettyCompileChunk(std::string input) {
return prettyCompileLLL(compileChunkToLLL(input));
}

35
libserpent/funcs.h
View File

@ -1,35 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include "bignum.h"
#include "util.h"
#include "parser.h"
#include "lllparser.h"
#include "compiler.h"
#include "rewriter.h"
#include "tokenize.h"
// Function listing:
//
// parseSerpent (serpent -> AST) std::string -> Node
// parseLLL (LLL -> AST) std::string -> Node
// rewrite (apply rewrite rules) Node -> Node
// compileToLLL (serpent -> LLL) std::string -> Node
// compileLLL (LLL -> EVMhex) Node -> std::string
// prettyCompileLLL (LLL -> EVMasm) Node -> std::vector<Node>
// prettyCompile (serpent -> EVMasm) std::string -> std::vector>Node>
// compile (serpent -> EVMhex) std::string -> std::string
// get_file_contents (filename -> file) std::string -> std::string
// exists (does file exist?) std::string -> bool
Node compileToLLL(std::string input);
Node compileChunkToLLL(std::string input);
std::string compile(std::string input);
std::vector<Node> prettyCompile(std::string input);
std::string compileChunk(std::string input);
std::vector<Node> prettyCompileChunk(std::string input);

203
libserpent/functions.cpp
View File

@ -1,203 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
#include "lllparser.h"
#include "bignum.h"
#include "optimize.h"
#include "rewriteutils.h"
#include "preprocess.h"
#include "functions.h"
std::string getSignature(std::vector<Node> args) {
std::string o;
for (unsigned i = 0; i < args.size(); i++) {
if (args[i].val == ":" && args[i].args[1].val == "s")
o += "s";
else if (args[i].val == ":" && args[i].args[1].val == "a")
o += "a";
else
o += "i";
}
return o;
}
// Convert a list of arguments into a node containing a
// < datastart, datasz > pair
Node packArguments(std::vector<Node> args, std::string sig,
int funId, Metadata m) {
// Plain old 32 byte arguments
std::vector<Node> nargs;
// Variable-sized arguments
std::vector<Node> vargs;
// Variable sizes
std::vector<Node> sizes;
// Is a variable an array?
std::vector<bool> isArray;
// Fill up above three argument lists
int argCount = 0;
for (unsigned i = 0; i < args.size(); i++) {
Metadata m = args[i].metadata;
if (args[i].val == "=") {
// do nothing
}
else {
// Determine the correct argument type
char argType;
if (sig.size() > 0) {
if (argCount >= (signed)sig.size())
err("Too many args", m);
argType = sig[argCount];
}
else argType = 'i';
// Integer (also usable for short strings)
if (argType == 'i') {
if (args[i].val == ":")
err("Function asks for int, provided string or array", m);
nargs.push_back(args[i]);
}
// Long string
else if (argType == 's') {
if (args[i].val != ":")
err("Must specify string length", m);
vargs.push_back(args[i].args[0]);
sizes.push_back(args[i].args[1]);
isArray.push_back(false);
}
// Array
else if (argType == 'a') {
if (args[i].val != ":")
err("Must specify array length", m);
vargs.push_back(args[i].args[0]);
sizes.push_back(args[i].args[1]);
isArray.push_back(true);
}
else err("Invalid arg type in signature", m);
argCount++;
}
}
int static_arg_size = 1 + (vargs.size() + nargs.size()) * 32;
// Start off by saving the size variables and calculating the total
msn kwargs;
kwargs["funid"] = tkn(utd(funId), m);
std::string pattern =
"(with _sztot "+utd(static_arg_size)+" "
" (with _sizes (alloc "+utd(sizes.size() * 32)+") "
" (seq ";
for (unsigned i = 0; i < sizes.size(); i++) {
std::string sizeIncrement =
isArray[i] ? "(mul 32 _x)" : "_x";
pattern +=
"(with _x $sz"+utd(i)+"(seq "
" (mstore (add _sizes "+utd(i * 32)+") _x) "
" (set _sztot (add _sztot "+sizeIncrement+" )))) ";
kwargs["sz"+utd(i)] = sizes[i];
}
// Allocate memory, and set first data byte
pattern +=
"(with _datastart (alloc (add _sztot 32)) (seq "
" (mstore8 _datastart $funid) ";
// Copy over size variables
for (unsigned i = 0; i < sizes.size(); i++) {
int v = 1 + i * 32;
pattern +=
" (mstore "
" (add _datastart "+utd(v)+") "
" (mload (add _sizes "+utd(v-1)+"))) ";
}
// Store normal arguments
for (unsigned i = 0; i < nargs.size(); i++) {
int v = 1 + (i + sizes.size()) * 32;
pattern +=
" (mstore (add _datastart "+utd(v)+") $"+utd(i)+") ";
kwargs[utd(i)] = nargs[i];
}
// Loop through variable-sized arguments, store them
pattern +=
" (with _pos (add _datastart "+utd(static_arg_size)+") (seq";
for (unsigned i = 0; i < vargs.size(); i++) {
std::string copySize =
isArray[i] ? "(mul 32 (mload (add _sizes "+utd(i * 32)+")))"
: "(mload (add _sizes "+utd(i * 32)+"))";
pattern +=
" (unsafe_mcopy _pos $vl"+utd(i)+" "+copySize+") "
" (set _pos (add _pos "+copySize+")) ";
kwargs["vl"+utd(i)] = vargs[i];
}
// Return a 2-item array containing the start and size
pattern += " (array_lit _datastart _sztot))))))))";
std::string prefix = "_temp_"+mkUniqueToken();
// Fill in pattern, return triple
return subst(parseLLL(pattern), kwargs, prefix, m);
}
// Create a node for argument unpacking
Node unpackArguments(std::vector<Node> vars, Metadata m) {
std::vector<std::string> varNames;
std::vector<std::string> longVarNames;
std::vector<bool> longVarIsArray;
// Fill in variable and long variable names, as well as which
// long variables are arrays and which are strings
for (unsigned i = 0; i < vars.size(); i++) {
if (vars[i].val == ":") {
if (vars[i].args.size() != 2)
err("Malformed def!", m);
longVarNames.push_back(vars[i].args[0].val);
std::string tag = vars[i].args[1].val;
if (tag == "s")
longVarIsArray.push_back(false);
else if (tag == "a")
longVarIsArray.push_back(true);
else
err("Function value can only be string or array", m);
}
else {
varNames.push_back(vars[i].val);
}
}
std::vector<Node> sub;
if (!varNames.size() && !longVarNames.size()) {
// do nothing if we have no arguments
}
else {
std::vector<Node> varNodes;
for (unsigned i = 0; i < longVarNames.size(); i++)
varNodes.push_back(token(longVarNames[i], m));
for (unsigned i = 0; i < varNames.size(); i++)
varNodes.push_back(token(varNames[i], m));
// Copy over variable lengths and short variables
for (unsigned i = 0; i < varNodes.size(); i++) {
int pos = 1 + i * 32;
std::string prefix = (i < longVarNames.size()) ? "_len_" : "";
sub.push_back(asn("untyped", asn("set",
token(prefix+varNodes[i].val, m),
asn("calldataload", tkn(utd(pos), m), m),
m)));
}
// Copy over long variables
if (longVarNames.size() > 0) {
std::vector<Node> sub2;
int pos = varNodes.size() * 32 + 1;
Node tot = tkn("_tot", m);
for (unsigned i = 0; i < longVarNames.size(); i++) {
Node var = tkn(longVarNames[i], m);
Node varlen = longVarIsArray[i]
? asn("mul", tkn("32", m), tkn("_len_"+longVarNames[i], m))
: tkn("_len_"+longVarNames[i], m);
sub2.push_back(asn("untyped",
asn("set", var, asn("alloc", varlen))));
sub2.push_back(asn("calldatacopy", var, tot, varlen));
sub2.push_back(asn("set", tot, asn("add", tot, varlen)));
}
std::string prefix = "_temp_"+mkUniqueToken();
sub.push_back(subst(
astnode("with", tot, tkn(utd(pos), m), asn("seq", sub2)),
msn(),
prefix,
m));
}
}
return asn("seq", sub, m);
}

39
libserpent/functions.h
View File

@ -1,39 +0,0 @@
#ifndef ETHSERP_FUNCTIONS
#define ETHSERP_FUNCTIONS
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
#include "lllparser.h"
#include "bignum.h"
#include "optimize.h"
#include "rewriteutils.h"
#include "preprocess.h"
class argPack {
public:
argPack(Node a, Node b, Node c) {
pre = a;
datastart = b;
datasz = c;
}
Node pre;
Node datastart;
Node datasz;
};
// Get a signature from a function
std::string getSignature(std::vector<Node> args);
// Convert a list of arguments into a <pre, mstart, msize> node
// triple, given the signature of a function
Node packArguments(std::vector<Node> args, std::string sig,
int funId, Metadata m);
// Create a node for argument unpacking
Node unpackArguments(std::vector<Node> vars, Metadata m);
#endif

70
libserpent/lllparser.cpp
View File

@ -1,70 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
#include "lllparser.h"
#include "tokenize.h"
struct _parseOutput {
Node node;
int newpos;
};
// Helper, returns subtree and position of start of next node
_parseOutput _parse(std::vector<Node> inp, int pos) {
Metadata met = inp[pos].metadata;
_parseOutput o;
// Bracket: keep grabbing tokens until we get to the
// corresponding closing bracket
if (inp[pos].val == "(" || inp[pos].val == "[") {
std::string fun, rbrack;
std::vector<Node> args;
pos += 1;
if (inp[pos].val == "[") {
fun = "access";
rbrack = "]";
}
else rbrack = ")";
// First argument is the function
while (inp[pos].val != ")") {
_parseOutput po = _parse(inp, pos);
if (fun.length() == 0 && po.node.type == 1) {
std::cerr << "Error: first arg must be function\n";
fun = po.node.val;
}
else if (fun.length() == 0) {
fun = po.node.val;
}
else {
args.push_back(po.node);
}
pos = po.newpos;
}
o.newpos = pos + 1;
o.node = astnode(fun, args, met);
}
// Normal token, return it and advance to next token
else {
o.newpos = pos + 1;
o.node = token(inp[pos].val, met);
}
return o;
}
// stream of tokens -> lisp parse tree
Node parseLLLTokenStream(std::vector<Node> inp) {
_parseOutput o = _parse(inp, 0);
return o.node;
}
// Parses LLL
Node parseLLL(std::string s, bool allowFileRead) {
std::string input = s;
std::string file = "main";
if (exists(s) && allowFileRead) {
file = s;
input = get_file_contents(s);
}
return parseLLLTokenStream(tokenize(s, Metadata(file, 0, 0), true));
}

13
libserpent/lllparser.h
View File

@ -1,13 +0,0 @@
#ifndef ETHSERP_LLLPARSER
#define ETHSERP_LLLPARSER
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
// LLL text -> parse tree
Node parseLLL(std::string s, bool allowFileRead=false);
#endif

154
libserpent/opcodes.cpp
View File

@ -1,154 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "opcodes.h"
#include "util.h"
#include "bignum.h"
Mapping mapping[] = {
Mapping("STOP", 0x00, 0, 0),
Mapping("ADD", 0x01, 2, 1),
Mapping("MUL", 0x02, 2, 1),
Mapping("SUB", 0x03, 2, 1),
Mapping("DIV", 0x04, 2, 1),
Mapping("SDIV", 0x05, 2, 1),
Mapping("MOD", 0x06, 2, 1),
Mapping("SMOD", 0x07, 2, 1),
Mapping("ADDMOD", 0x08, 3, 1),
Mapping("MULMOD", 0x09, 3, 1),
Mapping("EXP", 0x0a, 2, 1),
Mapping("SIGNEXTEND", 0x0b, 2, 1),
Mapping("LT", 0x10, 2, 1),
Mapping("GT", 0x11, 2, 1),
Mapping("SLT", 0x12, 2, 1),
Mapping("SGT", 0x13, 2, 1),
Mapping("EQ", 0x14, 2, 1),
Mapping("ISZERO", 0x15, 1, 1),
Mapping("AND", 0x16, 2, 1),
Mapping("OR", 0x17, 2, 1),
Mapping("XOR", 0x18, 2, 1),
Mapping("NOT", 0x19, 1, 1),
Mapping("BYTE", 0x1a, 2, 1),
Mapping("SHA3", 0x20, 2, 1),
Mapping("ADDRESS", 0x30, 0, 1),
Mapping("BALANCE", 0x31, 1, 1),
Mapping("ORIGIN", 0x32, 0, 1),
Mapping("CALLER", 0x33, 0, 1),
Mapping("CALLVALUE", 0x34, 0, 1),
Mapping("CALLDATALOAD", 0x35, 1, 1),
Mapping("CALLDATASIZE", 0x36, 0, 1),
Mapping("CALLDATACOPY", 0x37, 3, 0),
Mapping("CODESIZE", 0x38, 0, 1),
Mapping("CODECOPY", 0x39, 3, 0),
Mapping("GASPRICE", 0x3a, 0, 1),
Mapping("EXTCODESIZE", 0x3b, 1, 1),
Mapping("EXTCODECOPY", 0x3c, 4, 0),
Mapping("BLOCKHASH", 0x40, 1, 1),
Mapping("COINBASE", 0x41, 0, 1),
Mapping("TIMESTAMP", 0x42, 0, 1),
Mapping("NUMBER", 0x43, 0, 1),
Mapping("DIFFICULTY", 0x44, 0, 1),
Mapping("GASLIMIT", 0x45, 0, 1),
Mapping("POP", 0x50, 1, 0),
Mapping("MLOAD", 0x51, 1, 1),
Mapping("MSTORE", 0x52, 2, 0),
Mapping("MSTORE8", 0x53, 2, 0),
Mapping("SLOAD", 0x54, 1, 1),
Mapping("SSTORE", 0x55, 2, 0),
Mapping("JUMP", 0x56, 1, 0),
Mapping("JUMPI", 0x57, 2, 0),
Mapping("PC", 0x58, 0, 1),
Mapping("MSIZE", 0x59, 0, 1),
Mapping("GAS", 0x5a, 0, 1),
Mapping("JUMPDEST", 0x5b, 0, 0),
Mapping("LOG0", 0xa0, 2, 0),
Mapping("LOG1", 0xa1, 3, 0),
Mapping("LOG2", 0xa2, 4, 0),
Mapping("LOG3", 0xa3, 5, 0),
Mapping("LOG4", 0xa4, 6, 0),
Mapping("CREATE", 0xf0, 3, 1),
Mapping("CALL", 0xf1, 7, 1),
Mapping("CALLCODE", 0xf2, 7, 1),
Mapping("RETURN", 0xf3, 2, 0),
Mapping("SUICIDE", 0xff, 1, 0),
Mapping("---END---", 0x00, 0, 0),
};
std::map<std::string, std::vector<int> > opcodes;
std::map<int, std::string> reverseOpcodes;
// Fetches everything EXCEPT PUSH1..32
std::pair<std::string, std::vector<int> > _opdata(std::string ops, int opi) {
if (!opcodes.size()) {
int i = 0;
while (mapping[i].op != "---END---") {
Mapping mi = mapping[i];
opcodes[mi.op] = triple(mi.opcode, mi.in, mi.out);
i++;
}
for (i = 1; i <= 16; i++) {
opcodes["DUP"+unsignedToDecimal(i)] = triple(0x7f + i, i, i+1);
opcodes["SWAP"+unsignedToDecimal(i)] = triple(0x8f + i, i+1, i+1);
}
for (std::map<std::string, std::vector<int> >::iterator it=opcodes.begin();
it != opcodes.end();
it++) {
reverseOpcodes[(*it).second[0]] = (*it).first;
}
}
ops = upperCase(ops);
std::string op;
std::vector<int> opdata;
op = reverseOpcodes.count(opi) ? reverseOpcodes[opi] : "";
opdata = opcodes.count(ops) ? opcodes[ops] : triple(-1, -1, -1);
return std::pair<std::string, std::vector<int> >(op, opdata);
}
int opcode(std::string op) {
return _opdata(op, -1).second[0];
}
int opinputs(std::string op) {
return _opdata(op, -1).second[1];
}
int opoutputs(std::string op) {
return _opdata(op, -1).second[2];
}
std::string op(int opcode) {
return _opdata("", opcode).first;
}
std::string lllSpecials[][3] = {
{ "ref", "1", "1" },
{ "get", "1", "1" },
{ "set", "2", "2" },
{ "with", "3", "3" },
{ "comment", "0", "2147483647" },
{ "ops", "0", "2147483647" },
{ "lll", "2", "2" },
{ "seq", "0", "2147483647" },
{ "if", "3", "3" },
{ "unless", "2", "2" },
{ "until", "2", "2" },
{ "alloc", "1", "1" },
{ "---END---", "0", "0" },
};
std::map<std::string, std::pair<int, int> > lllMap;
// Is a function name one of the valid functions above?
bool isValidLLLFunc(std::string f, int argc) {
if (lllMap.size() == 0) {
for (int i = 0; ; i++) {
if (lllSpecials[i][0] == "---END---") break;
lllMap[lllSpecials[i][0]] = std::pair<int, int>(
dtu(lllSpecials[i][1]), dtu(lllSpecials[i][2]));
}
}
return lllMap.count(f)
&& argc >= lllMap[f].first
&& argc <= lllMap[f].second;
}

45
libserpent/opcodes.h
View File

@ -1,45 +0,0 @@
#ifndef ETHSERP_OPCODES
#define ETHSERP_OPCODES
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
class Mapping {
public:
Mapping(std::string Op, int Opcode, int In, int Out) {
op = Op;
opcode = Opcode;
in = In;
out = Out;
}
std::string op;
int opcode;
int in;
int out;
};
extern Mapping mapping[];
extern std::map<std::string, std::vector<int> > opcodes;
extern std::map<int, std::string> reverseOpcodes;
std::pair<std::string, std::vector<int> > _opdata(std::string ops, int opi);
int opcode(std::string op);
int opinputs(std::string op);
int opoutputs(std::string op);
std::string op(int opcode);
extern std::string lllSpecials[][3];
extern std::map<std::string, std::pair<int, int> > lllMap;
bool isValidLLLFunc(std::string f, int argc);
#endif

98
libserpent/optimize.cpp
View File

@ -1,98 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
#include "lllparser.h"
#include "bignum.h"
// Compile-time arithmetic calculations
Node optimize(Node inp) {
if (inp.type == TOKEN) {
Node o = tryNumberize(inp);
if (decimalGt(o.val, tt256, true))
err("Value too large (exceeds 32 bytes or 2^256)", inp.metadata);
return o;
}
for (unsigned i = 0; i < inp.args.size(); i++) {
inp.args[i] = optimize(inp.args[i]);
}
// Arithmetic-specific transform
if (inp.val == "+") inp.val = "add";
if (inp.val == "*") inp.val = "mul";
if (inp.val == "-") inp.val = "sub";
if (inp.val == "/") inp.val = "sdiv";
if (inp.val == "^") inp.val = "exp";
if (inp.val == "**") inp.val = "exp";
if (inp.val == "%") inp.val = "smod";
// Degenerate cases for add and mul
if (inp.args.size() == 2) {
if (inp.val == "add" && inp.args[0].type == TOKEN &&
inp.args[0].val == "0") {
Node x = inp.args[1];
inp = x;
}
if (inp.val == "add" && inp.args[1].type == TOKEN &&
inp.args[1].val == "0") {
Node x = inp.args[0];
inp = x;
}
if (inp.val == "mul" && inp.args[0].type == TOKEN &&
inp.args[0].val == "1") {
Node x = inp.args[1];
inp = x;
}
if (inp.val == "mul" && inp.args[1].type == TOKEN &&
inp.args[1].val == "1") {
Node x = inp.args[0];
inp = x;
}
}
// Arithmetic computation
if (inp.args.size() == 2
&& inp.args[0].type == TOKEN
&& inp.args[1].type == TOKEN) {
std::string o;
if (inp.val == "add") {
o = decimalMod(decimalAdd(inp.args[0].val, inp.args[1].val), tt256);
}
else if (inp.val == "sub") {
if (decimalGt(inp.args[0].val, inp.args[1].val, true))
o = decimalSub(inp.args[0].val, inp.args[1].val);
}
else if (inp.val == "mul") {
o = decimalMod(decimalMul(inp.args[0].val, inp.args[1].val), tt256);
}
else if (inp.val == "div" && inp.args[1].val != "0") {
o = decimalDiv(inp.args[0].val, inp.args[1].val);
}
else if (inp.val == "sdiv" && inp.args[1].val != "0"
&& decimalGt(tt255, inp.args[0].val)
&& decimalGt(tt255, inp.args[1].val)) {
o = decimalDiv(inp.args[0].val, inp.args[1].val);
}
else if (inp.val == "mod" && inp.args[1].val != "0") {
o = decimalMod(inp.args[0].val, inp.args[1].val);
}
else if (inp.val == "smod" && inp.args[1].val != "0"
&& decimalGt(tt255, inp.args[0].val)
&& decimalGt(tt255, inp.args[1].val)) {
o = decimalMod(inp.args[0].val, inp.args[1].val);
}
else if (inp.val == "exp") {
o = decimalModExp(inp.args[0].val, inp.args[1].val, tt256);
}
if (o.length()) return token(o, inp.metadata);
}
return inp;
}
// Is a node degenerate (ie. trivial to calculate) ?
bool isDegenerate(Node n) {
return optimize(n).type == TOKEN;
}
// Is a node purely arithmetic?
bool isPureArithmetic(Node n) {
return isNumberLike(optimize(n));
}

19
libserpent/optimize.h
View File

@ -1,19 +0,0 @@
#ifndef ETHSERP_OPTIMIZER
#define ETHSERP_OPTIMIZER
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
// Compile-time arithmetic calculations
Node optimize(Node inp);
// Is a node degenerate (ie. trivial to calculate) ?
bool isDegenerate(Node n);
// Is a node purely arithmetic?
bool isPureArithmetic(Node n);
#endif

437
libserpent/parser.cpp
View File

@ -1,437 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
#include "parser.h"
#include "tokenize.h"
// Extended BEDMAS precedence order
int precedence(Node tok) {
std::string v = tok.val;
if (v == ".") return -1;
else if (v == "!" || v == "not") return 1;
else if (v=="^" || v == "**") return 2;
else if (v=="*" || v=="/" || v=="%") return 3;
else if (v=="+" || v=="-") return 4;
else if (v=="<" || v==">" || v=="<=" || v==">=") return 5;
else if (v=="&" || v=="|" || v=="xor" || v=="==" || v == "!=") return 6;
else if (v=="&&" || v=="and") return 7;
else if (v=="||" || v=="or") return 8;
else if (v=="=") return 10;
else if (v=="+=" || v=="-=" || v=="*=" || v=="/=" || v=="%=") return 10;
else if (v==":" || v == "::") return 11;
else return 0;
}
// Token classification for shunting-yard purposes
int toktype(Node tok) {
if (tok.type == ASTNODE) return COMPOUND;
std::string v = tok.val;
if (v == "(" || v == "[" || v == "{") return LPAREN;
else if (v == ")" || v == "]" || v == "}") return RPAREN;
else if (v == ",") return COMMA;
else if (v == "!" || v == "~" || v == "not") return UNARY_OP;
else if (precedence(tok) > 0) return BINARY_OP;
else if (precedence(tok) < 0) return TOKEN_SPLITTER;
if (tok.val[0] != '"' && tok.val[0] != '\'') {
for (unsigned i = 0; i < tok.val.length(); i++) {
if (chartype(tok.val[i]) == SYMB) {
err("Invalid symbol: "+tok.val, tok.metadata);
}
}
}
return ALPHANUM;
}
// Converts to reverse polish notation
std::vector<Node> shuntingYard(std::vector<Node> tokens) {
std::vector<Node> iq;
for (int i = tokens.size() - 1; i >= 0; i--) {
iq.push_back(tokens[i]);
}
std::vector<Node> oq;
std::vector<Node> stack;
Node prev, tok;
int prevtyp = 0, toktyp = 0;
while (iq.size()) {
prev = tok;
prevtyp = toktyp;
tok = iq.back();
toktyp = toktype(tok);
iq.pop_back();
// Alphanumerics go straight to output queue
if (toktyp == ALPHANUM) {
oq.push_back(tok);
}
// Left parens go on stack and output queue
else if (toktyp == LPAREN) {
while (stack.size() && toktype(stack.back()) == TOKEN_SPLITTER) {
oq.push_back(stack.back());
stack.pop_back();
}
if (prevtyp != ALPHANUM && prevtyp != RPAREN) {
oq.push_back(token("id", tok.metadata));
}
stack.push_back(tok);
oq.push_back(tok);
}
// If rparen, keep moving from stack to output queue until lparen
else if (toktyp == RPAREN) {
while (stack.size() && toktype(stack.back()) != LPAREN) {
oq.push_back(stack.back());
stack.pop_back();
}
if (stack.size()) {
stack.pop_back();
}
oq.push_back(tok);
}
else if (toktyp == UNARY_OP) {
stack.push_back(tok);
}
// If token splitter, just push it to the stack
else if (toktyp == TOKEN_SPLITTER) {
while (stack.size() && toktype(stack.back()) == TOKEN_SPLITTER) {
oq.push_back(stack.back());
stack.pop_back();
}
stack.push_back(tok);
}
// If binary op, keep popping from stack while higher bedmas precedence
else if (toktyp == BINARY_OP) {
if (tok.val == "-" && prevtyp != ALPHANUM && prevtyp != RPAREN) {
stack.push_back(tok);
oq.push_back(token("0", tok.metadata));
}
else {
int prec = precedence(tok);
while (stack.size()
&& (toktype(stack.back()) == BINARY_OP
|| toktype(stack.back()) == UNARY_OP
|| toktype(stack.back()) == TOKEN_SPLITTER)
&& precedence(stack.back()) <= prec) {
oq.push_back(stack.back());
stack.pop_back();
}
stack.push_back(tok);
}
}
// Comma means finish evaluating the argument
else if (toktyp == COMMA) {
while (stack.size() && toktype(stack.back()) != LPAREN) {
oq.push_back(stack.back());
stack.pop_back();
}
}
}
while (stack.size()) {
oq.push_back(stack.back());
stack.pop_back();
}
return oq;
}
// Converts reverse polish notation into tree
Node treefy(std::vector<Node> stream) {
std::vector<Node> iq;
for (int i = stream.size() -1; i >= 0; i--) {
iq.push_back(stream[i]);
}
std::vector<Node> oq;
while (iq.size()) {
Node tok = iq.back();
iq.pop_back();
int typ = toktype(tok);
// If unary, take node off end of oq and wrap it with the operator
// If binary, do the same with two nodes
if (typ == UNARY_OP || typ == BINARY_OP || typ == TOKEN_SPLITTER) {
std::vector<Node> args;
int rounds = (typ == UNARY_OP) ? 1 : 2;
for (int i = 0; i < rounds; i++) {
if (oq.size() == 0) {
err("Line malformed, not enough args for "+tok.val,
tok.metadata);
}
args.push_back(oq.back());
oq.pop_back();
}
std::vector<Node> args2;
while (args.size()) {
args2.push_back(args.back());
args.pop_back();
}
oq.push_back(astnode(tok.val, args2, tok.metadata));
}
// If rparen, keep grabbing until we get to an lparen
else if (typ == RPAREN) {
std::vector<Node> args;
while (1) {
if (toktype(oq.back()) == LPAREN) break;
args.push_back(oq.back());
oq.pop_back();
if (!oq.size()) err("Bracket without matching", tok.metadata);
}
oq.pop_back();
args.push_back(oq.back());
oq.pop_back();
// We represent a[b] as (access a b)
if (tok.val == "]")
args.push_back(token("access", tok.metadata));
if (args.back().type == ASTNODE)
args.push_back(token("fun", tok.metadata));
std::string fun = args.back().val;
args.pop_back();
// We represent [1,2,3] as (array_lit 1 2 3)
if (fun == "access" && args.size() && args.back().val == "id") {
fun = "array_lit";
args.pop_back();
}
std::vector<Node> args2;
while (args.size()) {
args2.push_back(args.back());
args.pop_back();
}
// When evaluating 2 + (3 * 5), the shunting yard algo turns that
// into 2 ( id 3 5 * ) +, effectively putting "id" as a dummy
// function where the algo was expecting a function to call the
// thing inside the brackets. This reverses that step
if (fun == "id" && args2.size() == 1) {
oq.push_back(args2[0]);
}
else {
oq.push_back(astnode(fun, args2, tok.metadata));
}
}
else oq.push_back(tok);
// This is messy, but has to be done. Import/inset other files here
std::string v = oq.back().val;
if ((v == "inset" || v == "import" || v == "create")
&& oq.back().args.size() == 1
&& oq.back().args[0].type == TOKEN) {
int lastSlashPos = tok.metadata.file.rfind("/");
std::string root;
if (lastSlashPos >= 0)
root = tok.metadata.file.substr(0, lastSlashPos) + "/";
else
root = "";
std::string filename = oq.back().args[0].val;
filename = filename.substr(1, filename.length() - 2);
if (!exists(root + filename))
err("File does not exist: "+root + filename, tok.metadata);
if (v == "inset") {
oq.pop_back();
oq.push_back(parseSerpent(root + filename));
}
else {
oq.back().args.pop_back();
oq.back().args.push_back(
asn("outer", parseSerpent(root + filename), tok.metadata));
}
}
//Useful for debugging
//for (int i = 0; i < oq.size(); i++) {
// std::cerr << printSimple(oq[i]) << " ";
//}
//std::cerr << " <-\n";
}
// Output must have one argument
if (oq.size() == 0) {
err("Output blank", Metadata());
}
else if (oq.size() > 1) {
return asn("multi", oq, oq[0].metadata);
}
return oq[0];
}
// Parses one line of serpent
Node parseSerpentTokenStream(std::vector<Node> s) {
return treefy(shuntingYard(s));
}
// Count spaces at beginning of line
int spaceCount(std::string s) {
unsigned pos = 0;
while (pos < s.length() && (s[pos] == ' ' || s[pos] == '\t'))
pos++;
return pos;
}
// Is this a command that takes an argument on the same line?
bool bodied(std::string tok) {
return tok == "if" || tok == "elif" || tok == "while"
|| tok == "with" || tok == "def" || tok == "extern"
|| tok == "data" || tok == "assert" || tok == "return"
|| tok == "fun" || tok == "scope" || tok == "macro"
|| tok == "type";
}
// Are the two commands meant to continue each other?
bool bodiedContinued(std::string prev, std::string tok) {
return (prev == "if" && tok == "elif")
|| (prev == "elif" && tok == "else")
|| (prev == "elif" && tok == "elif")
|| (prev == "if" && tok == "else");
}
// Is a line of code empty?
bool isLineEmpty(std::string line) {
std::vector<Node> tokens = tokenize(line);
if (!tokens.size() || tokens[0].val == "#" || tokens[0].val == "//")
return true;
return false;
}
// Parse lines of serpent (helper function)
Node parseLines(std::vector<std::string> lines, Metadata metadata, int sp) {
std::vector<Node> o;
int origLine = metadata.ln;
unsigned i = 0;
while (i < lines.size()) {
metadata.ln = origLine + i;
std::string main = lines[i];
if (isLineEmpty(main)) {
i += 1;
continue;
}
int spaces = spaceCount(main);
if (spaces != sp) {
err("Indent mismatch", metadata);
}
// Tokenize current line
std::vector<Node> tokens = tokenize(main.substr(sp), metadata);
// Remove comments
std::vector<Node> tokens2;
for (unsigned j = 0; j < tokens.size(); j++) {
if (tokens[j].val == "#" || tokens[j].val == "//") break;
tokens2.push_back(tokens[j]);
}
bool expectingChildBlock = false;
if (tokens2.size() > 0 && tokens2.back().val == ":") {
tokens2.pop_back();
expectingChildBlock = true;
}
// Parse current line
Node out = parseSerpentTokenStream(tokens2);
// Parse child block
int childIndent = 999999;
std::vector<std::string> childBlock;
while (1) {
i++;
if (i >= lines.size())
break;
bool ile = isLineEmpty(lines[i]);
if (!ile) {
int spaces = spaceCount(lines[i]);
if (spaces <= sp) break;
childBlock.push_back(lines[i]);
if (spaces < childIndent) childIndent = spaces;
}
else childBlock.push_back("");
}
// Child block empty?
bool cbe = true;
for (unsigned i = 0; i < childBlock.size(); i++) {
if (childBlock[i].length() > 0) { cbe = false; break; }
}
// Add child block to AST
if (expectingChildBlock) {
if (cbe)
err("Expected indented child block!", out.metadata);
out.type = ASTNODE;
metadata.ln += 1;
out.args.push_back(parseLines(childBlock, metadata, childIndent));
metadata.ln -= 1;
}
else if (!cbe)
err("Did not expect indented child block!", out.metadata);
else if (out.args.size() && out.args[out.args.size() - 1].val == ":") {
Node n = out.args[out.args.size() - 1];
out.args.pop_back();
out.args.push_back(n.args[0]);
out.args.push_back(n.args[1]);
}
// Bring back if / elif into AST
if (bodied(tokens[0].val)) {
if (out.val != "multi") {
// token not being used in bodied form
}
else if (out.args[0].val == "id")
out = astnode(tokens[0].val, out.args[1].args, out.metadata);
else if (out.args[0].type == TOKEN) {
std::vector<Node> out2;
for (unsigned i = 1; i < out.args.size(); i++)
out2.push_back(out.args[i]);
out = astnode(tokens[0].val, out2, out.metadata);
}
else
out = astnode("fun", out.args, out.metadata);
}
// Multi not supported
if (out.val == "multi")
err("Multiple expressions or unclosed bracket", out.metadata);
// Convert top-level colon expressions into non-colon expressions;
// makes if statements and the like equivalent indented or not
//if (out.val == ":" && out.args[0].type == TOKEN)
// out = asn(out.args[0].val, out.args[1], out.metadata);
//if (bodied(tokens[0].val) && out.args[0].val == ":")
// out = asn(tokens[0].val, out.args[0].args);
if (o.size() == 0 || o.back().type == TOKEN) {
o.push_back(out);
continue;
}
// This is a little complicated. Basically, the idea here is to build
// constructions like [if [< x 5] [a] [elif [< x 10] [b] [else [c]]]]
std::vector<Node> u;
u.push_back(o.back());
if (bodiedContinued(o.back().val, out.val)) {
while (1) {
if (!bodiedContinued(u.back().val, out.val)) {
u.pop_back();
break;
}
if (!u.back().args.size()
|| !bodiedContinued(u.back().val, u.back().args.back().val)) {
break;
}
u.push_back(u.back().args.back());
}
u.back().args.push_back(out);
while (u.size() > 1) {
Node v = u.back();
u.pop_back();
u.back().args.pop_back();
u.back().args.push_back(v);
}
o.pop_back();
o.push_back(u[0]);
}
else o.push_back(out);
}
if (o.size() == 1)
return o[0];
else if (o.size())
return astnode("seq", o, o[0].metadata);
else
return astnode("seq", o, Metadata());
}
// Parses serpent code
Node parseSerpent(std::string s) {
std::string input = s;
std::string file = "main";
if (exists(s)) {
file = s;
input = get_file_contents(s);
}
return parseLines(splitLines(input), Metadata(file, 0, 0), 0);
}
using namespace std;

13
libserpent/parser.h
View File

@ -1,13 +0,0 @@
#ifndef ETHSERP_PARSER
#define ETHSERP_PARSER
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
// Serpent text -> parse tree
Node parseSerpent(std::string s);
#endif

327
libserpent/preprocess.cpp
View File

@ -1,327 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
#include "lllparser.h"
#include "bignum.h"
#include "rewriteutils.h"
#include "optimize.h"
#include "preprocess.h"
#include "functions.h"
#include "opcodes.h"
// Convert a function of the form (def (f x y z) (do stuff)) into
// (if (first byte of ABI is correct) (seq (setup x y z) (do stuff)))
Node convFunction(Node node, int functionCount) {
std::string prefix = "_temp"+mkUniqueToken()+"_";
Metadata m = node.metadata;
if (node.args.size() != 2)
err("Malformed def!", m);
// Collect the list of variable names and variable byte counts
Node unpack = unpackArguments(node.args[0].args, m);
// And the actual code
Node body = node.args[1];
// Main LLL-based function body
return astnode("if",
astnode("eq",
astnode("get", token("__funid", m), m),
token(unsignedToDecimal(functionCount), m),
m),
astnode("seq", unpack, body, m));
}
// Populate an svObj with the arguments needed to determine
// the storage position of a node
svObj getStorageVars(svObj pre, Node node, std::string prefix,
int index) {
Metadata m = node.metadata;
if (!pre.globalOffset.size()) pre.globalOffset = "0";
std::vector<Node> h;
std::vector<std::string> coefficients;
// Array accesses or atoms
if (node.val == "access" || node.type == TOKEN) {
std::string tot = "1";
h = listfyStorageAccess(node);
coefficients.push_back("1");
for (unsigned i = h.size() - 1; i >= 1; i--) {
// Array sizes must be constant or at least arithmetically
// evaluable at compile time
if (!isPureArithmetic(h[i]))
err("Array size must be fixed value", m);
// Create a list of the coefficient associated with each
// array index
coefficients.push_back(decimalMul(coefficients.back(), h[i].val));
}
}
// Tuples
else {
int startc;
// Handle the (fun <fun_astnode> args...) case
if (node.val == "fun") {
startc = 1;
h = listfyStorageAccess(node.args[0]);
}
// Handle the (<fun_name> args...) case, which
// the serpent parser produces when the function
// is a simple name and not a complex astnode
else {
startc = 0;
h = listfyStorageAccess(token(node.val, m));
}
svObj sub = pre;
sub.globalOffset = "0";
// Evaluate tuple elements recursively
for (unsigned i = startc; i < node.args.size(); i++) {
sub = getStorageVars(sub,
node.args[i],
prefix+h[0].val.substr(2)+".",
i-startc);
}
coefficients.push_back(sub.globalOffset);
for (unsigned i = h.size() - 1; i >= 1; i--) {
// Array sizes must be constant or at least arithmetically
// evaluable at compile time
if (!isPureArithmetic(h[i]))
err("Array size must be fixed value", m);
// Create a list of the coefficient associated with each
// array index
coefficients.push_back(decimalMul(coefficients.back(), h[i].val));
}
pre.offsets = sub.offsets;
pre.coefficients = sub.coefficients;
pre.nonfinal = sub.nonfinal;
pre.nonfinal[prefix+h[0].val.substr(2)] = true;
}
pre.coefficients[prefix+h[0].val.substr(2)] = coefficients;
pre.offsets[prefix+h[0].val.substr(2)] = pre.globalOffset;
pre.indices[prefix+h[0].val.substr(2)] = index;
if (decimalGt(tt176, coefficients.back()))
pre.globalOffset = decimalAdd(pre.globalOffset, coefficients.back());
return pre;
}
// Preprocess input containing functions
//
// localExterns is a map of the form, eg,
//
// { x: { foo: 0, bar: 1, baz: 2 }, y: { qux: 0, foo: 1 } ... }
//
// localExternSigs is a map of the form, eg,
//
// { x : { foo: iii, bar: iis, baz: ia }, y: { qux: i, foo: as } ... }
//
// Signifying that x.foo = 0, x.baz = 2, y.foo = 1, etc
// and that x.foo has three integers as arguments, x.bar has two
// integers and a variable-length string, and baz has an integer
// and an array
//
// globalExterns is a one-level map, eg from above
//
// { foo: 1, bar: 1, baz: 2, qux: 0 }
//
// globalExternSigs is a one-level map, eg from above
//
// { foo: as, bar: iis, baz: ia, qux: i}
//
// Note that globalExterns and globalExternSigs may be ambiguous
// Also, a null signature implies an infinite tail of integers
preprocessResult preprocessInit(Node inp) {
Metadata m = inp.metadata;
if (inp.val != "seq")
inp = astnode("seq", inp, m);
std::vector<Node> empty = std::vector<Node>();
Node init = astnode("seq", empty, m);
Node shared = astnode("seq", empty, m);
std::vector<Node> any;
std::vector<Node> functions;
preprocessAux out = preprocessAux();
out.localExterns["self"] = std::map<std::string, int>();
int functionCount = 0;
int storageDataCount = 0;
for (unsigned i = 0; i < inp.args.size(); i++) {
Node obj = inp.args[i];
// Functions
if (obj.val == "def") {
if (obj.args.size() == 0)
err("Empty def", m);
std::string funName = obj.args[0].val;
// Init, shared and any are special functions
if (funName == "init" || funName == "shared" || funName == "any") {
if (obj.args[0].args.size())
err(funName+" cannot have arguments", m);
}
if (funName == "init") init = obj.args[1];
else if (funName == "shared") shared = obj.args[1];
else if (funName == "any") any.push_back(obj.args[1]);
else {
// Other functions
functions.push_back(convFunction(obj, functionCount));
out.localExterns["self"][obj.args[0].val] = functionCount;
out.localExternSigs["self"][obj.args[0].val]
= getSignature(obj.args[0].args);
functionCount++;
}
}
// Extern declarations
else if (obj.val == "extern") {
std::string externName = obj.args[0].val;
Node al = obj.args[1];
if (!out.localExterns.count(externName))
out.localExterns[externName] = std::map<std::string, int>();
for (unsigned i = 0; i < al.args.size(); i++) {
if (al.args[i].val == ":") {
std::string v = al.args[i].args[0].val;
std::string sig = al.args[i].args[1].val;
out.globalExterns[v] = i;
out.globalExternSigs[v] = sig;
out.localExterns[externName][v] = i;
out.localExternSigs[externName][v] = sig;
}
else {
std::string v = al.args[i].val;
out.globalExterns[v] = i;
out.globalExternSigs[v] = "";
out.localExterns[externName][v] = i;
out.localExternSigs[externName][v] = "";
}
}
}
// Custom macros
else if (obj.val == "macro" || (obj.val == "fun" && obj.args[0].val == "macro")) {
// Rules for valid macros:
//
// There are only four categories of valid macros:
//
// 1. a macro where the outer function is something
// which is NOT an existing valid function/extern/datum
// 2. a macro of the form set(c(x), d) where c must NOT
// be an existing valid function/extern/datum
// 3. something of the form access(c(x)), where c must NOT
// be an existing valid function/extern/datum
// 4. something of the form set(access(c(x)), d) where c must
// NOT be an existing valid function/extern/datum
// 5. something of the form with(c(x), d, e) where c must
// NOT be an existing valid function/extern/datum
bool valid = false;
Node pattern;
Node substitution;
int priority;
// Priority not set: default zero
if (obj.val == "macro") {
pattern = obj.args[0];
substitution = obj.args[1];
priority = 0;
}
// Specified priority
else {
pattern = obj.args[1];
substitution = obj.args[2];
if (obj.args[0].args.size())
priority = dtu(obj.args[0].args[0].val);
else
priority = 0;
}
if (opcode(pattern.val) < 0 && !isValidFunctionName(pattern.val))
valid = true;
if (pattern.val == "set" &&
opcode(pattern.args[0].val) < 0 &&
!isValidFunctionName(pattern.args[0].val))
valid = true;
if (pattern.val == "access" &&
opcode(pattern.args[0].val) < 0 &&
!isValidFunctionName(pattern.args[0].val))
if (pattern.val == "set" &&
pattern.args[0].val == "access" &&
opcode(pattern.args[0].args[0].val) < 0 &&
!isValidFunctionName(pattern.args[0].args[0].val))
valid = true;
if (pattern.val == "with" &&
opcode(pattern.args[0].val) < 0 &&
!isValidFunctionName(pattern.args[0].val))
valid = true;
if (valid) {
if (!out.customMacros.count(priority))
out.customMacros[priority] = rewriteRuleSet();
out.customMacros[priority].addRule
(rewriteRule(pattern, substitution));
}
else warn("Macro does not fit valid template: "+printSimple(pattern), m);
}
// Variable types
else if (obj.val == "type") {
std::string typeName = obj.args[0].val;
std::vector<Node> vars = obj.args[1].args;
for (unsigned i = 0; i < vars.size(); i++)
out.types[vars[i].val] = typeName;
}
// Storage variables/structures
else if (obj.val == "data") {
out.storageVars = getStorageVars(out.storageVars,
obj.args[0],
"",
storageDataCount);
storageDataCount += 1;
}
else any.push_back(obj);
}
// Set up top-level AST structure
std::vector<Node> main;
if (shared.args.size()) main.push_back(shared);
if (init.args.size()) main.push_back(init);
std::vector<Node> code;
if (shared.args.size()) code.push_back(shared);
for (unsigned i = 0; i < any.size(); i++)
code.push_back(any[i]);
for (unsigned i = 0; i < functions.size(); i++)
code.push_back(functions[i]);
Node codeNode;
if (functions.size() > 0) {
codeNode = astnode("with",
token("__funid", m),
astnode("byte",
token("0", m),
astnode("calldataload", token("0", m), m),
m),
astnode("seq", code, m),
m);
}
else codeNode = astnode("seq", code, m);
main.push_back(astnode("~return",
token("0", m),
astnode("lll",
codeNode,
token("0", m),
m),
m));
Node result;
if (main.size() == 1) result = main[0];
else result = astnode("seq", main, inp.metadata);
return preprocessResult(result, out);
}
preprocessResult processTypes (preprocessResult pr) {
preprocessAux aux = pr.second;
Node node = pr.first;
if (node.type == TOKEN && aux.types.count(node.val))
node = asn(aux.types[node.val], node, node.metadata);
else if (node.val == "untyped")
return preprocessResult(node.args[0], aux);
else if (node.val == "outer")
return preprocessResult(node, aux);
else {
for (unsigned i = 0; i < node.args.size(); i++) {
node.args[i] =
processTypes(preprocessResult(node.args[i], aux)).first;
}
}
return preprocessResult(node, aux);
}
preprocessResult preprocess(Node n) {
return processTypes(preprocessInit(n));
}

50
libserpent/preprocess.h
View File

@ -1,50 +0,0 @@
#ifndef ETHSERP_PREPROCESSOR
#define ETHSERP_PREPROCESSOR
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
#include "rewriteutils.h"
// Storage variable index storing object
struct svObj {
std::map<std::string, std::string> offsets;
std::map<std::string, int> indices;
std::map<std::string, std::vector<std::string> > coefficients;
std::map<std::string, bool> nonfinal;
std::string globalOffset;
};
// Preprocessing result storing object
class preprocessAux {
public:
preprocessAux() {
globalExterns = std::map<std::string, int>();
localExterns = std::map<std::string, std::map<std::string, int> >();
localExterns["self"] = std::map<std::string, int>();
}
std::map<std::string, int> globalExterns;
std::map<std::string, std::string> globalExternSigs;
std::map<std::string, std::map<std::string, int> > localExterns;
std::map<std::string, std::map<std::string, std::string> > localExternSigs;
std::map<int, rewriteRuleSet > customMacros;
std::map<std::string, std::string> types;
svObj storageVars;
};
#define preprocessResult std::pair<Node, preprocessAux>
// Populate an svObj with the arguments needed to determine
// the storage position of a node
svObj getStorageVars(svObj pre, Node node, std::string prefix="",
int index=0);
// Preprocess a function (see cpp for details)
preprocessResult preprocess(Node inp);
#endif

905
libserpent/rewriter.cpp
View File

@ -1,905 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
#include "lllparser.h"
#include "bignum.h"
#include "optimize.h"
#include "rewriteutils.h"
#include "preprocess.h"
#include "functions.h"
#include "opcodes.h"
// Rewrite rules
std::string macros[][2] = {
{
"(seq $x)",
"$x"
},
{
"(seq (seq) $x)",
"$x"
},
{
"(+= $a $b)",
"(set $a (+ $a $b))"
},
{
"(*= $a $b)",
"(set $a (* $a $b))"
},
{
"(-= $a $b)",
"(set $a (- $a $b))"
},
{
"(/= $a $b)",
"(set $a (/ $a $b))"
},
{
"(%= $a $b)",
"(set $a (% $a $b))"
},
{
"(^= $a $b)",
"(set $a (^ $a $b))"
},
{
"(!= $a $b)",
"(iszero (eq $a $b))"
},
{
"(assert $x)",
"(unless $x (stop))"
},
{
"(min $a $b)",
"(with $1 $a (with $2 $b (if (lt $1 $2) $1 $2)))"
},
{
"(max $a $b)",
"(with $1 $a (with $2 $b (if (lt $1 $2) $2 $1)))"
},
{
"(smin $a $b)",
"(with $1 $a (with $2 $b (if (slt $1 $2) $1 $2)))"
},
{
"(smax $a $b)",
"(with $1 $a (with $2 $b (if (slt $1 $2) $2 $1)))"
},
{
"(if $cond $do (else $else))",
"(if $cond $do $else)"
},
{
"(code $code)",
"$code"
},
{
"(slice $arr $pos)",
"(add $arr (mul 32 $pos))",
},
{
"(array $len)",
"(alloc (mul 32 $len))"
},
{
"(while $cond $do)",
"(until (iszero $cond) $do)",
},
{
"(while (iszero $cond) $do)",
"(until $cond $do)",
},
{
"(if $cond $do)",
"(unless (iszero $cond) $do)",
},
{
"(if (iszero $cond) $do)",
"(unless $cond $do)",
},
{
"(access (. self storage) $ind)",
"(sload $ind)"
},
{
"(access $var $ind)",
"(mload (add $var (mul 32 $ind)))"
},
{
"(set (access (. self storage) $ind) $val)",
"(sstore $ind $val)"
},
{
"(set (sload $ind) $val)",
"(sstore $ind $val)"
},
{
"(set (access $var $ind) $val)",
"(mstore (add $var (mul 32 $ind)) $val)"
},
{
"(getch $var $ind)",
"(mod (mload (sub (add $var $ind) 31)) 256)"
},
{
"(setch $var $ind $val)",
"(mstore8 (add $var $ind) $val)",
},
{
"(send $to $value)",
"(~call (sub (gas) 25) $to $value 0 0 0 0)"
},
{
"(send $gas $to $value)",
"(~call $gas $to $value 0 0 0 0)"
},
{
"(sha3 $x)",
"(seq (set $1 $x) (~sha3 (ref $1) 32))"
},
{
"(sha3 $mstart (= chars $msize))",
"(~sha3 $mstart $msize)"
},
{
"(sha3 $mstart $msize)",
"(~sha3 $mstart (mul 32 $msize))"
},
{
"(id $0)",
"$0"
},
{
"(return $x)",
"(seq (set $1 $x) (~return (ref $1) 32))"
},
{
"(return $mstart (= chars $msize))",
"(~return $mstart $msize)"
},
{
"(return $start $len)",
"(~return $start (mul 32 $len))"
},
{
"(&& $x $y)",
"(if $x $y 0)"
},
{
"(|| $x $y)",
"(with $1 $x (if $1 $1 $y))"
},
{
"(>= $x $y)",
"(iszero (slt $x $y))"
},
{
"(<= $x $y)",
"(iszero (sgt $x $y))"
},
{
"(create $code)",
"(create 0 $code)"
},
{
"(create $endowment $code)",
"(with $1 (msize) (create $endowment (get $1) (lll $code (msize))))"
},
{
"(sha256 $x)",
"(with $1 (alloc 64) (seq (mstore (add (get $1) 32) $x) (pop (~call 101 2 0 (add (get $1) 32) 32 (get $1) 32)) (mload (get $1))))"
},
{
"(sha256 $arr (= chars $sz))",
"(with $1 (alloc 32) (seq (pop (~call 101 2 0 $arr $sz (get $1) 32)) (mload (get $1))))"
},
{
"(sha256 $arr $sz)",
"(with $1 (alloc 32) (seq (pop (~call 101 2 0 $arr (mul 32 $sz) (get $1) 32)) (mload (get $1))))"
},
{
"(ripemd160 $x)",
"(with $1 (alloc 64) (seq (mstore (add (get $1) 32) $x) (pop (~call 101 3 0 (add (get $1) 32) 32 (get $1) 32)) (mload (get $1))))"
},
{
"(ripemd160 $arr (= chars $sz))",
"(with $1 (alloc 32) (seq (pop (~call 101 3 0 $arr $sz (mload $1) 32)) (mload (get $1))))"
},
{
"(ripemd160 $arr $sz)",
"(with $1 (alloc 32) (seq (pop (~call 101 3 0 $arr (mul 32 $sz) (get $1) 32)) (mload (get $1))))"
},
{
"(ecrecover $h $v $r $s)",
"(with $1 (alloc 160) (seq (mstore (get $1) $h) (mstore (add (get $1) 32) $v) (mstore (add (get $1) 64) $r) (mstore (add (get $1) 96) $s) (pop (~call 101 1 0 (get $1) 128 (add (get $1 128)) 32)) (mload (add (get $1) 128))))"
},
{
"(inset $x)",
"$x"
},
{
"(create $x)",
"(with $1 (msize) (create $val (get $1) (lll $code (get $1))))"
},
{
"(with (= $var $val) $cond)",
"(with $var $val $cond)"
},
{
"(log $t1)",
"(~log1 0 0 $t1)"
},
{
"(log $t1 $t2)",
"(~log2 0 0 $t1 $t2)"
},
{
"(log $t1 $t2 $t3)",
"(~log3 0 0 $t1 $t2 $t3)"
},
{
"(log $t1 $t2 $t3 $t4)",
"(~log4 0 0 $t1 $t2 $t3 $t4)"
},
{
"(logarr $a $sz)",
"(~log0 $a (mul 32 $sz))"
},
{
"(logarr $a $sz $t1)",
"(~log1 $a (mul 32 $sz) $t1)"
},
{
"(logarr $a $sz $t1 $t2)",
"(~log2 $a (mul 32 $sz) $t1 $t2)"
},
{
"(logarr $a $sz $t1 $t2 $t3)",
"(~log3 $a (mul 32 $sz) $t1 $t2 $t3)"
},
{
"(logarr $a $sz $t1 $t2 $t3 $t4)",
"(~log4 $a (mul 32 $sz) $t1 $t2 $t3 $t4)"
},
{
"(save $loc $array (= chars $count))",
"(with $location (ref $loc) (with $c $count (with $end (div $c 32) (with $i 0 (seq (while (slt $i $end) (seq (sstore (add $i $location) (access $array $i)) (set $i (add $i 1)))) (sstore (add $i $location) (~and (access $array $i) (sub 0 (exp 256 (sub 32 (mod $c 32)))))))))))"
},
{
"(save $loc $array $count)",
"(with $location (ref $loc) (with $end $count (with $i 0 (while (slt $i $end) (seq (sstore (add $i $location) (access $array $i)) (set $i (add $i 1)))))))"
},
{
"(load $loc (= chars $count))",
"(with $location (ref $loc) (with $c $count (with $a (alloc $c) (with $i 0 (seq (while (slt $i (div $c 32)) (seq (set (access $a $i) (sload (add $location $i))) (set $i (add $i 1)))) (set (access $a $i) (~and (sload (add $location $i)) (sub 0 (exp 256 (sub 32 (mod $c 32)))))) $a)))))"
},
{
"(load $loc $count)",
"(with $location (ref $loc) (with $c $count (with $a (alloc $c) (with $i 0 (seq (while (slt $i $c) (seq (set (access $a $i) (sload (add $location $i))) (set $i (add $i 1)))) $a)))))"
},
{
"(unsafe_mcopy $to $from $sz)",
"(with _sz $sz (with _from $from (with _to $to (seq (comment STARTING UNSAFE MCOPY) (with _i 0 (while (lt _i _sz) (seq (mstore (add $to _i) (mload (add _from _i))) (set _i (add _i 32)))))))))"
},
{
"(mcopy $to $from $_sz)",
"(with _to $to (with _from $from (with _sz $sz (seq (comment STARTING MCOPY (with _i 0 (seq (while (lt (add _i 31) _sz) (seq (mstore (add _to _i) (mload (add _from _i))) (set _i (add _i 32)))) (with _mask (exp 256 (sub 32 (mod _sz 32))) (mstore (add $to _i) (add (mod (mload (add $to _i)) _mask) (and (mload (add $from _i)) (sub 0 _mask))))))))))))"
},
{ "(. msg sender)", "(caller)" },
{ "(. msg value)", "(callvalue)" },
{ "(. tx gasprice)", "(gasprice)" },
{ "(. tx origin)", "(origin)" },
{ "(. tx gas)", "(gas)" },
{ "(. $x balance)", "(balance $x)" },
{ "self", "(address)" },
{ "(. block prevhash)", "(prevhash)" },
{ "(. block coinbase)", "(coinbase)" },
{ "(. block timestamp)", "(timestamp)" },
{ "(. block number)", "(number)" },
{ "(. block difficulty)", "(difficulty)" },
{ "(. block gaslimit)", "(gaslimit)" },
{ "stop", "(stop)" },
{ "---END---", "" } //Keep this line at the end of the list
};
// Token synonyms
std::string synonyms[][2] = {
{ "or", "||" },
{ "and", "&&" },
{ "|", "~or" },
{ "&", "~and" },
{ "elif", "if" },
{ "!", "iszero" },
{ "~", "~not" },
{ "not", "iszero" },
{ "string", "alloc" },
{ "+", "add" },
{ "-", "sub" },
{ "*", "mul" },
{ "/", "sdiv" },
{ "^", "exp" },
{ "**", "exp" },
{ "%", "smod" },
{ "<", "slt" },
{ ">", "sgt" },
{ "=", "set" },
{ "==", "eq" },
{ ":", "kv" },
{ "---END---", "" } //Keep this line at the end of the list
};
std::map<std::string, std::string> synonymMap;
// Custom setters (need to be registered separately
// for use with managed storage)
std::string setters[][2] = {
{ "+=", "+" },
{ "-=", "-" },
{ "*=", "*" },
{ "/=", "/" },
{ "%=", "%" },
{ "^=", "^" },
{ "---END---", "" } //Keep this line at the end of the list
};
std::map<std::string, std::string> setterMap;
// Processes mutable array literals
Node array_lit_transform(Node node) {
std::string prefix = "_temp"+mkUniqueToken() + "_";
Metadata m = node.metadata;
std::map<std::string, Node> d;
std::string o = "(seq (set $arr (alloc "+utd(node.args.size()*32)+"))";
for (unsigned i = 0; i < node.args.size(); i++) {
o += " (mstore (add (get $arr) "+utd(i * 32)+") $"+utd(i)+")";
d[utd(i)] = node.args[i];
}
o += " (get $arr))";
return subst(parseLLL(o), d, prefix, m);
}
// Processes long text literals
Node string_transform(Node node) {
Metadata m = node.metadata;
if (!node.args.size())
err("Empty text!", m);
if (node.args[0].val.size() < 2
|| node.args[0].val[0] != '"'
|| node.args[0].val[node.args[0].val.size() - 1] != '"')
err("Text contents don't look like a string!", m);
std::string bin = node.args[0].val.substr(1, node.args[0].val.size() - 2);
unsigned sz = bin.size();
std::vector<Node> o;
for (unsigned i = 0; i < sz; i += 32) {
std::string t = binToNumeric(bin.substr(i, 32));
if ((sz - i) < 32 && (sz - i) > 0) {
while ((sz - i) < 32) {
t = decimalMul(t, "256");
i--;
}
i = sz;
}
o.push_back(token(t, node.metadata));
}
node = astnode("array_lit", o, node.metadata);
return array_lit_transform(node);
}
Node apply_rules(preprocessResult pr);
// Transform "<variable>.<fun>(args...)" into
// a call
Node dotTransform(Node node, preprocessAux aux) {
Metadata m = node.metadata;
// We're gonna make lots of temporary variables,
// so set up a unique flag for them
std::string prefix = "_temp"+mkUniqueToken()+"_";
// Check that the function name is a token
if (node.args[0].args[1].type == ASTNODE)
err("Function name must be static", m);
Node dotOwner = node.args[0].args[0];
std::string dotMember = node.args[0].args[1].val;
// kwargs = map of special arguments
std::map<std::string, Node> kwargs;
kwargs["value"] = token("0", m);
kwargs["gas"] = subst(parseLLL("(- (gas) 25)"), msn(), prefix, m);
// Search for as=? and call=code keywords, and isolate the actual
// function arguments
std::vector<Node> fnargs;
std::string as = "";
std::string op = "call";
for (unsigned i = 1; i < node.args.size(); i++) {
fnargs.push_back(node.args[i]);
Node arg = fnargs.back();
if (arg.val == "=" || arg.val == "set") {
if (arg.args[0].val == "as")
as = arg.args[1].val;
if (arg.args[0].val == "call" && arg.args[1].val == "code")
op = "callcode";
if (arg.args[0].val == "gas")
kwargs["gas"] = arg.args[1];
if (arg.args[0].val == "value")
kwargs["value"] = arg.args[1];
if (arg.args[0].val == "outsz")
kwargs["outsz"] = arg.args[1];
}
}
if (dotOwner.val == "self") {
if (as.size()) err("Cannot use \"as\" when calling self!", m);
as = dotOwner.val;
}
// Determine the funId and sig assuming the "as" keyword was used
int funId = 0;
std::string sig;
if (as.size() > 0 && aux.localExterns.count(as)) {
if (!aux.localExterns[as].count(dotMember))
err("Invalid call: "+printSimple(dotOwner)+"."+dotMember, m);
funId = aux.localExterns[as][dotMember];
sig = aux.localExternSigs[as][dotMember];
}
// Determine the funId and sig otherwise
else if (!as.size()) {
if (!aux.globalExterns.count(dotMember))
err("Invalid call: "+printSimple(dotOwner)+"."+dotMember, m);
std::string key = unsignedToDecimal(aux.globalExterns[dotMember]);
funId = aux.globalExterns[dotMember];
sig = aux.globalExternSigs[dotMember];
}
else err("Invalid call: "+printSimple(dotOwner)+"."+dotMember, m);
// Pack arguments
kwargs["data"] = packArguments(fnargs, sig, funId, m);
kwargs["to"] = dotOwner;
Node main;
// Pack output
if (!kwargs.count("outsz")) {
main = parseLLL(
"(with _data $data (seq "
"(pop (~"+op+" $gas $to $value (access _data 0) (access _data 1) (ref $dataout) 32))"
"(get $dataout)))");
}
else {
main = parseLLL(
"(with _data $data (with _outsz (mul 32 $outsz) (with _out (alloc _outsz) (seq "
"(pop (~"+op+" $gas $to $value (access _data 0) (access _data 1) _out _outsz))"
"(get _out)))))");
}
// Set up main call
Node o = subst(main, kwargs, prefix, m);
return o;
}
// Transform an access of the form self.bob, self.users[5], etc into
// a storage access
//
// There exist two types of objects: finite objects, and infinite
// objects. Finite objects are packed optimally tightly into storage
// accesses; for example:
//
// data obj[100](a, b[2][4], c)
//
// obj[0].a -> 0
// obj[0].b[0][0] -> 1
// obj[0].b[1][3] -> 8
// obj[45].c -> 459
//
// Infinite objects are accessed by sha3([v1, v2, v3 ... ]), where
// the values are a list of array indices and keyword indices, for
// example:
// data obj[](a, b[2][4], c)
// data obj2[](a, b[][], c)
//
// obj[0].a -> sha3([0, 0, 0])
// obj[5].b[1][3] -> sha3([0, 5, 1, 1, 3])
// obj[45].c -> sha3([0, 45, 2])
// obj2[0].a -> sha3([1, 0, 0])
// obj2[5].b[1][3] -> sha3([1, 5, 1, 1, 3])
// obj2[45].c -> sha3([1, 45, 2])
Node storageTransform(Node node, preprocessAux aux,
bool mapstyle=false, bool ref=false) {
Metadata m = node.metadata;
// Get a list of all of the "access parameters" used in order
// eg. self.users[5].cow[4][m[2]][woof] ->
// [--self, --users, 5, --cow, 4, m[2], woof]
std::vector<Node> hlist = listfyStorageAccess(node);
// For infinite arrays, the terms array will just provide a list
// of indices. For finite arrays, it's a list of index*coefficient
std::vector<Node> terms;
std::string offset = "0";
std::string prefix = "";
std::string varPrefix = "_temp"+mkUniqueToken()+"_";
int c = 0;
std::vector<std::string> coefficients;
coefficients.push_back("");
for (unsigned i = 1; i < hlist.size(); i++) {
// We pre-add the -- flag to parameter-like terms. For example,
// self.users[m] -> [--self, --users, m]
// self.users.m -> [--self, --users, --m]
if (hlist[i].val.substr(0, 2) == "--") {
prefix += hlist[i].val.substr(2) + ".";
std::string tempPrefix = prefix.substr(0, prefix.size()-1);
if (!aux.storageVars.offsets.count(tempPrefix))
return node;
if (c < (signed)coefficients.size() - 1)
err("Too few array index lookups", m);
if (c > (signed)coefficients.size() - 1)
err("Too many array index lookups", m);
coefficients = aux.storageVars.coefficients[tempPrefix];
// If the size of an object exceeds 2^176, we make it an infinite
// array
if (decimalGt(coefficients.back(), tt176) && !mapstyle)
return storageTransform(node, aux, true, ref);
offset = decimalAdd(offset, aux.storageVars.offsets[tempPrefix]);
c = 0;
if (mapstyle)
terms.push_back(token(unsignedToDecimal(
aux.storageVars.indices[tempPrefix])));
}
else if (mapstyle) {
terms.push_back(hlist[i]);
c += 1;
}
else {
if (c > (signed)coefficients.size() - 2)
err("Too many array index lookups", m);
terms.push_back(
astnode("mul",
hlist[i],
token(coefficients[coefficients.size() - 2 - c], m),
m));
c += 1;
}
}
if (aux.storageVars.nonfinal.count(prefix.substr(0, prefix.size()-1)))
err("Storage variable access not deep enough", m);
if (c < (signed)coefficients.size() - 1) {
err("Too few array index lookups", m);
}
if (c > (signed)coefficients.size() - 1) {
err("Too many array index lookups", m);
}
Node o;
if (mapstyle) {
std::string t = "_temp_"+mkUniqueToken();
std::vector<Node> sub;
for (unsigned i = 0; i < terms.size(); i++)
sub.push_back(asn("mstore",
asn("add",
tkn(utd(i * 32), m),
asn("get", tkn(t+"pos", m), m),
m),
terms[i],
m));
sub.push_back(tkn(t+"pos", m));
Node main = asn("with",
tkn(t+"pos", m),
asn("alloc", tkn(utd(terms.size() * 32), m), m),
asn("seq", sub, m),
m);
Node sz = token(utd(terms.size() * 32), m);
o = astnode("~sha3",
main,
sz,
m);
}
else {
// We add up all the index*coefficients
Node out = token(offset, node.metadata);
for (unsigned i = 0; i < terms.size(); i++) {
std::vector<Node> temp;
temp.push_back(out);
temp.push_back(terms[i]);
out = astnode("add", temp, node.metadata);
}
o = out;
}
if (ref) return o;
else return astnode("sload", o, node.metadata);
}
// Basic rewrite rule execution
std::pair<Node, bool> rulesTransform(Node node, rewriteRuleSet macros) {
std::string prefix = "_temp_"+mkUniqueToken();
bool changed = false;
if (!macros.ruleLists.count(node.val))
return std::pair<Node, bool>(node, false);
std::vector<rewriteRule> rules = macros.ruleLists[node.val];
for (unsigned pos = 0; pos < rules.size(); pos++) {
rewriteRule macro = rules[pos];
matchResult mr = match(macro.pattern, node);
if (mr.success) {
node = subst(macro.substitution, mr.map, prefix, node.metadata);
std::pair<Node, bool> o = rulesTransform(node, macros);
o.second = true;
return o;
}
}
return std::pair<Node, bool>(node, changed);
}
std::pair<Node, bool> synonymTransform(Node node) {
bool changed = false;
if (node.type == ASTNODE && synonymMap.count(node.val)) {
node.val = synonymMap[node.val];
changed = true;
}
return std::pair<Node, bool>(node, changed);
}
rewriteRuleSet nodeMacros;
rewriteRuleSet setterMacros;
bool dontDescend(std::string s) {
return s == "macro" || s == "comment" || s == "outer";
}
// Recursively applies any set of rewrite rules
std::pair<Node, bool> apply_rules_iter(preprocessResult pr, rewriteRuleSet rules) {
bool changed = false;
Node node = pr.first;
if (dontDescend(node.val))
return std::pair<Node, bool>(node, false);
std::pair<Node, bool> o = rulesTransform(node, rules);
node = o.first;
changed = changed || o.second;
if (node.type == ASTNODE) {
for (unsigned i = 0; i < node.args.size(); i++) {
std::pair<Node, bool> r =
apply_rules_iter(preprocessResult(node.args[i], pr.second), rules);
node.args[i] = r.first;
changed = changed || r.second;
}
}
return std::pair<Node, bool>(node, changed);
}
// Recursively applies rewrite rules and other primary transformations
std::pair<Node, bool> mainTransform(preprocessResult pr) {
bool changed = false;
Node node = pr.first;
// Anything inside "outer" should be treated as a separate program
// and thus recursively compiled in its entirety
if (node.val == "outer") {
node = apply_rules(preprocess(node.args[0]));
changed = true;
}
// Don't descend into comments, macros and inner scopes
if (dontDescend(node.val))
return std::pair<Node, bool>(node, changed);
// Special storage transformation
if (isNodeStorageVariable(node)) {
node = storageTransform(node, pr.second);
changed = true;
}
if (node.val == "ref" && isNodeStorageVariable(node.args[0])) {
node = storageTransform(node.args[0], pr.second, false, true);
changed = true;
}
if (node.val == "=" && isNodeStorageVariable(node.args[0])) {
Node t = storageTransform(node.args[0], pr.second);
if (t.val == "sload") {
std::vector<Node> o;
o.push_back(t.args[0]);
o.push_back(node.args[1]);
node = astnode("sstore", o, node.metadata);
}
changed = true;
}
// Main code
std::pair<Node, bool> pnb = synonymTransform(node);
node = pnb.first;
changed = changed || pnb.second;
// std::cerr << priority << " " << macros.size() << "\n";
std::pair<Node, bool> pnc = rulesTransform(node, nodeMacros);
node = pnc.first;
changed = changed || pnc.second;
// Special transformations
if (node.val == "array_lit") {
node = array_lit_transform(node);
changed = true;
}
if (node.val == "fun" && node.args[0].val == ".") {
node = dotTransform(node, pr.second);
changed = true;
}
if (node.val == "text") {
node = string_transform(node);
changed = true;
}
if (node.type == ASTNODE) {
unsigned i = 0;
// Arg 0 of all of these is a variable, so should not be changed
if (node.val == "set" || node.val == "ref"
|| node.val == "get" || node.val == "with") {
if (node.args[0].type == TOKEN &&
node.args[0].val.size() > 0 && node.args[0].val[0] != '\'') {
node.args[0].val = "'" + node.args[0].val;
changed = true;
}
i = 1;
}
// Convert arglen(x) to '_len_x
else if (node.val == "arglen") {
node.val = "get";
node.args[0].val = "'_len_" + node.args[0].val;
i = 1;
changed = true;
}
// Recursively process children
for (; i < node.args.size(); i++) {
std::pair<Node, bool> r =
mainTransform(preprocessResult(node.args[i], pr.second));
node.args[i] = r.first;
changed = changed || r.second;
}
}
// Add leading ' to variable names, and wrap them inside get
else if (node.type == TOKEN && !isNumberLike(node)) {
if (node.val.size() && node.val[0] != '\'' && node.val[0] != '$') {
Node n = astnode("get", tkn("'"+node.val), node.metadata);
node = n;
changed = true;
}
}
// Convert all numbers to normalized form
else if (node.type == TOKEN && isNumberLike(node) && !isDecimal(node.val)) {
node.val = strToNumeric(node.val);
changed = true;
}
return std::pair<Node, bool>(node, changed);
}
// Do some preprocessing to convert all of our macro lists into compiled
// forms that can then be reused
void parseMacros() {
for (int i = 0; i < 9999; i++) {
std::vector<Node> o;
if (macros[i][0] == "---END---") break;
nodeMacros.addRule(rewriteRule(
parseLLL(macros[i][0]),
parseLLL(macros[i][1])
));
}
for (int i = 0; i < 9999; i++) {
std::vector<Node> o;
if (setters[i][0] == "---END---") break;
setterMacros.addRule(rewriteRule(
asn(setters[i][0], tkn("$x"), tkn("$y")),
asn("=", tkn("$x"), asn(setters[i][1], tkn("$x"), tkn("$y")))
));
}
for (int i = 0; i < 9999; i++) {
if (synonyms[i][0] == "---END---") break;
synonymMap[synonyms[i][0]] = synonyms[i][1];
}
}
Node apply_rules(preprocessResult pr) {
// If the rewrite rules have not yet been parsed, parse them
if (!nodeMacros.ruleLists.size()) parseMacros();
// Iterate over macros by priority list
std::map<int, rewriteRuleSet >::iterator it;
std::pair<Node, bool> r;
for(it=pr.second.customMacros.begin();
it != pr.second.customMacros.end(); it++) {
while (1) {
// std::cerr << "STARTING ARI CYCLE: " << (*it).first <<"\n";
// std::cerr << printAST(pr.first) << "\n";
r = apply_rules_iter(pr, (*it).second);
pr.first = r.first;
if (!r.second) break;
}
}
// Apply setter macros
while (1) {
r = apply_rules_iter(pr, setterMacros);
pr.first = r.first;
if (!r.second) break;
}
// Apply all other mactos
while (1) {
r = mainTransform(pr);
pr.first = r.first;
if (!r.second) break;
}
return r.first;
}
// Pre-validation
Node validate(Node inp) {
Metadata m = inp.metadata;
if (inp.type == ASTNODE) {
int i = 0;
while(validFunctions[i][0] != "---END---") {
if (inp.val == validFunctions[i][0]) {
std::string sz = unsignedToDecimal(inp.args.size());
if (decimalGt(validFunctions[i][1], sz)) {
err("Too few arguments for "+inp.val, inp.metadata);
}
if (decimalGt(sz, validFunctions[i][2])) {
err("Too many arguments for "+inp.val, inp.metadata);
}
}
i++;
}
}
else if (inp.type == TOKEN) {
if (!inp.val.size()) err("??? empty token", m);
if (inp.val[0] == '_') err("Variables cannot start with _", m);
}
for (unsigned i = 0; i < inp.args.size(); i++) validate(inp.args[i]);
return inp;
}
Node postValidate(Node inp) {
// This allows people to use ~x as a way of having functions with the same
// name and arity as macros; the idea is that ~x is a "final" form, and
// should not be remacroed, but it is converted back at the end
if (inp.val.size() > 0 && inp.val[0] == '~') {
inp.val = inp.val.substr(1);
}
if (inp.type == ASTNODE) {
if (inp.val == ".")
err("Invalid object member (ie. a foo.bar not mapped to anything)",
inp.metadata);
else if (opcode(inp.val) >= 0) {
if ((signed)inp.args.size() < opinputs(inp.val))
err("Too few arguments for "+inp.val, inp.metadata);
if ((signed)inp.args.size() > opinputs(inp.val))
err("Too many arguments for "+inp.val, inp.metadata);
}
else if (isValidLLLFunc(inp.val, inp.args.size())) {
// do nothing
}
else err ("Invalid argument count or LLL function: "+printSimple(inp), inp.metadata);
for (unsigned i = 0; i < inp.args.size(); i++) {
inp.args[i] = postValidate(inp.args[i]);
}
}
return inp;
}
Node rewriteChunk(Node inp) {
return postValidate(optimize(apply_rules(
preprocessResult(
validate(inp), preprocessAux()))));
}
// Flatten nested sequence into flat sequence
Node flattenSeq(Node inp) {
std::vector<Node> o;
if (inp.val == "seq" && inp.type == ASTNODE) {
for (unsigned i = 0; i < inp.args.size(); i++) {
if (inp.args[i].val == "seq" && inp.args[i].type == ASTNODE)
o = extend(o, flattenSeq(inp.args[i]).args);
else
o.push_back(flattenSeq(inp.args[i]));
}
}
else if (inp.type == ASTNODE) {
for (unsigned i = 0; i < inp.args.size(); i++) {
o.push_back(flattenSeq(inp.args[i]));
}
}
else return inp;
return asn(inp.val, o, inp.metadata);
}
Node rewrite(Node inp) {
return postValidate(optimize(apply_rules(preprocess(flattenSeq(inp)))));
}
using namespace std;

16
libserpent/rewriter.h
View File

@ -1,16 +0,0 @@
#ifndef ETHSERP_REWRITER
#define ETHSERP_REWRITER
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
// Applies rewrite rules
Node rewrite(Node inp);
// Applies rewrite rules adding without wrapper
Node rewriteChunk(Node inp);
#endif

212
libserpent/rewriteutils.cpp
View File

@ -1,212 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include <string>
#include "util.h"
#include "lllparser.h"
#include "bignum.h"
#include "rewriteutils.h"
#include "optimize.h"
// Valid functions and their min and max argument counts
std::string validFunctions[][3] = {
{ "if", "2", "3" },
{ "unless", "2", "2" },
{ "while", "2", "2" },
{ "until", "2", "2" },
{ "alloc", "1", "1" },
{ "array", "1", "1" },
{ "call", "2", tt256 },
{ "callcode", "2", tt256 },
{ "create", "1", "4" },
{ "getch", "2", "2" },
{ "setch", "3", "3" },
{ "sha3", "1", "2" },
{ "return", "1", "2" },
{ "inset", "1", "1" },
{ "min", "2", "2" },
{ "max", "2", "2" },
{ "array_lit", "0", tt256 },
{ "seq", "0", tt256 },
{ "log", "1", "6" },
{ "outer", "1", "1" },
{ "set", "2", "2" },
{ "get", "1", "1" },
{ "ref", "1", "1" },
{ "declare", "1", tt256 },
{ "with", "3", "3" },
{ "outer", "1", "1" },
{ "mcopy", "3", "3" },
{ "unsafe_mcopy", "3", "3" },
{ "save", "3", "3" },
{ "load", "2", "2" },
{ "---END---", "", "" } //Keep this line at the end of the list
};
std::map<std::string, bool> vfMap;
// Is a function name one of the valid functions above?
bool isValidFunctionName(std::string f) {
if (vfMap.size() == 0) {
for (int i = 0; ; i++) {
if (validFunctions[i][0] == "---END---") break;
vfMap[validFunctions[i][0]] = true;
}
}
return vfMap.count(f) != 0;
}
// Cool function for debug purposes (named cerrStringList to make
// all prints searchable via 'cerr')
void cerrStringList(std::vector<std::string> s, std::string suffix) {
for (unsigned i = 0; i < s.size(); i++) std::cerr << s[i] << " ";
std::cerr << suffix << "\n";
}
// Convert:
// self.cow -> ["cow"]
// self.horse[0] -> ["horse", "0"]
// self.a[6][7][self.storage[3]].chicken[9] ->
// ["6", "7", (sload 3), "chicken", "9"]
std::vector<Node> listfyStorageAccess(Node node) {
std::vector<Node> out;
std::vector<Node> nodez;
nodez.push_back(node);
while (1) {
if (nodez.back().type == TOKEN) {
out.push_back(token("--" + nodez.back().val, node.metadata));
std::vector<Node> outrev;
for (int i = (signed)out.size() - 1; i >= 0; i--) {
outrev.push_back(out[i]);
}
return outrev;
}
if (nodez.back().val == ".")
nodez.back().args[1].val = "--" + nodez.back().args[1].val;
if (nodez.back().args.size() == 0)
err("Error parsing storage variable statement", node.metadata);
if (nodez.back().args.size() == 1)
out.push_back(token(tt256m1, node.metadata));
else
out.push_back(nodez.back().args[1]);
nodez.push_back(nodez.back().args[0]);
}
}
// Is the given node something of the form
// self.cow
// self.horse[0]
// self.a[6][7][self.storage[3]].chicken[9]
bool isNodeStorageVariable(Node node) {
std::vector<Node> nodez;
nodez.push_back(node);
while (1) {
if (nodez.back().type == TOKEN) return false;
if (nodez.back().args.size() == 0) return false;
if (nodez.back().val != "." && nodez.back().val != "access")
return false;
if (nodez.back().args[0].val == "self") return true;
nodez.push_back(nodez.back().args[0]);
}
}
// Main pattern matching routine, for those patterns that can be expressed
// using our standard mini-language above
//
// Returns two values. First, a boolean to determine whether the node matches
// the pattern, second, if the node does match then a map mapping variables
// in the pattern to nodes
matchResult match(Node p, Node n) {
matchResult o;
o.success = false;
if (p.type == TOKEN) {
if (p.val == n.val && n.type == TOKEN) o.success = true;
else if (p.val[0] == '$' || p.val[0] == '@') {
o.success = true;
o.map[p.val.substr(1)] = n;
}
}
else if (n.type==TOKEN || p.val!=n.val || p.args.size()!=n.args.size()) {
// do nothing
}
else {
for (unsigned i = 0; i < p.args.size(); i++) {
matchResult oPrime = match(p.args[i], n.args[i]);
if (!oPrime.success) {
o.success = false;
return o;
}
for (std::map<std::string, Node>::iterator it = oPrime.map.begin();
it != oPrime.map.end();
it++) {
o.map[(*it).first] = (*it).second;
}
}
o.success = true;
}
return o;
}
// Fills in the pattern with a dictionary mapping variable names to
// nodes (these dicts are generated by match). Match and subst together
// create a full pattern-matching engine.
Node subst(Node pattern,
std::map<std::string, Node> dict,
std::string varflag,
Metadata m) {
// Swap out patterns at the token level
if (pattern.metadata.ln == -1)
pattern.metadata = m;
if (pattern.type == TOKEN &&
pattern.val[0] == '$') {
if (dict.count(pattern.val.substr(1))) {
return dict[pattern.val.substr(1)];
}
else {
return token(varflag + pattern.val.substr(1), m);
}
}
// Other tokens are untouched
else if (pattern.type == TOKEN) {
return pattern;
}
// Substitute recursively for ASTs
else {
std::vector<Node> args;
for (unsigned i = 0; i < pattern.args.size(); i++) {
args.push_back(subst(pattern.args[i], dict, varflag, m));
}
return asn(pattern.val, args, m);
}
}
// Transforms a sequence containing two-argument with statements
// into a statement containing those statements in nested form
Node withTransform (Node source) {
Node o = token("--");
Metadata m = source.metadata;
std::vector<Node> args;
for (int i = source.args.size() - 1; i >= 0; i--) {
Node a = source.args[i];
if (a.val == "with" && a.args.size() == 2) {
std::vector<Node> flipargs;
for (int j = args.size() - 1; j >= 0; j--)
flipargs.push_back(args[i]);
if (o.val != "--")
flipargs.push_back(o);
o = asn("with", a.args[0], a.args[1], asn("seq", flipargs, m), m);
args = std::vector<Node>();
}
else {
args.push_back(a);
}
}
std::vector<Node> flipargs;
for (int j = args.size() - 1; j >= 0; j--)
flipargs.push_back(args[j]);
if (o.val != "--")
flipargs.push_back(o);
return asn("seq", flipargs, m);
}

76
libserpent/rewriteutils.h
View File

@ -1,76 +0,0 @@
#ifndef ETHSERP_REWRITEUTILS
#define ETHSERP_REWRITEUTILS
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
// Valid functions and their min and max argument counts
extern std::string validFunctions[][3];
extern std::map<std::string, bool> vfMap;
bool isValidFunctionName(std::string f);
// Converts deep array access into ordered list of the arguments
// along the descent
std::vector<Node> listfyStorageAccess(Node node);
// Cool function for debug purposes (named cerrStringList to make
// all prints searchable via 'cerr')
void cerrStringList(std::vector<std::string> s, std::string suffix="");
// Is the given node something of the form
// self.cow
// self.horse[0]
// self.a[6][7][self.storage[3]].chicken[9]
bool isNodeStorageVariable(Node node);
// Applies rewrite rules adding without wrapper
Node rewriteChunk(Node inp);
// Match result storing object
struct matchResult {
bool success;
std::map<std::string, Node> map;
};
// Match node to pattern
matchResult match(Node p, Node n);
// Substitute node using pattern
Node subst(Node pattern,
std::map<std::string, Node> dict,
std::string varflag,
Metadata m);
Node withTransform(Node source);
class rewriteRule {
public:
rewriteRule(Node p, Node s) {
pattern = p;
substitution = s;
}
Node pattern;
Node substitution;
};
class rewriteRuleSet {
public:
rewriteRuleSet() {
ruleLists = std::map<std::string, std::vector<rewriteRule> >();
}
void addRule(rewriteRule r) {
if (!ruleLists.count(r.pattern.val))
ruleLists[r.pattern.val] = std::vector<rewriteRule>();
ruleLists[r.pattern.val].push_back(r);
}
std::map<std::string, std::vector<rewriteRule> > ruleLists;
};
#endif

115
libserpent/tokenize.cpp
View File

@ -1,115 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
// These appear as independent tokens even if inside a stream of symbols
const std::string atoms[] = { "#", "//", "(", ")", "[", "]", "{", "}" };
const int numAtoms = 8;
// Is the char alphanumeric, a space, a bracket, a quote, a symbol?
int chartype(char c) {
if (c >= '0' && c <= '9') return ALPHANUM;
else if (c >= 'a' && c <= 'z') return ALPHANUM;
else if (c >= 'A' && c <= 'Z') return ALPHANUM;
else if (std::string("~_$@").find(c) != std::string::npos) return ALPHANUM;
else if (c == '\t' || c == ' ' || c == '\n' || c == '\r') return SPACE;
else if (std::string("()[]{}").find(c) != std::string::npos) return BRACK;
else if (c == '"') return DQUOTE;
else if (c == '\'') return SQUOTE;
else return SYMB;
}
// "y = f(45,124)/3" -> [ "y", "f", "(", "45", ",", "124", ")", "/", "3"]
std::vector<Node> tokenize(std::string inp, Metadata metadata, bool lispMode) {
int curtype = SPACE;
unsigned pos = 0;
int lastNewline = 0;
metadata.ch = 0;
std::string cur;
std::vector<Node> out;
inp += " ";
while (pos < inp.length()) {
int headtype = chartype(inp[pos]);
if (lispMode) {
if (inp[pos] == '\'') headtype = ALPHANUM;
}
// Are we inside a quote?
if (curtype == SQUOTE || curtype == DQUOTE) {
// Close quote
if (headtype == curtype) {
cur += inp[pos];
out.push_back(token(cur, metadata));
cur = "";
metadata.ch = pos - lastNewline;
curtype = SPACE;
pos += 1;
}
// eg. \xc3
else if (inp.length() >= pos + 4 && inp.substr(pos, 2) == "\\x") {
cur += (std::string("0123456789abcdef").find(inp[pos+2]) * 16
+ std::string("0123456789abcdef").find(inp[pos+3]));
pos += 4;
}
// Newline
else if (inp.substr(pos, 2) == "\\n") {
cur += '\n';
pos += 2;
}
// Backslash escape
else if (inp.length() >= pos + 2 && inp[pos] == '\\') {
cur += inp[pos + 1];
pos += 2;
}
// Normal character
else {
cur += inp[pos];
pos += 1;
}
}
else {
// Handle atoms ( '//', '#', brackets )
for (int i = 0; i < numAtoms; i++) {
int split = cur.length() - atoms[i].length();
if (split >= 0 && cur.substr(split) == atoms[i]) {
if (split > 0) {
out.push_back(token(cur.substr(0, split), metadata));
}
metadata.ch += split;
out.push_back(token(cur.substr(split), metadata));
metadata.ch = pos - lastNewline;
cur = "";
curtype = SPACE;
}
}
// Special case the minus sign
if (cur.length() > 1 && (cur.substr(cur.length() - 1) == "-"
|| cur.substr(cur.length() - 1) == "!")) {
out.push_back(token(cur.substr(0, cur.length() - 1), metadata));
out.push_back(token(cur.substr(cur.length() - 1), metadata));
cur = "";
}
// Boundary between different char types
if (headtype != curtype) {
if (curtype != SPACE && cur != "") {
out.push_back(token(cur, metadata));
}
metadata.ch = pos - lastNewline;
cur = "";
}
cur += inp[pos];
curtype = headtype;
pos += 1;
}
if (inp[pos] == '\n') {
lastNewline = pos;
metadata.ch = 0;
metadata.ln += 1;
}
}
return out;
}

16
libserpent/tokenize.h
View File

@ -1,16 +0,0 @@
#ifndef ETHSERP_TOKENIZE
#define ETHSERP_TOKENIZE
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
int chartype(char c);
std::vector<Node> tokenize(std::string inp,
Metadata meta=Metadata(),
bool lispMode=false);
#endif

333
libserpent/util.cpp
View File

@ -1,333 +0,0 @@
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include "util.h"
#include "bignum.h"
#include <fstream>
#include <cerrno>
//Token or value node constructor
Node token(std::string val, Metadata met) {
Node o;
o.type = 0;
o.val = val;
o.metadata = met;
return o;
}
//AST node constructor
Node astnode(std::string val, std::vector<Node> args, Metadata met) {
Node o;
o.type = 1;
o.val = val;
o.args = args;
o.metadata = met;
return o;
}
//AST node constructors for a specific number of children
Node astnode(std::string val, Metadata met) {
std::vector<Node> args;
return astnode(val, args, met);
}
Node astnode(std::string val, Node a, Metadata met) {
std::vector<Node> args;
args.push_back(a);
return astnode(val, args, met);
}
Node astnode(std::string val, Node a, Node b, Metadata met) {
std::vector<Node> args;
args.push_back(a);
args.push_back(b);
return astnode(val, args, met);
}
Node astnode(std::string val, Node a, Node b, Node c, Metadata met) {
std::vector<Node> args;
args.push_back(a);
args.push_back(b);
args.push_back(c);
return astnode(val, args, met);
}
Node astnode(std::string val, Node a, Node b, Node c, Node d, Metadata met) {
std::vector<Node> args;
args.push_back(a);
args.push_back(b);
args.push_back(c);
args.push_back(d);
return astnode(val, args, met);
}
// Print token list
std::string printTokens(std::vector<Node> tokens) {
std::string s = "";
for (unsigned i = 0; i < tokens.size(); i++) {
s += tokens[i].val + " ";
}
return s;
}
// Prints a lisp AST on one line
std::string printSimple(Node ast) {
if (ast.type == TOKEN) return ast.val;
std::string o = "(" + ast.val;
std::vector<std::string> subs;
for (unsigned i = 0; i < ast.args.size(); i++) {
o += " " + printSimple(ast.args[i]);
}
return o + ")";
}
// Number of tokens in a tree
int treeSize(Node prog) {
if (prog.type == TOKEN) return 1;
int o = 0;
for (unsigned i = 0; i < prog.args.size(); i++) o += treeSize(prog.args[i]);
return o;
}
// Pretty-prints a lisp AST
std::string printAST(Node ast, bool printMetadata) {
if (ast.type == TOKEN) return ast.val;
std::string o = "(";
if (printMetadata) {
o += ast.metadata.file + " ";
o += unsignedToDecimal(ast.metadata.ln) + " ";
o += unsignedToDecimal(ast.metadata.ch) + ": ";
}
o += ast.val;
std::vector<std::string> subs;
for (unsigned i = 0; i < ast.args.size(); i++) {
subs.push_back(printAST(ast.args[i], printMetadata));
}
unsigned k = 0;
std::string out = " ";
// As many arguments as possible go on the same line as the function,
// except when seq is used
while (k < subs.size() && o != "(seq") {
if (subs[k].find("\n") != std::string::npos || (out + subs[k]).length() >= 80) break;
out += subs[k] + " ";
k += 1;
}
// All remaining arguments go on their own lines
if (k < subs.size()) {
o += out + "\n";
std::vector<std::string> subsSliceK;
for (unsigned i = k; i < subs.size(); i++) subsSliceK.push_back(subs[i]);
o += indentLines(joinLines(subsSliceK));
o += "\n)";
}
else {
o += out.substr(0, out.size() - 1) + ")";
}
return o;
}
// Splits text by line
std::vector<std::string> splitLines(std::string s) {
unsigned pos = 0;
int lastNewline = 0;
std::vector<std::string> o;
while (pos < s.length()) {
if (s[pos] == '\n') {
o.push_back(s.substr(lastNewline, pos - lastNewline));
lastNewline = pos + 1;
}
pos = pos + 1;
}
o.push_back(s.substr(lastNewline));
return o;
}
// Inverse of splitLines
std::string joinLines(std::vector<std::string> lines) {
std::string o = "\n";
for (unsigned i = 0; i < lines.size(); i++) {
o += lines[i] + "\n";
}
return o.substr(1, o.length() - 2);
}
// Indent all lines by 4 spaces
std::string indentLines(std::string inp) {
std::vector<std::string> lines = splitLines(inp);
for (unsigned i = 0; i < lines.size(); i++) lines[i] = " "+lines[i];
return joinLines(lines);
}
// Binary to hexadecimal
std::string binToNumeric(std::string inp) {
std::string o = "0";
for (unsigned i = 0; i < inp.length(); i++) {
o = decimalAdd(decimalMul(o,"256"), unsignedToDecimal((unsigned char)inp[i]));
}
return o;
}
// Converts string to simple numeric format
std::string strToNumeric(std::string inp) {
std::string o = "0";
if (inp == "") {
o = "";
}
else if ((inp[0] == '"' && inp[inp.length()-1] == '"')
|| (inp[0] == '\'' && inp[inp.length()-1] == '\'')) {
for (unsigned i = 1; i < inp.length() - 1; i++) {
o = decimalAdd(decimalMul(o,"256"), unsignedToDecimal((unsigned char)inp[i]));
}
}
else if (inp.substr(0,2) == "0x") {
for (unsigned i = 2; i < inp.length(); i++) {
int dig = std::string("0123456789abcdef0123456789ABCDEF").find(inp[i]) % 16;
if (dig < 0) return "";
o = decimalAdd(decimalMul(o,"16"), unsignedToDecimal(dig));
}
}
else {
bool isPureNum = true;
for (unsigned i = 0; i < inp.length(); i++) {
isPureNum = isPureNum && inp[i] >= '0' && inp[i] <= '9';
}
o = isPureNum ? inp : "";
}
return o;
}
// Does the node contain a number (eg. 124, 0xf012c, "george")
bool isNumberLike(Node node) {
if (node.type == ASTNODE) return false;
return strToNumeric(node.val) != "";
}
// Is the number decimal?
bool isDecimal(std::string inp) {
for (unsigned i = 0; i < inp.length(); i++) {
if (inp[i] < '0' || inp[i] > '9') return false;
}
return true;
}
//Normalizes number representations
Node nodeToNumeric(Node node) {
std::string o = strToNumeric(node.val);
return token(o == "" ? node.val : o, node.metadata);
}
Node tryNumberize(Node node) {
if (node.type == TOKEN && isNumberLike(node)) return nodeToNumeric(node);
return node;
}
//Converts a value to an array of byte number nodes
std::vector<Node> toByteArr(std::string val, Metadata metadata, int minLen) {
std::vector<Node> o;
int L = 0;
while (val != "0" || L < minLen) {
o.push_back(token(decimalMod(val, "256"), metadata));
val = decimalDiv(val, "256");
L++;
}
std::vector<Node> o2;
for (int i = o.size() - 1; i >= 0; i--) o2.push_back(o[i]);
return o2;
}
int counter = 0;
//Makes a unique token
std::string mkUniqueToken() {
counter++;
return unsignedToDecimal(counter);
}
//Does a file exist? http://stackoverflow.com/questions/12774207
bool exists(std::string fileName) {
std::ifstream infile(fileName.c_str());
return infile.good();
}
//Reads a file: http://stackoverflow.com/questions/2602013
std::string get_file_contents(std::string filename)
{
std::ifstream in(filename.c_str(), std::ios::in | std::ios::binary);
if (in)
{
std::string contents;
in.seekg(0, std::ios::end);
contents.resize(in.tellg());
in.seekg(0, std::ios::beg);
in.read(&contents[0], contents.size());
in.close();
return(contents);
}
throw(errno);
}
//Report error
void err(std::string errtext, Metadata met) {
std::string err = "Error (file \"" + met.file + "\", line " +
unsignedToDecimal(met.ln + 1) + ", char " + unsignedToDecimal(met.ch) +
"): " + errtext;
std::cerr << err << std::endl;
throw(err);
}
//Report warning
void warn(std::string errtext, Metadata met) {
std::string err = "Warning (file \"" + met.file + "\", line " +
unsignedToDecimal(met.ln + 1) + ", char " + unsignedToDecimal(met.ch) +
"): " + errtext;
std::cerr << err << std::endl;
}
//Bin to hex
std::string binToHex(std::string inp) {
std::string o = "";
for (unsigned i = 0; i < inp.length(); i++) {
unsigned char v = inp[i];
o += std::string("0123456789abcdef").substr(v/16, 1)
+ std::string("0123456789abcdef").substr(v%16, 1);
}
return o;
}
//Hex to bin
std::string hexToBin(std::string inp) {
std::string o = "";
for (unsigned i = 0; i+1 < inp.length(); i+=2) {
char v = (char)(std::string("0123456789abcdef").find(inp[i]) * 16 +
std::string("0123456789abcdef").find(inp[i+1]));
o += v;
}
return o;
}
//Lower to upper
std::string upperCase(std::string inp) {
std::string o = "";
for (unsigned i = 0; i < inp.length(); i++) {
if (inp[i] >= 97 && inp[i] <= 122) o += inp[i] - 32;
else o += inp[i];
}
return o;
}
//Three-int vector
std::vector<int> triple(int a, int b, int c) {
std::vector<int> v;
v.push_back(a);
v.push_back(b);
v.push_back(c);
return v;
}
//Extend node vector
std::vector<Node> extend(std::vector<Node> a, std::vector<Node> b) {
for (unsigned i = 0; i < b.size(); i++) a.push_back(b[i]);
return a;
}

137
libserpent/util.h
View File

@ -1,137 +0,0 @@
#ifndef ETHSERP_UTIL
#define ETHSERP_UTIL
#include <stdio.h>
#include <iostream>
#include <vector>
#include <map>
#include <fstream>
#include <cerrno>
const int TOKEN = 0,
ASTNODE = 1,
SPACE = 2,
BRACK = 3,
SQUOTE = 4,
DQUOTE = 5,
SYMB = 6,
ALPHANUM = 7,
LPAREN = 8,
RPAREN = 9,
COMMA = 10,
COLON = 11,
UNARY_OP = 12,
BINARY_OP = 13,
COMPOUND = 14,
TOKEN_SPLITTER = 15;
// Stores metadata about each token
class Metadata {
public:
Metadata(std::string File="main", int Ln=-1, int Ch=-1) {
file = File;
ln = Ln;
ch = Ch;
fixed = false;
}
std::string file;
int ln;
int ch;
bool fixed;
};
std::string mkUniqueToken();
// type can be TOKEN or ASTNODE
class Node {
public:
int type;
std::string val;
std::vector<Node> args;
Metadata metadata;
};
Node token(std::string val, Metadata met=Metadata());
Node astnode(std::string val, std::vector<Node> args, Metadata met=Metadata());
Node astnode(std::string val, Metadata met=Metadata());
Node astnode(std::string val, Node a, Metadata met=Metadata());
Node astnode(std::string val, Node a, Node b, Metadata met=Metadata());
Node astnode(std::string val, Node a, Node b, Node c, Metadata met=Metadata());
Node astnode(std::string val, Node a, Node b,
Node c, Node d, Metadata met=Metadata());
// Number of tokens in a tree
int treeSize(Node prog);
// Print token list
std::string printTokens(std::vector<Node> tokens);
// Prints a lisp AST on one line
std::string printSimple(Node ast);
// Pretty-prints a lisp AST
std::string printAST(Node ast, bool printMetadata=false);
// Splits text by line
std::vector<std::string> splitLines(std::string s);
// Inverse of splitLines
std::string joinLines(std::vector<std::string> lines);
// Indent all lines by 4 spaces
std::string indentLines(std::string inp);
// Converts binary to simple numeric format
std::string binToNumeric(std::string inp);
// Converts string to simple numeric format
std::string strToNumeric(std::string inp);
// Does the node contain a number (eg. 124, 0xf012c, "george")
bool isNumberLike(Node node);
//Normalizes number representations
Node nodeToNumeric(Node node);
//If a node is numeric, normalize its representation
Node tryNumberize(Node node);
//Converts a value to an array of byte number nodes
std::vector<Node> toByteArr(std::string val, Metadata metadata, int minLen=1);
//Reads a file
std::string get_file_contents(std::string filename);
//Does a file exist?
bool exists(std::string fileName);
//Report error
void err(std::string errtext, Metadata met);
//Report warning
void warn(std::string errtext, Metadata met);
//Bin to hex
std::string binToHex(std::string inp);
//Hex to bin
std::string hexToBin(std::string inp);
//Lower to upper
std::string upperCase(std::string inp);
//Three-int vector
std::vector<int> triple(int a, int b, int c);
//Extend node vector
std::vector<Node> extend(std::vector<Node> a, std::vector<Node> b);
// Is the number decimal?
bool isDecimal(std::string inp);
#define asn astnode
#define tkn token
#define msi std::map<std::string, int>
#define msn std::map<std::string, Node>
#define mss std::map<std::string, std::string>
#endif

3
libweb3jsonrpc/WebThreeStubServerBase.cpp
View File

@ -41,9 +41,6 @@
#include <libethcore/CommonJS.h> #include <libethcore/CommonJS.h>
#include <libwhisper/Message.h> #include <libwhisper/Message.h>
#include <libwhisper/WhisperHost.h> #include <libwhisper/WhisperHost.h>
#if ETH_SERPENT || !ETH_TRUE
#include <libserpent/funcs.h>
#endif
#include "AccountHolder.h" #include "AccountHolder.h"
#include "JsonHelper.h" #include "JsonHelper.h"
using namespace std; using namespace std;

12
pullSerpent.sh
View File

@ -1,12 +0,0 @@
#!/bin/bash
opwd="$PWD"
cd ../serpent
git stash
git pull
git stash pop
cp bignum.* compiler.* funcs.* lllparser.* opcodes.h parser.* rewriter.* tokenize.* util.* ../cpp-ethereum/libserpent/
cp cmdline.* "$opwd/sc/"
cd "$opwd"
perl -i -p -e 's:include "funcs.h":include <libserpent/funcs.h>:gc' sc/*

13
pysol/MANIFEST.in
View File

@ -1,13 +0,0 @@
include pysol/*.cpp
include *.py
include libdevcore/*cpp
include libdevcore/*h
include libdevcrypto/*cpp
include libdevcrypto/*h
include libethcore/*cpp
include libethcore/*h
include libsolidity/*cpp
include libsolidity/*h
include libevmcore/*cpp
include libevmcore/*h
include pysol/README.md

0
pysol/README.md
View File

115
pysol/pysolidity.cpp
View File

@ -1,115 +0,0 @@
#include <Python.h>
#include "structmember.h"
#include <stdlib.h>
#include <stdio.h>
#include <iostream>
#include <vector>
#include "../libdevcore/CommonData.h"
#include <libsolidity/Compiler.h>
#include <libsolidity/CompilerStack.h>
#include <libsolidity/CompilerUtils.h>
#include <libsolidity/SourceReferenceFormatter.h>
std::string compile(std::string src) {
dev::solidity::CompilerStack compiler;
try
{
std::vector<uint8_t> m_data = compiler.compile(src, false);
return std::string(m_data.begin(), m_data.end());
}
catch (dev::Exception const& exception)
{
std::ostringstream error;
dev::solidity::SourceReferenceFormatter::printExceptionInformation(error, exception, "Error", compiler);
std::string e = error.str();
throw(e);
}
}
std::string mk_full_signature(std::string src) {
dev::solidity::CompilerStack compiler;
try
{
compiler.compile(src);
return compiler.getInterface("");
}
catch (dev::Exception const& exception)
{
std::ostringstream error;
dev::solidity::SourceReferenceFormatter::printExceptionInformation(error, exception, "Error", compiler);
std::string e = error.str();
throw(e);
}
}
std::string bob(std::string src) { return src + src; }
#define PYMETHOD(name, FROM, method, TO) \
static PyObject * name(PyObject *, PyObject *args) { \
try { \
FROM(med) \
return TO(method(med)); \
} \
catch (std::string e) { \
PyErr_SetString(PyExc_Exception, e.c_str()); \
return NULL; \
} \
}
#define FROMSTR(v) \
const char *command; \
int len; \
if (!PyArg_ParseTuple(args, "s#", &command, &len)) \
return NULL; \
std::string v = std::string(command, len); \
// Convert string into python wrapper form
PyObject* pyifyString(std::string s) {
return Py_BuildValue("s#", s.c_str(), s.length());
}
// Convert integer into python wrapper form
PyObject* pyifyInteger(unsigned int i) {
return Py_BuildValue("i", i);
}
// Convert pyobject int into normal form
int cppifyInt(PyObject* o) {
int out;
if (!PyArg_Parse(o, "i", &out))
throw("Argument should be integer");
return out;
}
// Convert pyobject string into normal form
std::string cppifyString(PyObject* o) {
const char *command;
if (!PyArg_Parse(o, "s", &command))
throw("Argument should be string");
return std::string(command);
}
int fh(std::string i) {
return dev::fromHex(i[0]);
}
PYMETHOD(ps_compile, FROMSTR, compile, pyifyString)
PYMETHOD(ps_mk_full_signature, FROMSTR, mk_full_signature, pyifyString)
static PyMethodDef PyextMethods[] = {
{"compile", ps_compile, METH_VARARGS,
"Compile code."},
{"mk_full_signature", ps_mk_full_signature, METH_VARARGS,
"Get the signature of a piece of code."},
{NULL, NULL, 0, NULL} /* Sentinel */
};
PyMODINIT_FUNC initsolidity(void)
{
Py_InitModule( "solidity", PyextMethods );
}

41
pysol/setup.py
View File

@ -1,41 +0,0 @@
import os
os.chdir('..')
from setuptools import setup, Extension
from distutils.sysconfig import get_config_vars
(opt,) = get_config_vars('OPT')
os.environ['OPT'] = " ".join(
flag for flag in opt.split() if flag != '-Wstrict-prototypes'
)
setup(
# Name of this package
name="ethereum-solidity",
# Package version
version='1.8.0',
description='Solidity compiler python wrapper',
maintainer='Vitalik Buterin',
maintainer_email='v@buterin.com',
license='WTFPL',
url='http://www.ethereum.org/',
# Describes how to build the actual extension module from C source files.
ext_modules=[
Extension(
'solidity', # Python name of the module
sources= ['libdevcore/Common.cpp', 'libdevcore/CommonData.cpp', 'libdevcore/CommonIO.cpp', 'libdevcore/FixedHash.cpp', 'libdevcore/Guards.cpp', 'libdevcore/Log.cpp', 'libdevcore/RangeMask.cpp', 'libdevcore/RLP.cpp', 'libdevcore/Worker.cpp', 'libdevcrypto/AES.cpp', 'libdevcrypto/Common.cpp', 'libdevcrypto/CryptoPP.cpp', 'libdevcrypto/ECDHE.cpp', 'libdevcrypto/FileSystem.cpp', 'libdevcrypto/MemoryDB.cpp', 'libdevcrypto/OverlayDB.cpp', 'libdevcrypto/SHA3.cpp', 'libdevcrypto/TrieCommon.cpp', 'libdevcrypto/TrieDB.cpp', 'libethcore/CommonEth.cpp', 'libethcore/CommonJS.cpp', 'libethcore/Exceptions.cpp', 'libsolidity/AST.cpp', 'libsolidity/ASTJsonConverter.cpp', 'libsolidity/ASTPrinter.cpp', 'libsolidity/CompilerContext.cpp', 'libsolidity/Compiler.cpp', 'libsolidity/CompilerStack.cpp', 'libsolidity/CompilerUtils.cpp', 'libsolidity/DeclarationContainer.cpp', 'libsolidity/ExpressionCompiler.cpp', 'libsolidity/GlobalContext.cpp', 'libsolidity/InterfaceHandler.cpp', 'libsolidity/NameAndTypeResolver.cpp', 'libsolidity/Parser.cpp', 'libsolidity/Scanner.cpp', 'libsolidity/SourceReferenceFormatter.cpp', 'libsolidity/Token.cpp', 'libsolidity/Types.cpp', 'libevmcore/Assembly.cpp', 'libevmcore/Instruction.cpp', 'pysol/pysolidity.cpp'],
libraries=['boost_python', 'boost_filesystem', 'boost_chrono', 'boost_thread', 'cryptopp', 'leveldb', 'jsoncpp'],
include_dirs=['/usr/include/boost', '..', '../..', '.'],
extra_compile_args=['--std=c++11', '-Wno-unknown-pragmas']
)],
py_modules=[
],
scripts=[
],
entry_points={
}
),

19
sc/CMakeLists.txt
View File

@ -1,19 +0,0 @@
cmake_policy(SET CMP0015 NEW)
aux_source_directory(. SRC_LIST)
include_directories(BEFORE ..)
set(EXECUTABLE sc)
add_executable(${EXECUTABLE} ${SRC_LIST})
if (NOT ("${CMAKE_CXX_COMPILER_ID}" STREQUAL "MSVC"))
target_link_libraries(${EXECUTABLE} serpent)
endif()
target_link_libraries(${EXECUTABLE} lll)
target_link_libraries(${EXECUTABLE} evmcore)
target_link_libraries(${EXECUTABLE} devcore)
install( TARGETS ${EXECUTABLE} DESTINATION bin )

129
sc/cmdline.cpp
View File

@ -1,129 +0,0 @@
#include <stdio.h>
#include <string>
#include <iostream>
#include <vector>
#include <map>
#include <libserpent/funcs.h>
int main(int argv, char** argc) {
if (argv == 1) {
std::cerr << "Must provide a command and arguments! Try parse, rewrite, compile, assemble\n";
return 0;
}
if (argv == 2 && (std::string(argc[1]) == "--help" || std::string(argc[1]) == "-h" )) {
std::cout << argc[1] << "\n";
std::cout << "serpent command input\n";
std::cout << "where input -s for from stdin, a file, or interpreted as serpent code if does not exist as file.";
std::cout << "where command: \n";
std::cout << " parse: Just parses and returns s-expression code.\n";
std::cout << " rewrite: Parse, use rewrite rules print s-expressions of result.\n";
std::cout << " compile: Return resulting compiled EVM code in hex.\n";
std::cout << " assemble: Return result from step before compilation.\n";
return 0;
}
std::string flag = "";
std::string command = argc[1];
std::string input;
std::string secondInput;
if (std::string(argc[1]) == "-s") {
flag = command.substr(1);
command = argc[2];
input = "";
std::string line;
while (std::getline(std::cin, line)) {
input += line + "\n";
}
secondInput = argv == 3 ? "" : argc[3];
}
else {
if (argv == 2) {
std::cerr << "Not enough arguments for serpent cmdline\n";
throw(0);
}
input = argc[2];
secondInput = argv == 3 ? "" : argc[3];
}
bool haveSec = secondInput.length() > 0;
if (command == "parse" || command == "parse_serpent") {
std::cout << printAST(parseSerpent(input), haveSec) << "\n";
}
else if (command == "rewrite") {
std::cout << printAST(rewrite(parseLLL(input, true)), haveSec) << "\n";
}
else if (command == "compile_to_lll") {
std::cout << printAST(compileToLLL(input), haveSec) << "\n";
}
else if (command == "rewrite_chunk") {
std::cout << printAST(rewriteChunk(parseLLL(input, true)), haveSec) << "\n";
}
else if (command == "compile_chunk_to_lll") {
std::cout << printAST(compileChunkToLLL(input), haveSec) << "\n";
}
else if (command == "build_fragtree") {
std::cout << printAST(buildFragmentTree(parseLLL(input, true))) << "\n";
}
else if (command == "compile_lll") {
std::cout << binToHex(compileLLL(parseLLL(input, true))) << "\n";
}
else if (command == "dereference") {
std::cout << printTokens(dereference(parseLLL(input, true))) <<"\n";
}
else if (command == "pretty_assemble") {
std::cout << printTokens(prettyAssemble(parseLLL(input, true))) <<"\n";
}
else if (command == "pretty_compile_lll") {
std::cout << printTokens(prettyCompileLLL(parseLLL(input, true))) << "\n";
}
else if (command == "pretty_compile") {
std::cout << printTokens(prettyCompile(input)) << "\n";
}
else if (command == "pretty_compile_chunk") {
std::cout << printTokens(prettyCompileChunk(input)) << "\n";
}
else if (command == "assemble") {
std::cout << assemble(parseLLL(input, true)) << "\n";
}
else if (command == "serialize") {
std::cout << binToHex(serialize(tokenize(input, Metadata(), false))) << "\n";
}
else if (command == "deserialize") {
std::cout << printTokens(deserialize(hexToBin(input))) << "\n";
}
else if (command == "compile") {
std::cout << binToHex(compile(input)) << "\n";
}
else if (command == "compile_chunk") {
std::cout << binToHex(compileChunk(input)) << "\n";
}
else if (command == "encode_datalist") {
std::vector<Node> tokens = tokenize(input);
std::vector<std::string> o;
for (int i = 0; i < (int)tokens.size(); i++) {
o.push_back(tokens[i].val);
}
std::cout << binToHex(encodeDatalist(o)) << "\n";
}
else if (command == "decode_datalist") {
std::vector<std::string> o = decodeDatalist(hexToBin(input));
std::vector<Node> tokens;
for (int i = 0; i < (int)o.size(); i++)
tokens.push_back(token(o[i]));
std::cout << printTokens(tokens) << "\n";
}
else if (command == "tokenize") {
std::cout << printTokens(tokenize(input));
}
else if (command == "biject") {
if (argv == 3)
std::cerr << "Not enough arguments for biject\n";
int pos = decimalToUnsigned(secondInput);
std::vector<Node> n = prettyCompile(input);
if (pos >= (int)n.size())
std::cerr << "Code position too high\n";
Metadata m = n[pos].metadata;
std::cout << "Opcode: " << n[pos].val << ", file: " << m.file <<
", line: " << m.ln << ", char: " << m.ch << "\n";
}
}
Write Preview
Loading…
Cancel
Save