From aa042ccb8777c15c871a533f65e9cc760de1e01c Mon Sep 17 00:00:00 2001
From: Alex Beregszaszi <alex@rtfs.hu>
Date: Fri, 11 Dec 2020 17:00:28 +0000
Subject: [PATCH] Disallow leading zeroes in sized-types

Also avoid using boost::lexical_cast for parseSize.
---
 Changelog.md                                  |  1 +
 liblangutil/Token.cpp                         | 43 +++++++++++--------
 .../syntaxTests/types/weird_sized_types.sol   |  1 -
 3 files changed, 27 insertions(+), 18 deletions(-)

diff --git a/Changelog.md b/Changelog.md
index fa01287e4..c9f4d2318 100644
--- a/Changelog.md
+++ b/Changelog.md
@@ -26,6 +26,7 @@ Bugfixes:
  * SMTChecker: Fix internal error when trying to generate counterexamples with old z3.
  * SMTChecker: Fix segmentation fault that could occur on certain SMT-enabled sources when no SMT solver was available.
  * Type Checker: ``super`` is not available in libraries.
+ * Type Checker: Disallow leading zeroes in sized-types (e.g. ``bytes000032``), but allow them to be treated as identifiers.
  * Yul Optimizer: Fix a bug in NameSimplifier where a new name created by NameSimplifier could also be created by NameDispenser.
  * Yul Optimizer: Removed NameSimplifier from optimization steps available to users.
 
diff --git a/liblangutil/Token.cpp b/liblangutil/Token.cpp
index 28180514a..31f1f310e 100644
--- a/liblangutil/Token.cpp
+++ b/liblangutil/Token.cpp
@@ -41,7 +41,6 @@
 // along with solidity.  If not, see <http://www.gnu.org/licenses/>.
 
 #include <liblangutil/Token.h>
-#include <boost/range/iterator_range.hpp>
 #include <map>
 
 using namespace std;
@@ -126,22 +125,6 @@ int precedence(Token tok)
 }
 #undef T
 
-namespace
-{
-int parseSize(string::const_iterator _begin, string::const_iterator _end)
-{
-	try
-	{
-		int m = boost::lexical_cast<int>(boost::make_iterator_range(_begin, _end));
-		return m;
-	}
-	catch(boost::bad_lexical_cast const&)
-	{
-		return -1;
-	}
-}
-}
-
 static Token keywordByName(string const& _name)
 {
 	// The following macros are used inside TOKEN_LIST and cause non-keyword tokens to be ignored
@@ -162,6 +145,32 @@ bool isYulKeyword(string const& _literal)
 
 tuple<Token, unsigned int, unsigned int> fromIdentifierOrKeyword(string const& _literal)
 {
+	// Used for `bytesM`, `uintM`, `intM`, `fixedMxN`, `ufixedMxN`.
+	// M/N must be shortest representation. M can never be 0. N can be zero.
+	auto parseSize = [](string::const_iterator _begin, string::const_iterator _end) -> int
+	{
+		// No number.
+		if (distance(_begin, _end) == 0)
+			return -1;
+
+		// Disallow leading zero.
+		if (distance(_begin, _end) > 1 && *_begin == '0')
+			return -1;
+
+		int ret = 0;
+		for (auto it = _begin; it != _end; it++)
+		{
+			if (*it < '0' || *it > '9')
+				return -1;
+			//  Overflow check. The largest acceptable value is 256 in the callers.
+			if (ret >= 256)
+				return -1;
+			ret *= 10;
+			ret += *it - '0';
+		}
+		return ret;
+	};
+
 	auto positionM = find_if(_literal.begin(), _literal.end(), ::isdigit);
 	if (positionM != _literal.end())
 	{
diff --git a/test/libsolidity/syntaxTests/types/weird_sized_types.sol b/test/libsolidity/syntaxTests/types/weird_sized_types.sol
index 83f9362e0..563778d7e 100644
--- a/test/libsolidity/syntaxTests/types/weird_sized_types.sol
+++ b/test/libsolidity/syntaxTests/types/weird_sized_types.sol
@@ -40,4 +40,3 @@ contract C {
     uint uint300 = 0;
 }
 // ----
-// ParserError 2314: (22-29): Expected identifier but got 'bytes1'