ethereum / serpent Goto Github PK

License: Other

Makefile 0.80% C++ 94.23% Python 4.78% C 0.19%

serpent's Introduction

Introduction

Serpent is an assembly language that compiles to EVM code that is extended with various high-level features. It can be useful for writing code that requires low-level opcode manipulation as well as access to high-level primitives like the ABI.

Being a low-level language, Serpent is NOT RECOMMENDED for building applications unless you really really know what you're doing. The creator recommends Solidity as a default choice, LLL if you want close-to-the-metal optimizations, or Viper if you like its features though it is still experimental.

Installation:

make && sudo make install

Testing

Testing is done using pytest and tox.

$ pip install tox -r requirements-dev.txt

To run the test suite in your current python version:

$ py.test

To run the full test suite across all supported python versions:

$ tox

serpent's People

Contributors

Stargazers

Watchers

Forkers

jaytoday guyz grimderp heikoheiko vmatekole tjade273 cyberco glaspe mind-blank kenkappler joeykrug ethers rfikki reorder cdetrio counterpartyxcp hfeeki davidwan csbitcoin whitj00 gitter-badger huangjiehua hiroshi1tanaka fierozen zebbra2014 chfast janx ambadruddin pipermerriam rubensayshi hackaugusto nneerr123 pbnj-io quasi-dev j1o1h1n chriscalderon ethereumxt wintermooch simulationcoin alphablockchain jingqiangliu arasharchor enumatech rhodiumlabs brndnblck serpentchris keeeeeeeks konark-cliqz whoshy sffej balgan avineshwar chaneybenjamini my2house bionicles raina2017 chenrui2014 come-maiz beaugunderson stanxii dkbradley etherbanking psdgithub amitshah changwu-tw adozenlines hwwhww wanghs09 myneworder rue-foundation dsaplbg sriharikapu fancy2110 usrcoin-interesting redsquirrel kylingood dongsam stnbu lrettig mswezey23 tokenaires aanwhs baophucct daeplatform ysun94 mahendarkarthikeya crypto0sy80 roninkaizen fsk783529 m4nh puppet1982 tubbz-alt jo-cho redrarri43 ray-chunkit-chung isabella232 pinchaschaim keccak384 adimi93 dima-brook

serpent's Issues

how to return array containing an array

how to do this:
return([firstElemNotArray, secondElemWhichIsAnArray], 2)

behavior when refactoring inset to extern

sand2.py has:

data heaviestBlock

def bar(n):
    log(111)
    log(self.heaviestBlock)
    self.heaviestBlock = n*10

    log(222)
    log(self.heaviestBlock)

sand3.py has:

extern two: [bar:i:_]
TWO = create('sand2.py')

def foo(n):
    TWO.bar(n)

output is:

>>> c=s.abi_contract('sand3.py')
>>> c.foo(3)
('LOG', '511025535e12a356197fa0c1c2a53a6617186f93', [111], [])
('LOG', '511025535e12a356197fa0c1c2a53a6617186f93', [0], [])
('LOG', '511025535e12a356197fa0c1c2a53a6617186f93', [222], [])
('LOG', '511025535e12a356197fa0c1c2a53a6617186f93', [30L], [])
>>> c.foo(4)
('LOG', '83891c69d6a98ba3a83cc6b9dab025b4c5ad2325', [111], [])
('LOG', '83891c69d6a98ba3a83cc6b9dab025b4c5ad2325', [0], [])
('LOG', '83891c69d6a98ba3a83cc6b9dab025b4c5ad2325', [222], [])
('LOG', '83891c69d6a98ba3a83cc6b9dab025b4c5ad2325', [40L], [])

If inset was used and extern were not used, c.foo(4) would log 30 after 111:

>>> c.foo(4)
('LOG', '83891c69d6a98ba3a83cc6b9dab025b4c5ad2325', [111], [])
('LOG', '83891c69d6a98ba3a83cc6b9dab025b4c5ad2325', [30], [])

I am opening this suggestion because it would be nice if replacing code that used inset, would still behave similarly when using extern.

undefined reference error when compiling latest code (0e4f5) on Ubuntu 14.04

After running "make":

cmdline.o: In function `main':
cmdline.cpp:(.text+0xcd2): undefined reference to `tokenize(std::string, Metadata)'

Apologies if this is user error on my part...

mk_web3_signature returns invalid json on argumentless functions

The returned JSON for mk_web3_signature is invalid when there are functions without arguments:

$ serpent mk_web3_signature examples/schellingcoin/quicksort_pairs.se
[{
    name: 'kall',
    inputs: [
    outputs: [{ name: 'out', type: 'string' }]
}]

No JUMPDESTs

poc7: accessing property of a struct in an array

data a[](b, c)
def foo():
    self.a[0].b = 33
    self.a[0].c = 55
    return(self.a[0].b)

EXPECTED is 33
OBSERVED is 55. (always return the value of the last assignment)

Compiler doesnt spot nonexistant variables.

This means a simple typo can have consequences that are only spotted later when the value matters, instead of immediately. (return(x) to see it)

Probably it allocates memory for the variable meaning the value starts at zero.

Private functions? The danger of exposing functions?!

Is there a way to make private functions? Might if msg.sender == self: suffice?

Infact accidentally not having some functions not private can be downright disasterous. For instance, i am upgrading my assurance contracts to the newer serpent/Ethereum, and i have a refund() option for the creator, and a finish() function that decides whether to refunds.

God forbid that someone decides to refactor the internals, _refund(), with no checking on whois asking. I think this warrants a warning.. Infact, perhaps functions need to be defaultly private?!

unable to modify string

def f2(string:s):
    setch(string, 0, "a")
    setch(string, 1, "b")

def t2():
    string = text("cd")
    self.f2(string)
    return([getch(string,0), getch(string,1)]:a)  # should return [97,98]

when calling t2()
actual: [99,100]
expected: [97,98]

Does ``return`` only return a single 256 bit word?

Note: The example uses serpent on the develop branch with revision c9563cff.

The following example tries to return a long text string. An examination of the LLL output makes it seem as if only 32 bytes are returned in the expression (return (ref '_temp_131) 32). Am I misreading this LLL expression?

$ cat > long_string_return.se
return text("This is 53 characters, which is longer than 256 bits.")

$ serpent compile_to_lll !$
serpent compile_to_lll long_string_return.se
(return 0
  (lll
    (seq
      (set '_temp_131
        (with '_temp17_str (alloc 85)
          (seq
            (mstore (get '_temp17_str) 53)
            (mstore (add (get '_temp17_str) 32)
              38178759162904981153227459547417441039765882805204013944776380134389117973280
            )
            (mstore (add (get '_temp17_str) 64)
              49046669650839314774563824355363137219616139645592700948219339121638676365312
            )
            (add (get '_temp17_str) 32)
          )
        )
      )
      (return (ref '_temp_131) 32)
    )
    0
  )
)

Unexpected array behaviour

this code:

a = array(1)
a[0] = 1
return(a, 1)

returns 0 (unexpected)

this code:

a = array(1)
something = 2
a[0] = 1
return(a, 1)

for some reason returns 1 (expected)

Name: ethereum-serpent
Version: 1.6.3

poc9 ECRECOVER gas

ECRECOVER calls still happen with a 500 gas (instead of 3000)
https://github.com/ethereum/serpent/blob/develop/rewriter.cpp#L286

I have created a (failing) test case for this ethereum/pyethereum#219

modifying array in another function

def f2(arr:a):
    arr[0] = 1
    arr[1] = 2

def t2():
    a = array(2)
    a[0] = 3
    a[1] = 4
    self.f2(a)
    return(a:a)

when calling t2()
actual: [3,4]
expected: [1,2]

return([1,2,3], items=3) # No array returned latest devel commit

Hi, thanks for your great work Vitalik. I am using serpent for some exchange contracts. I encountered the problem that functions don't return arrays but integers. I am working on the latest develop commit.

def return_array():
......return([1,2,3], items=3)

def main():
.....log(self.return_array()) # logs ('LOG', 'c305c901078781c232a2a521c2af7980f8385ee9', [1L], [])

examples/quicksort_pairs.py gives wrong results

Sorting does not work as expected. Afik, this issue is not related to #8.

from pyethereum import tester
def test_quicksort():
    s = tester.state()
    c = s.contract('quicksort_pairs.se')
    data = [30, 1, 90, 2, 70, 3, 50, 4]
    expected = [30, 1, 50, 4, 70, 3, 90, 2]
    r = s.send(tester.k0, c, 0, data=data)
    assert expected == r, (expected, r)
test_quicksort()

(default) ~/dev/ethereum/fuelbattles/serpent (git)-[master] % python test_quicksort_pairs.py 
Traceback (most recent call last):
  File "test_quicksort_pairs.py", line 11, in <module>
    test_quicksort()
  File "test_quicksort_pairs.py", line 9, in test_quicksort
    assert expected == r, (expected, r)
AssertionError: ([30, 1, 50, 4, 70, 3, 90, 2], [30, 1, 0, 0, 70, 3, 90, 2])

space saver

when serpent compiles, it gives things like: PUSH2 0 125 JUMP
more efficient equivalent EVM code: PUSH1 125 JUMP

idk where in the code to fix it from.
The trick to tell how many bytes you need to store number N: log(N)/log(256)

storing an array to storage

what's the recommended way to do this?

data gArr[]

def f2():
    arr = array(2)
    arr[0] = 1
    arr[1] = 2
    return(arr:a)

def t2():
    self.gArr = self.f2(outsz=2)
    # return(self.gArr:a)

As is, it's a compile error on the line with self.gArr: Too few array index lookups

Padding zeros of strings at the beginning rather than the end

JS API of cpp-ethereum pads at the end, seems like that makes more sense to me.

setch not working?

def f1(string:s):
    setch(string, 0, "a")
    setch(string, 1, "b")
    return(string)

def t1():
    string = text("cd")
    res = self.f1(string)
    return([getch(res,0), getch(res,1)]:a)  # should return [97,98]

when call t1()
actual: [99, 100]
expected: [97, 98]

help installing serpent please

hi, can anyone link me to a step by step guide for installing/testing Serpent? I have a VM Ubuntu box set up already. I'm really new to all this, but really excited to play around with the technology! Please help!

0xfe invalid opcode should be removed?

I have a contract that when compiled has some parts JUMPI 0xfe. ~~Alethzero does not allow contract creation with invalid opcode 254~~:

<INVALID_INSTRUCTION: 254 >
 Dynamic exception type: std::out_of_range
std::exception::what: map::at

For a repro, paste this in AZ "New Transaction" 6006600957fe5b60035b6004 and click Debug. Look at AZ log for error above.

inset() is changing behavior of function that writes to storage

When inset() is not used, calling testingonlySetGenesis() in code below writes data to storage and things are good.

However, when top of file is placed in another file and inset() is used (as mentioned in code comment), testingonlySetGenesis does not store data: it is also unclear what happens because this tx cannot be debugged by Alethzero (AZ shows blank http://imgur.com/DMArfoC )

data numAncestorDepths
self.numAncestorDepths = 9  # if change, look at defn of ancestor_depths and block in btcrelay.py
data ancestor_depths[9]

self.ancestor_depths[0] = 1
self.ancestor_depths[1] = 4
self.ancestor_depths[2] = 16
self.ancestor_depths[3] = 64
self.ancestor_depths[4] = 256
self.ancestor_depths[5] = 1024
self.ancestor_depths[6] = 4096
self.ancestor_depths[7] = 16384
self.ancestor_depths[8] = 65536
########### BUG IS IF ABOVE IS PUT IN ANOTHER FILE and replaced with inset() here ######

# block with the highest score (the end of the blockchain)
data heaviestBlock

# highest score among all blocks (so far)
data highScore

# note: _ancestor[9]
data block[2^256](_height, _score, _ancestor[9], _blockHeader[])

extern btc_eth: [processTransfer:s:i]


#TODO for testing only
def testingonlySetGenesis(blockHash):
    self.heaviestBlock = blockHash
    self.block[blockHash]._height = 1
    ancLen = self.numAncestorDepths
    i = 0
    while i < ancLen:
        self.block[blockHash]._ancestor[i] = blockHash
        i += 1

getting length of array

since the calling syntax is foo(array:size) maybe inside foo(), the length of array can be obtained by a syntax like len(array)

Segmentation fault when rewriting

On my OS X Mavericks laptop, the PoC7 version of Serpent crashes with a segmentation fault when compiling this example code:

http://ether.fund/contract/e9fc5/crowdfund-in-serpent-20

Stacktrace:

http://pastebin.com/qH7Uz0WL

Compiled Serpent subcurrency example turns into STOP STOP STOP in AlethZero

Taking a Serpent compiled contract, submitting it to AlethZero by copy/pasting the hexstring into the "(Create Contract)" Data field returns into a persisted contract with a lot of STOP STOP STOP instructions up front... which won't execute.

Serpent sourcecode:

init:
    contract.storage[msg.sender] = 1000000
code:
    if msg.datasize == 1:
        addr = msg.data[0]
        return(contract.storage[addr])
    else:
        from = msg.sender
        fromvalue = contract.storage[from]
        to = msg.data[0]
        value = msg.data[1]
        if fromvalue >= value:
            contract.storage[from] = fromvalue - value
            contract.storage[to] = contract.storage[to] + value
            return(1)
        else:
            return(0)

Compiling with serpent compile (serpent version 51cd8f4)

$ serpent compile examples/subcurrency.se
620f424033576100775160006100143961008b586001602036040e0f610022596000356020546020535660405460206040f261007758336060546060535660805460003560a05460203560c05460c0536080530c0f0f61006b5960c053608053036060535760c05360a053560160a05357600160e054602060e0f2610077586000610100546020610100f26000f2

Or the pretty version:

$ serpent pretty_compile examples/subcurrency.se
[PUSH3, 15, 66, 64, CALLER, SSTORE, PUSH2, 0, 119, DUP, PUSH1, 0, PUSH2, 0, 20, CODECOPY, PUSH2, 0, 139, JUMP, PUSH1, 1, PUSH1, 32, CALLDATASIZE, DIV, EQ, NOT, PUSH2, 0, 34, JUMPI, PUSH1, 0, CALLDATALOAD, PUSH1, 32, MSTORE, PUSH1, 32, MLOAD, SLOAD, PUSH1, 64, MSTORE, PUSH1, 32, PUSH1, 64, RETURN, PUSH2, 0, 119, JUMP, CALLER, PUSH1, 96, MSTORE, PUSH1, 96, MLOAD, SLOAD, PUSH1, 128, MSTORE, PUSH1, 0, CALLDATALOAD, PUSH1, 160, MSTORE, PUSH1, 32, CALLDATALOAD, PUSH1, 192, MSTORE, PUSH1, 192, MLOAD, PUSH1, 128, MLOAD, SLT, NOT, NOT, PUSH2, 0, 107, JUMPI, PUSH1, 192, MLOAD, PUSH1, 128, MLOAD, SUB, PUSH1, 96, MLOAD, SSTORE, PUSH1, 192, MLOAD, PUSH1, 160, MLOAD, SLOAD, ADD, PUSH1, 160, MLOAD, SSTORE, PUSH1, 1, PUSH1, 224, MSTORE, PUSH1, 32, PUSH1, 224, RETURN, PUSH2, 0, 119, JUMP, PUSH1, 0, PUSH2, 1, 0, MSTORE, PUSH1, 32, PUSH2, 1, 0, RETURN, PUSH1, 0, RETURN]

The contract initialization seems to work, but then the contract is rendered useless by the STOPs, trailing instructions of the contract are also trimmed.

(This issue was previously reported on the now deleted ethereum/langs repository)

byte() is not working on loaded or spliced strings

Problem is in the code and comment bellow:

data buf[]

def foo():
    mystr=text("81cd02ab7e569e8bcd9317e2fe99f2de44d49ab2b8851ba4a308000000000000e320b6c2fffc8d750423db8b1eb942ae710e951ed797f7affc8892b0f1fc122b81cd02ab7e569e8bcd9317e2fe99f2de44d49ab2b8851ba4a308000000000000e320b6c2fffc8d750423db8b1eb942ae710e951ed797f7affc8892b0f1fc122b")
    save(self.buf[0], mystr, chars=len(mystr))

    self.g()

def g():
    jstr = load(self.buf[0], chars=132)
    log(datastr=jstr)

    j = slice(jstr, chars=128, chars=132)
    log(datastr=j)

    b = byte(0, jstr)  # gets 0 when jstr or j is used.  byte only works when self.buf[0] is used instead of jstr
    log(b)

poc7 serpent not recognizing 'if' statement

When trying to compile the stock short_namecoin.se example with the POC7 branch of serpent:

$ serpent compile short_namecoin.se
Error (file "short_namecoin.se", line 2, char 720899): Not a function or opcode: if

Char numbering seems off as well.

Using log(data=a) with an array prints each byte, not each item

If my array is [1, 2, 3], shouldn't log print [1, 2, 3], and not [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3]?

how to save string to global

data buf

def foo():
    mystr = text("01ab")
    save(self.buf, mystr:str)

combinations of using buf[] and save(self.buf, mystr, chars=4) also do not work

Returning array with first element as 0 fails

The following serpent test code fails when the first returned array element is a 0:

def get_stops(rnd):
    stops = array(REEL_COUNT)
    i = 0
    while i < REEL_COUNT:
        stops[i] = rnd % REEL_POSITIONS
        rnd = rnd / REEL_POSITIONS
        i += 1
    return(stops:a)

Full serpent + test code: https://gist.github.com/jorisbontje/99a30029a3d15f065e13/757dded2ae95835010b6c09f1878a951a324278e

_______________________________________________________ TestStops.test_get_stops ________________________________________________________

self = <test_stops.TestStops object at 0x110ce0490>

    def test_get_stops(self):
        assert self._get_stops(23888)[:3] == [16, 10, 23]
        assert self._get_stops(1606)[:3] == [6, 18, 1]
        assert self._get_stops(23888)[:3] == [16, 10, 23]
>       assert self._get_stops(30464)[:3] == [0, 24, 29]
E       assert [] == [0, 24, 29]
E         Right contains more items, first extra item: 0
E         Full diff:
E         - []
E         + [0, 24, 29]

test/test_stops.py:35: AssertionError

This is with serpent 1.7.8

Expected would be that the array [0, 24, 29] would be returned.

global variable scoping/modifying

see inline comment in code below:

data pos
self.pos=0


def readVarintNum():
    log(self.pos)  # shouldn't self.pos be 4 here ?  or how can I update self.pos from readUInt32LE() ?

def readUInt32LE():
    self.pos += 4

def run():
    self.pos = 0
    self.readUInt32LE()
    log(self.pos)
    ins = self.readVarintNum()

Potential optimizations

Compiler macros are general functions that apply on an expression first, and return nothing if they do not apply, or the resulting expression.

A lot can already be achieved with the replacement rules already, of course.

Find-the-function.

Perhaps instead of having a lot of (when (eq (ge '__funid) ...) you can see if (lt (ge '__funid ...) < ) and go through branches of the Tree. Is log2(N) instead of N, no disadvantage that i see.

I think it is probably best if LLL has a (case ...) and then you make case figure it out.

`unless (iszero..)`

What's wrong with when?

Orring:

As observed.(whitespace altered)

(with '_temp_10201 (slt (get 'i) 1) 
    (if (get '_temp_10201) 
        (get '_temp_10201) 
        (sgt (get 'i) (sload 5))))

This might be fine for a value-or, returning the first non-false value. But if the value is just used as a boolean, should use the bitwise or. For simpler cases compiler macros can recognize it.

Memory values

Mark variables that the compiler makes and you know wont change.(For instance in the above or case.)

When can variables be substituted with their expressions?

Variable is never set before that.
Expressions are constant. (no call, mload, sload(AFAIK) although you might be able to check whether the latter two matter in this case)
Expressions do not make changes. (no call, sstore, mstore(AFAIK))

If you know when substituting is allowed, choosing whether to do it is figuring out what is more efficient.

I think i.e. (calldataload ..constant..), (number), (timestamp), constant all qualify if the variable is not changed. This might already help a lot!

More complicatedly, could estimate the gas cost. Either for contract creation or when running the code. (can 'settings' wrt optimization at branch in the AST-tree. I.e (optimize-for-run and (optimize-for-codelen?) Or (optimization-ratio r, i.e. saving r running gas is worth 1 creation-gas.

Arrays

Not sure.. If there is just one array with no particular length you can just use (+ (* itemsz i) last_index) instead of going the sha3 route?

Or do low indexes cause pain? ("continuous area?") Another argument against is that it makes where-things-are-stored less predictable. (dont think that is very important tho)

Powers of two

That is 2²²⁴ ... Probably better to denote that way?

(with '__funid 
    (div (calldataload 0)
        26959946667150639794667015087019630673637144422540572481103610249216)
    ....

Right EXP is not expensive. Could be cheaper though, if there was a EXP2

Calling any extern method keeps asking to use the "as" keyword

Even if the "as" keyword is already being used, serpent ignores the extern definition and returns "_" as signature, yielding the Method x has multiple definitions. Please use the "as" keyword... error

global string value is lost

the log in foo() shows correct, but the log in g() is empty... ?

data buf

def foo():
    mystr = text("01ab")
    self.buf = mystr
    log(datastr=self.buf)
    self.g()

def g():
    log(datastr=self.buf)  # this logs empty

Add NatSpec support

poc7 empty def on crowdfund example

Using poc7 build of serpent on the stock crowdfund example:

$ serpent compile crowdfund.se
Error (file "crowdfund.se", line 1, char 107): Empty def

crowdfund.se example gives: Multiple expressions or unclosed bracket

$ serpent compile crowdfund.se
Error (file "crowdfund.se", line 1, char 107): Multiple expressions or unclosed bracket
Traceback (most recent call last):
  File "/usr/local/bin/serpent", line 9, in <module>
    load_entry_point('ethereum-serpent==1.7.3', 'console_scripts', 'serpent')()
  File "/usr/local/bin/serpent.py", line 192, in main
    o = globals()[cmd](*args)
  File "/usr/local/bin/serpent.py", line 74, in <lambda>
    compile = lambda x: pyext.compile(x)
Exception: Error (file "crowdfund.se", line 1, char 107): Multiple expressions or unclosed bracket

broken compilation

With ethereum-serpent==1.78:

> serpent compile 'return(msg.data[0] * 2)'
Error (file "main", line 1, char 10): Invalid object member (ie. a foo.bar not mapped to anything)
Traceback (most recent call last):
  File "/usr/bin/serpent", line 9, in <module>
    load_entry_point('ethereum-serpent==1.7.8', 'console_scripts', 'serpent')()
  File "/usr/bin/serpent.py", line 197, in main
    o = globals()[cmd](*args)
  File "/usr/bin/serpent.py", line 74, in <lambda>
    compile = lambda x: pyext.compile(x)
Exception: Error (file "main", line 1, char 10): Invalid object member (ie. a foo.bar not mapped to anything)

inset at top of file doesn't work

inset('bar.py')

# this is what's in bar.py
# def g(n):
#     return(n+10)
#
# def f(n):
#     return(n*2)


def foo():
    res = self.g(12)

    return(res)

On my machine, the return value is always [10], regardless of which function is called.

Moving inset() to the bottom of the file is a workaround.

log with data arrays not copying content

def test_log_data():
    log(data=[1,2,3])

This returns:

('LOG', 'c305c901078781c232a2a521c2af7980f8385ee9', [], [0, 0, 0])

instead of [1, 2, 3]

Full test code https://github.com/jorisbontje/sleth/blob/master/test/test_log.py

unable to return a string from a function

def getString():
    str = text("abc")
    return(str)  # cannot use str:s here since Exception Transaction failed

def run():
    res = self.getString()
    return(getch(res, 0))

when calling run()
actual: 0
expected: 97

value of variable is clobbered

call test2() and the expected output is to log 2 twice, but numAncestorDepths gets clobbered.

data block[2^256](_blockHeader(_prevBlock))

data numAncestorDepths
def initAncestorDepths():
    self.numAncestorDepths = 2

def testStoreB(number, blockHash, hashPrevBlock):
    self.block[blockHash]._blockHeader._prevBlock = hashPrevBlock

    log(self.numAncestorDepths)


def test2():
    self.initAncestorDepths()
    self.testStoreB(45, 45, 44)
    self.testStoreB(46, 46, 45)  # this will log 45 (expected is to still log 2)

no `PyExt.compile`?

Get this with all of PyExt(master, 51bd6f2 and 1df7914) and 0.7 here. Got this trying to run the test of cryptocoinwatch. The following reproduces for me.(which particular file doesnt matter)

from serpent import compile
file = '/home/jasper/oproj/ethereum/contracts/cryptocoinwatch/cryptocoinwatch.se'
print(compile(open(file).read()).encode('hex'))

Searching the code of PyExt, i dont see compile in there, and checked that pip install pyext gets the same 0.7 files.

serpent abi json output missing 'type' for functions

each definition should include a type (function or event).

Compiled serpent code with insets doesn't work in js abi

When I use the JS API to connect to alethzero I can publish a compiled hex serpent contract without insets and a simple function call like double(2) will return 4. If I have an inset, even if the inset doesn't do anything, it returns 0

Loops of Length 311 are Too Big for init Functions

I am having problems with using while loops in def init(): functions in Contracts.

Specifically, I have installed the following versions of serpent and pyethereum:

pip install -q --upgrade \
  git+git://github.com/ethereum/serpent.git@86649c54c209a662c943d664819fe01b8ebc3e2b \
  git+git://github.com/ethereum/pyethereum.git@23c7d1e551c3c94d11fdea66cfdb7d29b17b8a50

Here's my contract:

# array_init.se
data arr[312]

def init():
    i = 0
    while i < 311:
        self.arr[i] = 1
        i = i + 1

def foo():
    return 1

312 happens to be the length of the buffer used by Mersenne Twister.

If I replace 311 with 310, there are no problems.

Here's how I am testing this code:

from pyethereum import tester as t
s = t.state()
c = s.contract('array_init.se')
for _ in range(5):
    gas_start = s.block.gas_used
    print s.send(t.k0, c, 0, funid=0, abi=[])[0]

Should print 1 five times. Instead here's my exception and stacktrace:

---------------------------------------------------------------------------
Exception                                 Traceback (most recent call last)
<ipython-input-37-fa2ff0671689> in <module>()
      1 from pyethereum import tester as t
      2 s = t.state()
----> 3 c = s.contract('array_init.se')
      4 for _ in range(5):
      5     gas_start = s.block.gas_used

/usr/local/lib/python2.7/site-packages/pyethereum/tester.pyc in contract(self, code, sender, endowment)
     60     def contract(self, code, sender=k0, endowment=0):
     61         evm = serpent.compile(code)
---> 62         o = self.evm(evm, sender, endowment)
     63         assert len(self.block.get_code(o)), "Contract code empty"
     64         return o

/usr/local/lib/python2.7/site-packages/pyethereum/tester.pyc in evm(self, evm, sender, endowment)
     70         (s, a) = pb.apply_transaction(self.block, tx)
     71         if not s:
---> 72             raise Exception("Contract creation failed")
     73         return a
     74 

Exception: Contract creation failed

Method signatures are incompatible with ethereum.js

By that I mean the 4 bytes "prefix", which should not be from sha3(method_name + input_signatures) but simply from sha3(method_name). The "as" keyword should also be brought back instead of using input signatures to differentiate methods for externs.

Inconsistency in (string) padding between serpent/pyethereum and cpp-ethereum/AlethZero

With Serpent 1.5.4:

>>> serpent.encode_datalist(["set", 1, 2])
'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00
\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00set
\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00
\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01
\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00
\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x02'

With AlethZero (POC5 and develop):

With Serpent the string "set" is left padded with NULLs. On AlethZero strings are right padded.
I would expect the padding to be the same for compatibility reasons.

collatz example is broken

it refers to msg.data, but msg.data isn't supported (or is it?)
the first equality if (x % 2) == 0 does not parse, it needs to be if ((x % 2) == 0).
it's the first example in the directory. heh.

Unable to compile develop branch

I'm trying to compile develop. When I run make I get the following error (running Xubuntu 14.04, previous versions have compiled successfully):

jack@heavy:~/src/serpent$ make
g++ -fPIC   -c -o bignum.o bignum.cpp
g++ -fPIC   -c -o util.o util.cpp
g++ -fPIC   -c -o tokenize.o tokenize.cpp
g++ -fPIC   -c -o lllparser.o lllparser.cpp
g++ -fPIC   -c -o parser.o parser.cpp
g++ -fPIC   -c -o opcodes.o opcodes.cpp
g++ -fPIC   -c -o optimize.o optimize.cpp
g++ -fPIC   -c -o functions.o functions.cpp
g++ -fPIC   -c -o rewriteutils.o rewriteutils.cpp
g++ -fPIC   -c -o preprocess.o preprocess.cpp
g++ -fPIC   -c -o rewriter.o rewriter.cpp
g++ -fPIC   -c -o compiler.o compiler.cpp
g++ -fPIC   -c -o funcs.o funcs.cpp
g++ -fPIC   -c -o cmdline.o cmdline.cpp
cmdline.cpp: In function ‘int main(int, char**)’:
cmdline.cpp:62:54: error: ‘compileChunkToLLL’ was not declared in this scope
         std::cout << printAST(compileChunkToLLL(input), haveSec) << "\n";
                                                      ^
cmdline.cpp:83:58: error: ‘prettyCompileChunk’ was not declared in this scope
         std::cout << printTokens(prettyCompileChunk(input)) << "\n";
                                                          ^
cmdline.cpp:98:49: error: ‘compileChunk’ was not declared in this scope
         std::cout << binToHex(compileChunk(input)) << "\n";
                                                 ^
make: *** [cmdline.o] Error 1