Hashmal - Transaction Script IDE (alpha)

Kefkius

member

Activity: 64

Merit: 20

Quote from: JackH on December 31, 2015, 09:24:07 AM

By inclusion I just ment on how the entire path from op_code creation to block inclusion by a miner would look. But I guess there also needs to be a limit to how much you can keep expanding it, without ending up with a full copy of Bitcoin

Yeah, I see what you mean.

OP_CHECKLOCKTIMEVERIFY is implemented for the Clams blockchain in branch develop. It will take a little more effort to get it implemented for Bitcoin. The way I've implemented it is such that you can specify data about the block that a script is in via a new tab, "Block", of the Stack Evaluator. In the following screenshot, I've set "Block height" to 800000.

JackH

sr. member

Activity: 381

Merit: 255

By inclusion I just ment on how the entire path from op_code creation to block inclusion by a miner would look. But I guess there also needs to be a limit to how much you can keep expanding it, without ending up with a full copy of Bitcoin

Kefkius

member

Activity: 64

Merit: 20

Quote from: JackH on December 30, 2015, 05:47:23 AM

There is a similar tool written in Javascript for Forth here: http://skilldrick.github.io/easyforth/

What would be interesting you implement is a mock miner that can be configured. This way we can test the more exotic op_codes and understand how they would behave from a inclusion point of view in the blocks. Do you have support for BIP65 already?

Seems pretty cool. About the mock miner: That could potentially be a useful plugin. I'm not sure exactly what you mean by an "inclusion point of view" though.

BIP65 support isn't implemented. I've been putting it off in hopes that somebody else would implement it in python-bitcoinlib. Tongue

But the underlying script evaluation API in python-bitcoinlib will take some changes to support including block height and/or block time in the evaluation, so I guess other people have the same idea that I do. Ultimately I'll implement it regardless, since I need to put in a way to override opcode behavior based on what chainparams are active. For example, a blockchain like CLAMS uses a different CHECKLOCKTIMEVERIFY implementation than Bitcoin does.

JackH

sr. member

Activity: 381

Merit: 255

There is a similar tool written in Javascript for Forth here: http://skilldrick.github.io/easyforth/

What would be interesting you implement is a mock miner that can be configured. This way we can test the more exotic op_codes and understand how they would behave from a inclusion point of view in the blocks. Do you have support for BIP65 already?

Kefkius

member

Activity: 64

Merit: 20

Alright, so I've done a lot on Hashmal since I last posted here. It really grew from a "Transaction Script IDE" to a full-fledged toolbox for cryptocurrency development. You can deserialize transactions/blocks, encode/decode addresses, etc. There's an entire plugin system. I've tried to keep the Github Wiki up-to-date with the all the plugins that Hashmal comes with.

Here's a short list of what the new plugins that Hashmal includes do:

Deserialize transactions/blocks
Encode/Decode addresses
Retrieve blockchain data from block explorers or via RPC
Create transactions from scratch

I'm currently working on a large refactor of the Stack Evaluator. Here is how it looks now (apologies for the big image):

The issue I'm currently dealing with is displaying the contents of an inner script (as in Pay-To-Script-Hash scriptSigs) in human-readable form. When I have issues like that solved, I'll merge it into the master branch.

--
edit: For your viewing pleasure, here are a couple of examples with the latest changes:

7 7 7:

string literal:

Kefkius

member

Activity: 64

Merit: 20

Quote from: benehsv on September 21, 2015, 06:22:52 PM

I am also getting an error on osx 10.10

Traceback (most recent call last):
File "./hashmal", line 17, in
gui.main()
File "/Users/bbuenz/Programmierung/CryptoWorkspace/hashmal/hashmal_lib/__init__.py", line 14, in main
self.main_window = HashmalMain(self.app)
File "/Users/bbuenz/Programmierung/CryptoWorkspace/hashmal/hashmal_lib/main_window.py", line 38, in __init__
self.dock_handler.do_default_layout()
File "/Users/bbuenz/Programmierung/CryptoWorkspace/hashmal/hashmal_lib/dock_handler.py", line 90, in do_default_layout
self.dock_widgets['Variables'].setVisible(True)
KeyError: 'Variables'

Did you run setup.py?

benehsv

newbie

Activity: 5

Merit: 0

I am also getting an error on osx 10.10

Traceback (most recent call last):
File "./hashmal", line 17, in
gui.main()
File "/Users/bbuenz/Programmierung/CryptoWorkspace/hashmal/hashmal_lib/__init__.py", line 14, in main
self.main_window = HashmalMain(self.app)
File "/Users/bbuenz/Programmierung/CryptoWorkspace/hashmal/hashmal_lib/main_window.py", line 38, in __init__
self.dock_handler.do_default_layout()
File "/Users/bbuenz/Programmierung/CryptoWorkspace/hashmal/hashmal_lib/dock_handler.py", line 90, in do_default_layout
self.dock_widgets['Variables'].setVisible(True)
KeyError: 'Variables'

Kefkius

member

Activity: 64

Merit: 20

Alright, there's now a plugin system. I'm looking for ideas for plugins, so if you have any, please post them. Note that plugin basically refers to one or more tools in the form of dock widgets, which are detachable widgets. The purpose of plugins is to provide extra functionality when you have a certain requirement installed.

The proof-of-concept plugin that I'm working on is Pycoin, which provides a GUI for pycoin.

Also, there's a "chainparams" system that allows you to specify the format of transactions. For example, in Peercoin and its forks, transactions have a timestamp field after their version field. By selecting the Peercoin chainparams preset, you can deserialize Peercoin transactions, and even set the timestamp for a transaction in the Transaction Builder tool.

Anyway, ideas for useful plugins are welcome!

Kefkius

member

Activity: 64

Merit: 20

Getting closer to a real 0.2.0a release. I've written about and uploaded screenshots of tools on the Github Wiki.

There won't be a changelog until after 0.2.0a since so many things are changing. Here are some key recent changes:

Variables can be substituted into a transaction (see Variables tool documentation on Github wiki).
Variables can be saved to the config file.
Some stack log messages are now better (e.g. grammatically improved).

If you'd like to try Hashmal but don't know where you would start, check out the Example Scripts page.

Kefkius

member

Activity: 64

Merit: 20

Quote from: chrisrico on September 04, 2015, 12:59:02 PM

I'm having difficulty getting it to run. Here's what I'm getting after installing from git on OSX Yosemite.

Code:

$ hashmal
Traceback (most recent call last):
File "/usr/local/bin/hashmal", line 13, in
import hashmal_lib
File "/usr/local/lib/python2.7/site-packages/hashmal_lib/__init__.py", line 5, in
from main_window import HashmalMain
File "/usr/local/lib/python2.7/site-packages/hashmal_lib/main_window.py", line 8, in
from dock_handler import DockHandler
File "/usr/local/lib/python2.7/site-packages/hashmal_lib/dock_handler.py", line 4, in
from docks.addr_encoder import AddrEncoder
File "/usr/local/lib/python2.7/site-packages/hashmal_lib/docks/__init__.py", line 6, in
import tx_deserializer
ImportError: No module named tx_deserializer

My bad. I hadn't included that file in the list of modules in setup.py. Fixed now.

chrisrico

hero member

Activity: 496

Merit: 500

I'm having difficulty getting it to run. Here's what I'm getting after installing from git on OSX Yosemite.

Code:

$ hashmal
Traceback (most recent call last):
File "/usr/local/bin/hashmal", line 13, in
import hashmal_lib
File "/usr/local/lib/python2.7/site-packages/hashmal_lib/__init__.py", line 5, in
from main_window import HashmalMain
File "/usr/local/lib/python2.7/site-packages/hashmal_lib/main_window.py", line 8, in
from dock_handler import DockHandler
File "/usr/local/lib/python2.7/site-packages/hashmal_lib/dock_handler.py", line 4, in
from docks.addr_encoder import AddrEncoder
File "/usr/local/lib/python2.7/site-packages/hashmal_lib/docks/__init__.py", line 6, in
import tx_deserializer
ImportError: No module named tx_deserializer

Kefkius

member

Activity: 64

Merit: 20

Quote from: mustyoshi on September 03, 2015, 03:03:17 PM

Any chance of porting this to Javascript?

Not really, no. The whole thing was designed with Python in mind.

Anyway, there are now some basic usage docs here. I've also put in a basic Transaction Deserializer tool which can be upgraded with the functionality talked about earlier in the thread at a later time.

The most significant update is that you can now specify a "spending transaction" in the Stack Evaluator, and select one of its inputs. That input's scriptSig will be pushed to the stack before your script starts. "OP_CHECKSIG" and "OP_CHECKSIGVERIFY" should work now.

Pretty soon I'll have to start incrementing the version number and providing changelogs.

mustyoshi

sr. member

Activity: 287

Merit: 250

Any chance of porting this to Javascript?

Kefkius

member

Activity: 64

Merit: 20

Thanks for the feedback. Repo transferred to its new home. We're now in #hashmal on freenode.

georgem

legendary

Activity: 1484

Merit: 1007

spreadcoin.info

exactly the tool I need right now. Thanks!

CanaryInTheMine

donator

Activity: 2352

Merit: 1060

between a rock and a block!

Great tool!! It's very useful, wish it was made sooner

Kefkius

member

Activity: 64

Merit: 20

Quote from: antanst on September 03, 2015, 04:06:16 AM

Very nice, thank you!

Should there be a list of the Python dependencies somewhere, like a requirements file?

Yeah, there's a requirements.txt now.

Quote from: SteamGamesBTC.com on September 03, 2015, 04:58:48 AM

Python isn't so bad at all, Electrum is also written in Python and works. ;-) However the Windows version of Electrum is nicely compiled to the one executable file. Do you plan the same?

When it's a little more mature, definitely.

SteamGamesBTC.com

hero member

Activity: 734

Merit: 507

Python isn't so bad at all, Electrum is also written in Python and works. ;-) However the Windows version of Electrum is nicely compiled to the one executable file. Do you plan the same?

antanst

sr. member

Activity: 294

Merit: 260

Very nice, thank you!

Should there be a list of the Python dependencies somewhere, like a requirements file?

anton000

newbie

Activity: 21

Merit: 0

Wow. Pretty awesome contribution op.

If this were only available 2 years go... Damn Cheesy

walletrecoveryservices

member

Activity: 89

Merit: 21

Nice idea OP. Thanks for your contribution to the community.

sAt0sHiFanClub

hero member

Activity: 546

Merit: 500

Warning: Confrmed Gavinista

Well done, this looks really good. I think you are going in the right direction with this. Wink

Luke-Jr

legendary

Activity: 2576

Merit: 1186

Quote from: Kefkius on September 03, 2015, 02:15:39 AM

Quote from: Luke-Jr on September 02, 2015, 10:52:36 PM

Would be nice if it could use libbitcoinconsensus for evaluation just to avoid any potential bugs.
I recall one time when Valgrind messed up an x86 instruction, it totally confused my debugging efforts for hours. Sad

... and that was without money at stake ... :|

Disclosure: Shameless try-to-build-demand for reopening Bitcoin Core pull request #6178, "Refactor EvalScript into a CScriptExecution class, so single-stepping can be done"

I hadn't thought of that. It's a good idea, albeit one that would complicate starting Hashmal (e.g. compiling the C++ extension). The current method of stepping through scripts - via re-implementing EvalScript() from python-bitcoinlib - is rather inelegant. I'm also relatively unfamiliar with C++ (as opposed to C) extensions in Python. Anyway, there probably should be a better implementation of script evaluation, so that's currently the only option on the table.

libbitcoinconsensus, while it does use C++, exports only a C API

Should be able to just use ctypes I would think (with stock libbitcoinconsensus lib).
(Of course, this only works with #6178 merged Sad

)

Kefkius

member

Activity: 64

Merit: 20

Quote from: Luke-Jr on September 02, 2015, 10:52:36 PM

Would be nice if it could use libbitcoinconsensus for evaluation just to avoid any potential bugs.
I recall one time when Valgrind messed up an x86 instruction, it totally confused my debugging efforts for hours. Sad

... and that was without money at stake ... :|

Disclosure: Shameless try-to-build-demand for reopening Bitcoin Core pull request #6178, "Refactor EvalScript into a CScriptExecution class, so single-stepping can be done"

I hadn't thought of that. It's a good idea, albeit one that would complicate starting Hashmal (e.g. compiling the C++ extension). The current method of stepping through scripts - via re-implementing EvalScript() from python-bitcoinlib - is rather inelegant. I'm also relatively unfamiliar with C++ (as opposed to C) extensions in Python. Anyway, there probably should be a better implementation of script evaluation, so that's currently the only option on the table.

Quote from: Uptrenda on September 02, 2015, 11:17:10 PM

Quote from: Kefkius on September 02, 2015, 10:36:44 PM

Quote from: Uptrenda on September 01, 2015, 06:40:51 PM

First POST and OP congratulations because you just solved a problem that every Bitcoin developer of the past few years has grappled with. I'm definitely going to be checking this out.

Edit: seems like an easy improvement to this is add another field for raw transactions to check that the TX is valid and to inspect the script works as intended. I just noticed what you wrote about multi-sig but should be quite easy to code this (pybitcoin-lib already has all the code.)

Thanks! The thing about CHECKSIG is that the Stack View tool was getting pretty crowded so I didn't put in any way to include a transaction. But you're definitely right in that it wouldn't be very difficult to do. I think the best way to accomplish that is to create a tool called "Transaction Deserializer" or some such, and have it be able to interact with the Stack View tool. I'll give that a shot.

Oh man, where were you 2 years ago when I was getting started? I'm wishing good things for this project

Thanks for the good wishes.

Uptrenda

member

Activity: 114

Merit: 16

Quote from: Kefkius on September 02, 2015, 10:36:44 PM

Quote from: Uptrenda on September 01, 2015, 06:40:51 PM

First POST and OP congratulations because you just solved a problem that every Bitcoin developer of the past few years has grappled with. I'm definitely going to be checking this out.

Edit: seems like an easy improvement to this is add another field for raw transactions to check that the TX is valid and to inspect the script works as intended. I just noticed what you wrote about multi-sig but should be quite easy to code this (pybitcoin-lib already has all the code.)

Thanks! The thing about CHECKSIG is that the Stack View tool was getting pretty crowded so I didn't put in any way to include a transaction. But you're definitely right in that it wouldn't be very difficult to do. I think the best way to accomplish that is to create a tool called "Transaction Deserializer" or some such, and have it be able to interact with the Stack View tool. I'll give that a shot.

Oh man, where were you 2 years ago when I was getting started? I'm wishing good things for this project

Luke-Jr

legendary

Activity: 2576

Merit: 1186

Would be nice if it could use libbitcoinconsensus for evaluation just to avoid any potential bugs.
I recall one time when Valgrind messed up an x86 instruction, it totally confused my debugging efforts for hours. Sad

... and that was without money at stake ... :|

Disclosure: Shameless try-to-build-demand for reopening Bitcoin Core pull request #6178, "Refactor EvalScript into a CScriptExecution class, so single-stepping can be done"

Kefkius

member

Activity: 64

Merit: 20

Quote from: Uptrenda on September 01, 2015, 06:40:51 PM

First POST and OP congratulations because you just solved a problem that every Bitcoin developer of the past few years has grappled with. I'm definitely going to be checking this out.

Edit: seems like an easy improvement to this is add another field for raw transactions to check that the TX is valid and to inspect the script works as intended. I just noticed what you wrote about multi-sig but should be quite easy to code this (pybitcoin-lib already has all the code.)

Thanks! The thing about CHECKSIG is that the Stack View tool was getting pretty crowded so I didn't put in any way to include a transaction. But you're definitely right in that it wouldn't be very difficult to do. I think the best way to accomplish that is to create a tool called "Transaction Deserializer" or some such, and have it be able to interact with the Stack View tool. I'll give that a shot.

Uptrenda

member

Activity: 114

Merit: 16

First POST and OP congratulations because you just solved a problem that every Bitcoin developer of the past few years has grappled with. I'm definitely going to be checking this out.

Edit: seems like an easy improvement to this is add another field for raw transactions to check that the TX is valid and to inspect the script works as intended. I just noticed what you wrote about multi-sig but should be quite easy to code this (pybitcoin-lib already has all the code.)

Kefkius

member

Activity: 64

Merit: 20

Hashmal

Transaction Script IDE (Github | IRC)

Hashmal is an IDE for Bitcoin transaction scripts. Its purpose is to make it easier to write, evaluate, and learn about transaction scripts. There are very few tools that help people learn about how transaction scripts work; hopefully Hashmal can do this.

Demonstration of stack evaluation.

Hashmal is intended for cryptocurrency developers and power users. It uses python-bitcoinlib for most of its functionality.

Key Features

Write scripts in human-readable form.
Observe the stack as scripts are executed.
See explanations of script operations as they are evaluated.

Notice

Hashmal is a project in its infancy. Please be careful if you intend to use any scripts you make in Hashmal with an actual blockchain. This is alpha software.

You can download Hashmal from Github.

Topic: Hashmal - Transaction Script IDE (alpha) (Read 9198 times)