VanityTXID-Plugin

Generate custom SHA256 Checksums for any file (either by appending an 8B tail or else by varying 8B starting from an internal target). Also generate txn IDs starting with a specific pattern, using a standard wallet + plugin & watching-only wallet. Also generate vanity addresses for any Script (smart contract). Available for Electron Cash (incl. SLP Edition) on Windows, Linux & macOS. Written in Python, & C++ for the generators. To install the latest version download "VanityTXID-Plugin.zip" above, or from the latest release. Using this plugin users can create and send SLP tokens with custom token/txn ID, like this PoW NFT-child (00000002Dad... minted in about 30secs):

v1.6.1 token genesis used nonce '07000000038a0bcb', which corresponded to the 8th thread, & had hash rate 1.2 MH/s for 433B txn. I suspect assembly code might be a few times faster than sha256.cpp (BCHN). For my i7-2600 CPU, I've read estimates ranging from 5 to 24 MH/s for an 80B block header. For 197B I get nearly 2.0 MH/s, & 6.2 MH/s is for address generation (quadruple the speed of VanitygenCash).

SLP Edition versions 3.6.7-dev7+ don't use up a CPU processor in the background, unlike 3.6.6. The pre-releases also have newer code. The CPU usage issue can arise on all 3 OSs.

One issue is that 0-conf may not apply to the TXID itself. The payment amount can't be changed, but the TXID & message can change before confirmation. If it ever fails, the contract can be improved, or else mining pools can mine the vanity TXID directly in return for a fee. A different version can include simple nonce checksum/s, s.t. miners can't easily deduce a vanity contract is being used. A "smarter" contract could also sign 0-conf message using OP_CHECKDATASIG.

Code::Blocks project files with working example parameters are included in the src folder, so others can build & run immediately, for themselves. There's a serious issue when it comes to deterministic builds which are verifiably identical to the source code. Checksums change with every build, whereas the exact number of Bytes stays the same, e.g. 86016B+8B.

Linux requires eSpeak for TTS (enter 'sudo apt install espeak' in terminal). The latest version of VirtualBox gives good hash rates for all OSs. I did a native Linux test and speed as the same as for WIN10, so Code::Blocks' MinGW is probably fine.

Here is an example of a vanity TXID spending from a vanity address whose Script is exactly 520B, containing 505 0-bytes (1010 0s in hex). Remember that the stack must finish with only 1 item, which can't be a vector containing only 0s. Unfortunately that exact address isn't secure, due to the custom Script I made (security just takes longer to code). An example of a 520B message with vanity TXID is here. A vanity covenant, signed by private key=1, is spent here.

Windows users can compare 64-bit to 32-bit performance by replacing all binaries manually from the zip. 64-bit binaries were 12% faster in a test (1.95/1.74).

VanityTXID.cpp & VanityTXID.rc are compiled together using -O3 -s -march=corei7 g++.exe compiler flags. Same for VanityP2SH. All .dll libraries are extracted from 'codeblocks-20.03mingw-nosetup.zip' & 'codeblocks-20.03-32bit-mingw-32bit-nosetup.zip'. Linux compiling doesn't use Icon.rc, and requires linking pthread library (-lpthread) in Code::Blocks ('sudo apt install codeblocks'). In macOS don't use Code::Blocks. Instead extract/copy src to home folder, then open Terminal.app, enter 'cd src', then 'g++ -std=c++17 -O3 ./VanityTXID.cpp'. macOS will download & install g++ if needed. Then rename the resulting 'a.out' to 'VanityTXID-Plugin' and it's ready to go inside the zip if you want to check your own build's hash rate. Same for VanityP2SH & VanityHash.

v1.6.2 notes:

Custom target for nonce placement in files (e.g. place "<#Nonce>" in a file somewhere). Only the VanityHash binaries are new.
P2SH nonce position at either end or start. OP_CODESEPARATOR prefers the nonce at the start. Needed for AutoCove.
Difficulty labels, in MH. Divide MH by MH/s to guess time.
Bugfix for Message size 255B.
Bugfix for drag & drop in Linux.
CashAddr toggle connection.
Message Byte counter (e.g. 字 is 3B).
(text)/(hex) Message conversion now by highlighting the combo-box.
Bugfix for invalid (non-hex) checksum pattern entry.
SHA256 Checksum 0000888ce25d5f2dc8a84322227c345b370fd667e0a8b006ef6d7affc590fbfe (0.4 kH/s · 5mins). Update via re-install requires restarting EC.

v1.6.1: SHA256 Checksum 0000a04841563dd64bb2930ca7f296a2c152643de5b9df6c4c21c9a14d96b8eb

Support for EC v3.6.6 re-enabled. v1.6.0 added a tail to its own zip, which wasn't allowed by EC-v3.6.6 (nor is adding archive comments using WinRAR).
VanityHash now optionally uses a password, "VanityHashNonceF", to decide which bytes to vary, instead of just adding a tail every time.
All new builds with proper memory allocation. Max file size ~2GB. Unfortunately using too many threads can cause a power failure to the CPU. Each thread demands enough RAM to work a copy of the file. I've tested over 1.7GB with 4 threads (under 3 minutes for Pattern '0', but .iso requires the "password" be placed somewhere within it).
Can now vanitize multiple files (document batch). Assumes wallet has permission to create a new folder. No files are ever renamed.
Drag & Drop for many files! Dropping during current process terminates the old one 1st. Files only, no folders. Drag & drop may be broken on Linux, but works fine on macOS.
Read-only QCheckBox for Script.
Bugfix where the VanityHash program keeps running if someone closes the wallet (oops). Added digit to hash rate, round(_,6). Much more elegant Python script, including .join()
Button icons which tick only during execution. (Pausable movies...)
All executable & WebP files now have SHA256 checksums starting with 0000, generated using 8B tails. The plugin itself is a zip & uses the "password" technique instead (hash rate over 300 H/s).

v1.6.0: SHA256 Checksum: 00003449a1e19a94d82dc7185c1845802c6c3c8aebd67e8083243f5415d9dde1 (0.4 kH/s · 7 mins)

VanityHash, now included, allows vanity checksums for any file. Try it out! The plugin's .zip's checksum now starts with 0000. This allows vanity checksums for Token Documents, BFP uploads, etc. Drag & drop file may be in the next version. An 8B nonce is appended to the end of the file.
Button icons. Address generator MessageBox improved. No nonce for CashAddr Pattern 'P' (same as 'p').
In C++, 'goto Finish' moved to better place.

v1.5.1:

Support for SLP Ed. v3.6.6 re-instated! I just prefer 3.6.7 (pre-releases). Screenie for macOS added.
Max sized Scripts can now be displayed after vanity address generation.
All new executables now enable any number of threads without any bugs. Surprisingly 8 are just as fast as 64, so we weren't missing out on anything. Instead of including mutex, it's much better to pass pointer/s to each thread so they report back. I prefer to time only in Python.
Combined some simple lines of Python code, and reversed some simple 'for' loops using set notation.
Technically a new Icon.webp where the ₿ is animated & shrunk to 32p before being placed on top of the colored circles.

v1.5.0:

Vanity CashAddr generator now included (VanityP2SH). It can "vanitize" any smart contract. It's about quadruple the speed of VanitygenCash, if using only CPU. Only 1 address at a time can be generated, currently. e.g. www.blockchain.com/bch/address/pqqqqqqfucku9gl2l5vtsu8dzmllqg9xn5z34kcu80
Contacts instead of Address Labels. To use old VanityTXID addresses please re-generate them by 1st clearing the CashAddr Pattern, enter correct conversion address, then press Generate button and maybe delete old label. User should make a backup of the wallet, which contains the VanityTXID addresses, to avoid a lot of work reproducing vanity addresses (deterministically) from seed phrase. Eventually the contacts should be labelled too, but the SLP Ed. doesn't show Contact labels.
Color changing TabIcon. New shade of green from bitcoincashpodcast.com
Copy-pasteable labels, incl. hash rate.
.notify & .activateWindow combined.
New buttons, e.g. 'Search Contacts'. Still under 500 lines of Python! I've also slightly improved C++ code.
Users of SLP Ed. 3.6.6 would need to update to a 3.6.7 pre-release (eg dev5) to continue using this plugin.

v1.4.1:

Example button near title, which immediately demonstrates VanityTXID with a real example.
Appended time to hash rate. More decimals so at least one non-0 digit occurs.
window.show_message whenever there's a problem.
.setToolTip for many QtWidgets.
Animated Icon.webp. I've checked it uses 0% of CPU even with multiple wallets. It's just a BCH logo minute clock (1 FPS). I hope other plugins can do something more creative with the idea (playing movies inside a crypto wallet).
Added std::once_flag in C++ which now simplifies Python decoding of subprocess communication. Provide error msg if someone double clicks on exe. Simpler C++, with goto. Eliminated std::stringstream and compat/cpuid.h. Some improvements are thanks to pull request by cculianu. Speed may be like 0.1% faster; build sizes are smaller. A serious issue is whether we should switch to vectors instead of arrays. I chose arrays because vectors seem less quick.
Lack of espeak in Linux now unchecks the TTS box.

v1.4.0:

~31% speed increase by using Bitcoin Core's CSHA256 C++ code. Binary sizes are all much larger now. I've removed the zedwood license. In a future update I might bring it back as a UI selection, since zedwood probably wins on simplicity (imagine having to write every line yourself). Users might want to select between CryptoPP, OpenSSL, Bitcoin Core & zedwood, to check the hash rates.
BugFix: Windows TTS now uses PowerShell -C instead of MSHTA (MicroSoft HTml App.) since the latter isn't allowed on 32-bit WIN10 Home N.
Improved Python script. Only ever extract binaries from zip once, on enable (faster startup). Disabling plugin still removes all binaries.
BugFix: Hash rate was always over by ~1% (3*1/255) due to 'for' loops failing to catch a few used nonce bytes at #255 (or -1). I was just about to switch back from the new 'do' loops, due to them being 1% slower!
Random espeak pitch now in Linux, which is the synth version of random voice. BugFix: Lack of espeak no longer throws an error.
WPM (POSix) now exponential against Rate index. Default @ Rate 5. Max WPM reduced to 450.
Language Translator not working (module can only handle one word at a time).

v1.3.4:

Windows 64 bit binary (with i7-AVX tuning) slightly faster on my CPU (1.5 MH/s instead of 1.4 MH/s). Back in v1.1.0 I downgraded to 32-bit, before I could check the MH/s. Just because EC is 32-bit, doesn't mean its plugins should always be! TBH I haven't tested a 32-bit VM yet. The plugin has now doubled in size. I haven't rebuilt the posix binaries.
Message size limit increaed to 520 Bytes, instead of 512B. Technically for data a better scriptcode could involve 3 OP_NIPs (777777) at the end, since it takes a few to max out the scriptsig limit (1650B). 1650B may correspond to something like a 15-of-15 multisig input, or a few OP_NIPs. 3 OP_NIPs, instead of 2, may be a bit like signing with a middle name, as well as first and last names. 77777777 may be even better (2 middle names), due to an added nonce being separate, for a vanity TXID. 520B example: https://blockchain.com/bch/tx/00000003ace42ee6d165eb3b37d27b42703cb0a56ce2990c60a84e01e78ae6d7
1337 off by default, now dict uses 'OlZEASGTBP'.
Unchecking TTS now disables TTS controls, etc.
Simpler Python script.

v1.3.3:

PrivKey & Password now mutable bytearrays instead of immutable strings.
User can now set how many TXID digits to pronounce, as well as TTS Rate. On POSIX 1->10 corresponds to 175->720 WPM. In a future version I might enable slower speeds, since that could help with 1337.
Random voice in macOS, chosen from 10! I tested in Catalina. To hear hex/1337 speak in any language or accent, a macOS user can uncomment the long Voices line in qt.py
More elegant HashRate timing, & formula which handles all 8B of nonce. More elegant script.

v1.3.2:

Approximate hash rate in MH/s. It assumes all threads are equal, the nonce doesn't make it to the 8th byte, & it's not reliable if there are too many threads. 1.4 MH/s is a good rate for me, for 197B.
1337 option for TTS.
Bugfix: Message size exactly 78 bytes now always works.
.activateWindow may help if someone's watching a video. Fixes an issue where Windows mshta steals focus for the javascript TTS. It involves a 60ms lag, to recapture focus.

v1.3.1:

Max message size now 512 Bytes instead of 75B. e.g. https://blockchain.com/bch/tx/00000073d648302417ad306912c4a43ea1aa91907921d13d95844f58637329c0
Optional Notifications. User can change their mind after mining begins.
Optional TTS on every OS! I've set it to pronounce the pattern + 4 extra digits. Linux requires eSpeak or else there's no voice (sudo apt install espeak). eSpeak comes with Ubuntu. In the future I might add a speed setting, and conceivably l33t (pronounce 0 as O etc).
Message input in either hex or text. Works as a hex converter! It even works with Chinese.
Removed/improved IsHex() method.
Bold heading, altered placeholder text.

v1.3.0:

Full SLP Edition support! That means VanityTXID addresses can now take SLPAddr format. We could already mint vanity token IDs by using a watching-only wallet in the SLP edition, but now the plugin works natively in the SLP Edition. I've tested both Windows & Linux, and will try to update to at least macOS Catalina tomorrow. One zip for all editions & OSs.
multiprocessing module wasn't required for os.cpu_count(). No change to binaries (C++).

v1.2.1:

Bug fix for disable &/or uninstall error. To uninstall v1.2.0 users need to close and then re-open the wallet.
Bug fix when user attempts to mine a txn which can't be mined.
Will now sign both P2SH & P2PKH, in any order, wherever possible. They can be combined using the watching-only wallet, along with other inputs which are signed separately.
I've changed the binaries to use a 3 Byte nonce position, exit() function and more constants.

v1.2.0:

Full support for macOS, version High Sierra and newer.
Bug fixed where closing irrelevant wallets cancels mining. Simplified Python code.
Password & PrivKey are now overwritten before being deleted, and then Garbage Collected.
C++ code simplified by using const nonce position.
All C++ warnings resolved.
New license file mentions copied plugin template code. github.com/KarolTrzeszczkowski/Electron-Cash-Plugin-Template

v1.1.0:

Linux now supported. Linux binary is about 17% faster. I dunno why.
Downgraded Windows binary to 32 bit from 64 bit, since EC is only 32 bit on Windows.
Simplified C++ using exit(0).

v1.0.3:

We can now instantly convert any number of addresses by first selecting them in the Addresses tab and then copy-paste all of them over to the Address Converter in VanityTXID.
Now have a complete list of P2SH addresses, which is updated whenever user does anything to the Converter box. These are guaranteed to be correct, unlike random labels in the Addresses tab.
I fixed an error from v1.0.2 where the user types in a wrong password. Now it asks again, and if user hits cancel then we return to normal. I simplified a lot of Python script, but no change to binary.

v1.0.2:

Nonce position can now be up to 256**2, so the input being "mined" doesn't have to be the first. (There was also a bug in the C++ binary of v1.0.1) e.g. now in www.blockchain.com/bch/tx/0000006727815232d1fd48e1988b9aea8b3e4cd060dfbe44c4a52239c71b5cd5 'deadbeef' doesn't appear until the 2nd input.
Improved code so that raw TX hex can be repeatedly mined (copy-paste after mining, and then again etc).
Full support for legacy P2SH form (starting with '3' instead of 'p').
Multiple P2SH address display, but generating lots of addresses at once will be in next update.
Added version title.

v1.0.1:

Now support uncompressed keys. e.g. www.blockchain.com/bch/tx/000000821afbda9e137ab90bf5de3cad8bf8a3bbe218b7f9c26566ec5779fd8f
Have improved UI with labels.
Have fixed many error reports by using try:/except: functions. I had crash reporter disabled so didn't realize there were errors.
Found Python bitcoin.int_to_hex
del(Password) after it's used.
128<Thread#<256 now allowed (was bug in C++ code). I've found 256 threads might be a tiny bit faster than 8 for patterns six long.

TinosNitso / VanityTXID-Plugin

VanityTXID-Plugin

About

Languages