Welcome to my project called flow for defining and running parallel,
dataflow programs like this one:
It currently consists of libraries for compiling (flowclib) and running (flowrlib) flows as
well as command line binaries (flowc and flowr) that use the libraries.
As data-flow programming fits very well with a visualization of processes operating on data flowing from other processes, a visual programming method would be great. But I've only touched the surface of an IDE, disappointed by the current state of cross-platform UI programming. That's potentially a great (but massive) project in itself.
In this project I work on the programming "semantics" as I implement some sample programs. It's a journey of discovery of writing something like this (for me), learning rust in the process and learning how such a paradigm could work.
Learn more about what it is and why I created it by reading the Inspirations section of the book, or dive right in and see if you can understand Your First Flow with zero previous knowledge and just your programmer's intuition.
This README.md forms part of the 'flow' Guide published using gh-pages.
If you want to encourage me then you can "patreonize me" :-)
You need Git to clone the repo.
From your command line:
bash
git clone https://github.com/andrewdavidmackenzie/flow.git
You need make and a rust toolchain (rustup, cargo, rustc) to build from source
(I suggest using rustup).
Once you have those, you can install the remaining pre-requisites using:
bash
make config
These include libraries like ssl and tools like mdbook, mdbook-linkcheck and graphviz to build the docs.
The make config target should install them all for you. It should work on macos and linux variants using apt-get
or yum package managers (PRs to Makefile are welcome for other linux package managers).
To build and test, including building and testing docs and running the samples and checking their output is correct:
bash
make
NOTE
The first time you build, it will take a long time. This is due to a large number of function implementations
in the flowstdlib standard library each being compiled to WASM as individual projects. After the first build,
dependencies are tracked by the flowc compiler and implementations are only re-compiled when required.
Run the 'fibonacci' sample flow using:
cargo run -p flowc -- samples/fibonacci
You should get a fibonacci series of numbers output to the terminal.
The first flow section of the guide walks you through it.
After a PR is merged, and the build of master succeeds, the guide is rebuilt and the resulting html is deployed for
viewing here
The Project is structured into a number of rust crates that form part of a rust cargo "workspace".
Currently, two binaries are built: flowc the flow compiler and flowr the flow runner.
See the Project Components and Structure section of the Guide for more details.
I organize all issues in a Github Project and chose things to work on from the "Next" column. I have only marked a couple of issues with "help wanted" label but I can do more if there is interest. If in doubt reach out to me by email, or GitHub issue.
I have been exploring how best to build a cross-platform IDE for flow that can be used to visually build/inspect/debug flows. The main "problem" is choosing the best cross-platform, rust-friendly GUI toolkit to use. I have tried Iced and other emerging rust-native GUI toolkits, but they are not ready yet. * flowide-gtk - First attempt at creating a flow IDE with a GUI, all written in rust. It is not visual yet, just text based but can load and compile and then run a flow. To get this to work I also spun-off: * gtk-rs-state - a small library to help do a GTK UI in rust, accessing UI from background threads.
Currently, I'm experimenting with flutter. FLTK looks relatively easy to use, has rust native bindings for the UI, but it doesn't look so modern and is not so well known.
Once I get a solution I like, I may well bring it back into this repo, although build-times for so many crates may become a problem again if this repo includes too much.