ngaro alternatives and similar packages
Based on the "Embeddable Scripting Languages" category.
Alternatively, view ngaro alternatives based on common mentions on social networks and blogs.
7.5 6.4 ngaro VS cel-goFast, portable, non-Turing complete expression evaluation with gradual typing (Go)
* Code Quality Rankings and insights are calculated and provided by Lumnify.
They vary from L1 to L5 with "L5" being the highest. Visit our partner's website for more details.
Do you think we are missing an alternative of ngaro or a related project?
This is an embeddable Go implementation of the Ngaro Virtual Machine.
This repository contains the embeddable virtual machine, a rudimentary symbolic assembler for easy bootstrapping of projects written in Ngaro machine language, and the retro command line tool that can be used as a replacement for the Retro reference implementations.
Please visit http://forthworks.com/retro/ to get you started about the Retro language and the Ngaro Virtual Machine.
The main purpose of this implementation is to allow customization and communication between Retro programs and Go programs via custom opcodes and I/O handlers (i.e. scripting Go programs in Retro). The package examples demonstrate various use cases. For more details on I/O handling in the Ngaro VM, please refer to http://retroforth.org/docs/The_Ngaro_Virtual_Machine.html.
Another goal is to make the VM core as neutral as possible regarding the higher level language running on it. For example, the in-memory string encoding scheme is fully customizable. Retro specific behaviors are provided via the lang/retro package.
Custom opcodes are implemented by intercepting implicit calls to negative memory addresses. This allows the VM to be fully backwards compatible with existing Retro images while still providing enhanced capabilities. The maximum number of addressable cells is 231 when running in 32 bits mode (that's 8GiB or memory on the host). The range [-231 - 1, -1] is available for custom opcodes.
This implementation passes all tests from the retro-language test suite and its performance when running tests/core.rx is slightly better than with the reference implementations:
1.08s for this implementation, no custom opcodes, compiled with Go 1.7, linux/amd64 1.15s for the reference assembly implementation, linux/386 1.30s for the reference Go implementation, compiled with Go 1.7, linux/amd64 2.00s for the reference C implementation, compiled with gcc-5.4 -O3 -fomit-frame-pointer
Yes, Go 1.7's new SSA backend is THAT good on this type of workload :)
For all intents and purposes, the VM behaves according to the specification. This is of particular importance to implementors of custom opcodes: the VM always increments the PC after each opcode, thus opcodes altering the PC must adjust it accordingly (i.e. set it to its real target minus one).
Install the retro command line tool:
go get -u github.com/db47h/ngaro/cmd/retro
go test -i github.com/db47h/ngaro/vm go test -v github.com/db47h/ngaro/vm/...
Build a retroImage:
cd $GOPATH/github.com/db47h/ngaro/cmd/retro make retroImage
Test the retro command line tool:
./retro --with vm/testdata/core.rx
Should generate a lot of output. Just check that the last lines look like this:
ok summary 360 tests run: 360 passed, 0 failed. 186 words checked, 0 words unchecked, 37 i/o words ignored. ok bye
Support for 32/64 bits memory images on all architectures
Since v2.0.0, the default Cell type (the base data type in Ngaro VM) is Go's int. This means that depending on the target you compile for, it will be either 32 or 64 bits. The retro command line tool supports loading and saving retro memory images where Cells can be either size. For example, to quickly get started you can do this:
echo "save bye" | \ retro -image vm/testdata/retroImage -ibits 32 -o retroImage
This will load the memory image file
vm/testdata/retroImage which we know to
be encoded using 32 bits cells, and save it in the current directory with
whatever encoding is the default for your platform. You could also force a
specific output Cell size with the
Loading and saving with encodings different from the target platform is safe: it will work or generate an error, but never create a corrupted memory image file. For example, with a 64 bits retro binary, saving to 32 bits cells will check that written values fit in a 32 bit int. If not, it will generate an error.
If for some reason you need a specific cell size, regardless of the target
platform's native int size, you can force it by compiling with the tags
go install -tags ngaro32 github.com/db47h/ngaro/cmd/retro
will build a version of retro that uses 32 bits cells, regardless of
GOARCH. Likewise, the
ngaro64 tag will force 64 bits cells, even on
32 bits targets (it'll be twice as slow though).
This project uses semantic versioning and tries to adhere to it.
See the releases page.
For a detailed change log, see the commit log.
No rules, just common sense. Bells, wristles and any other preformance improvements are very welcome. The only changes that will never be accepted are those that will break compatibility with the VM specification.
PRs are a good place to discuss changes so do not hesitate to send PRs directly. Or an issue, if your really have an issue.
Places to Go
Have a look at @dobegor's fork. It breaks compatibility with the spec, but it's got some nice features like floating point support, pausable VM, and a few others in their TODO list.
The Retro language and Ngaro Virtual Machine are Copyright (c) 2008-2016 Charles Childers (and many others), licensed under the ISC license. See the file LICENSE-RETRO at the root of this repository for more details as well as a full list of contributors.
Note that all files in the _misc and vm/testdata folders are verbatim copies from the retro-language project. As such, only Retro's ISC license applies to these files.
*Note that all licence references and agreements mentioned in the ngaro README section above are relevant to that project's source code only.