This repository contains Conda build recipes for all dropbot driver dependencies. Pre-built win-32/win-64 Conda packages for Python 2.7 and Python 3.6 are available on the dropbot Anaconda channel.

The packages are split into two types, based on whether or not they are platform-dependent (i.e., Windows, Linux, MacOS). Platform-dependent packages typically either include a compile C-extension component, or write files to include, bin, etc. since the Conda include, bin, etc. paths differ based on architecture.

For example, on Windows, the path is:

<conda prefix>/Library/include

On Linux, the path is:

<conda prefix>/include

1. Install conda-build:

   conda install -n root conda-build -c conda-forge
   

2. Build DropBot dependencies (Windows Powershell or Bash). Note that you will need to install this package to use conda with Powershell:

   conda build --croot C:\bld -c dropbot -c sci-bots -c conda-forge --skip-existing -m variants.yaml $(cat bootstrap-build-order.txt)
   

Step (2) uses:

• $(cat bootstrap-build-order.txt) to build recipes in dependency order, such that each recipe will only be built after all dependency packages have been built.

• -m variants.yaml to build against both Python 2.7 and Python 3.6 (in recipes that use {{ python }} variant tag).

• --croot C:\bld to shorten the file path length of the build environment to avoid "filename too long" or "file does not exist" type errors.

  This is a known issue.

The following dropbot driver dependencies may be built as noarch and are compatible across platforms (i.e., Windows, Linux, MacOS) and Python versions:

json-tricks/meta.yaml
onoff/meta.yaml
paho-mqtt/meta.yaml
platformio/click/meta.yaml
platformio/platform-atmelavr/meta.yaml
platformio/platform-teensy/meta.yaml
platformio/platformio-core/meta.yaml
python-clang/meta.yaml
si-prefix/meta.yaml
wheezy.routing/meta.yaml
_source/arduino-helpers/.conda-recipe/meta.yaml
_source/asyncserial.py27/conda/asyncserial/meta.yaml **
_source/asyncserial.py3/conda/asyncserial/meta.yaml **
_source/clang-helpers/.conda-recipe/meta.yaml
_source/conda-helpers/.conda-recipe/meta.yaml
_source/logging-helpers/.conda-recipe/meta.yaml
_source/mqtt-messages-python/.conda-recipe/meta.yaml
_source/nanopb-helpers/.conda-recipe/meta.yaml
_source/or-event/.conda-recipe/meta.yaml
_source/pandas-helpers/.conda-recipe/meta.yaml
_source/serial-device/.conda-recipe/meta.yaml

**: asyncserial has a different version for Python 2.7 and Python 3.5+, but both versions are compatible across platforms (i.e., Windows, Linux, MacOS).

The following dropbot driver dependencies must be built for each platform and Python version:

arduino/fast-digital/meta.yaml
arduino/linked-list/meta.yaml
arduino/slow-soft-i2c-master/meta.yaml
arduino/slow-soft-wire/meta.yaml
clang/meta.yaml
matplotlib/meta.yaml
nanopb/meta.yaml
ntfsutils/meta.yaml
path_helpers/meta.yaml
platformio/framework-arduinoavr/meta.yaml
platformio/framework-arduinoteensy/meta.yaml
platformio/tool-avrdude/meta.yaml
platformio/tool-scons/meta.yaml
platformio/tool-teensy/meta.yaml
platformio/toolchain-atmelavr/meta.yaml
platformio/toolchain-gccarmnoneeabi/meta.yaml
protoc/meta.yaml
ragel/meta.yaml
_source/arduino-memory/.conda-recipe/meta.yaml
_source/arduino-rpc/.conda-recipe/meta.yaml
_source/base-node/.conda-recipe/meta.yaml
_source/base-node-rpc/.conda-recipe/meta.yaml
_source/c-array-defs/.conda-recipe/meta.yaml
_source/nadamq/.conda-recipe/meta.yaml
_source/teensy-minimal-rpc/.conda-recipe/meta.yaml

The following commands can be used to graft an existing recipe git repo to a subdirectory of this repo:

git remote add temp ...other recipe...
git fetch --all
git branch --no-track '"feat(<recipe package name>)"' temp/master
git remote rm temp

git merge -s ours --no-commit '"feat(<recipe package name>)"' --allow-unrelated-histories
git read-tree -u --prefix=<recipe package name> '"feat(<recipe package name>)"'

git commit

Example message to append to the merge commit:

Add `<recipe package name>` recipe.

fix(conda): do not build as noarch

Do not build as `noarch` package since the Conda `include` path differs
based on architecture.

For example, on Windows, the path is:

    <conda prefix>/Library/include

On Linux, the path is:

    <conda prefix>/include