Example Legacy Library for Sublime Text

Legacy libraries (former "dependencies") bundle python packages, so they can be re-used by Sublime Text packages.

Folder Structure

Platform specific variants are organized in sub folders.

folder	description
all	no restrictions, universally compatible with all versions of Sublime Text, python and platform
st3	requires at least ST3
st3_{os}	... and specific os
st3_{os}_{arch}	... and specific architecture
st4	requires at least ST4
st4_{py}	... and specific python version
st4_{py}_{os}	... and specific os
st4_{py}_{os}_{arch}	... and specific architecture

variable	valid values
{os}	"osx", "linux", "windows"
{arch}	"x32", "x64", "arm64"
{py}	"py33", "py38"

Package Control installs most specific and best matching variant of a library.

When running Sublime Text 3,

• all and st3 are actually only synonyms as Sublime Text 2 is no longer supported.
• all folders beginning with st4 are ignored.

When running Sublime Text 4,

• all folders beginning with st4 are prefered, if present.
• all and st3 are installed for python 3.3 and 3.8
• st3_{os} and st3_{os}_{arch} are considdered containing binaries and thus are installed for python 3.3, only.

When running ST4 on 64bit Windows, it prefers ...

1. st4 over st3
2. st4_py38 over st4
3. st4_py38_windows over st4_py38
4. st4_py38_windows_x64 over st4_py38_windows

Registration

Dependencies must opt-in to python 3.8 via repository.

{
  "$schema": "sublime://packagecontrol.io/schemas/repository",
  "schema_version": "4.0.0",
  "libraries": [
    {
       "name": "example-dependency",
       "author": "packagecontrol",
       "issues": "https://github.com/packagecontrol/example-dependency/issues"
       "releases":[
           {
               "base": "https://github.com/packagecontrol/example-dependency"
               "python_versions": ["3.3", "3.8"]
           }
       ]
    }
  ]
}

Note

JSON repository scheme 4.0.0 is required to specify "python_versions": [].

packagecontrol.io supports up to scheme 3.0.0 only!

Warning

Package Control downloads the whole repository including all folders and drops all but the required variant when installing the library.

This may result in reasonable bandwith being wasted, depending on amount of variants and size of packages.

It is therefore recommended to ship libraries as platform specific python wheels using asset based releases, instead.