Closed 05nelsonm closed 8 months ago
This also makes publishing things much simpler, as darwin targets always take forever. If can publish an update for tor only (and not all modules) it'd save a lot of headache.
Should also rename binary-initializer
to binary-core-initializer
so publication name space is consistent.
If split versioning is going to be a thing, mine as well migrate everything to kmp-tor
instead of having 2 repositories..... :thinking:
kmp-tor
before publication.1.x.x
branch can still publish under the same git tag to give people lead time to migrate to 2.0.0
until it can then be archived and dropped.If split versioning is going to be a thing, mine as well migrate everything to
kmp-tor
instead of having 2 repositories..... 🤔
- This repository's version control is blown up from the inclusion of resources (my bad).
- Will be able to perform integration tests on produced binaries with
kmp-tor
before publication.- Everything will be housed in a single repo making it easier for people to grok (for the most part).
- The
1.x.x
branch can still publish under the same git tag to give people lead time to migrate to2.0.0
until it can then be archived and dropped.
Actually, that is confusing as hell and presents a real problem tracking things. What would be better is to move those modules to a separate repository, kmp-tor-core
, with it's own release cycle and what not.
binary-core
andbinary-core-api
modules should be released as0.1.0
, while thebinary
module (and futurebinary-pt
modules) follow the versions for the binaries they contain.binary-core
will only every change when support for a new architecture is added forJvm
/Node.js
binary-core-api
will only every change if newInstaller.Paths
are needed (such as adding support forobsf4
andsnowflake
pluggable transports. ThoseInstaller.Paths
can be added beforehand as well, so a new release for it is not neededbinary-core
andbinary-core-api
release versions in sync, even thoughbinary-core
does not depend onbinary-core-api
.