Open good-idea opened 5 years ago
Thanks for your quick reply. For now, I'm rolling back and just going to use Lerna as I was before. I've removed all of my tags & commits generated while working on this from my master branch, but if you'd like to see what my config looked like when I left off, it's on this branch:
https://github.com/good-idea/sane-shopify/tree/release-config
Also, thanks for all of your work putting this together! I'm not super familiar with how semantic release works, but I'd be happy to help out with a PR when I can -- I'll keep an eye on the issues.
@good-idea. Experience has shown that we lack the resources to support several monorepo release tools, sorry. In defense, I can say that we have made significant progress in another project — dhoulb/multi-semantic-release. MSR covers all of our cases.
I've set up my repo following the instructions, and matching the @qiwi/substrate repo as closely as possible - but I can't seem to get this to work.
At the end of each log in Travis, I see:
I have been able to publish a single package, but it is unable to authenticate with Github for later packages. Here is an example of a build that included a single
fix
commit, in one of the packages, then is unable to authenticate.Here is the next build, which included multiple
fix
commits. (But, I pushed these commits before the previous build had made it'schore(release)
commit back to the repo - so, semantic release saidThe local branch master is behind the remote one, therefore a new version won't be published.
). Maybe not too relevant --- but you'll see at the end thathooks is not a function
Here is the next build. This included fix commits for the
server
,sanity-plugin
, andsync-utils
packages -- but, it found a change intypes
from the previous tag, and deployed that instead.Finally, this build was run after adding
GH_USER
to the environment variables -- which then breaks here.I'm not sure about the
hooks is not a function
problem, but, as far as the other issues go - I'm guessing that one problem is that the tags are being added to the repo before the rest of the packages have been figured out by semantic-release. I guess I'm a little lost overall - - am I missing something obvious?