-
-
Notifications
You must be signed in to change notification settings - Fork 76
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[fixed] 2.1.0 exports broken in 2.2.0 #15
Comments
Yeah this should have been a major release.... Either use // @jbnicolai |
Version 2.2.0 has been unpublished. We'll do a major bump sometime tomorrow. Thanks for bringing this to our attention @ajwhite 👍 |
Thanks for the quick recovery! |
So you are aware, unpublishing breaks builds that have used shrinkwrap to pin to |
@hulbert unpublishing breaks a lot of things, though it was necessary in this case I believe. Our next release will be 3.0.0 anyway. Would you or @sindresorhus be willing to write some steps on how to fix the shrinkwrap issue if others are facing it? I've not used shrinkwrap personally. Sorry if this inconvenienced anyone. |
@hulbert The 2.2.0 version was broken. So I don't really see how not unpublishing would have helped when it's pinned? |
Reopening for discovery. I did a 2.2.1 patch release with the content of the old 2.1.0, just in case someone already upped it to |
Post Mortem2.2.0 was actually released on Feb 21. Some breaking changes that were finally merged (ref #13) were sitting in master until we wanted to do a major version bump. However these changes were inadvertently released in 2.2.0 when the next release should have been 3.0.0 (minor bump instead of a major bump). Since chalk wasn't affected by the otherwise breaking changes (this was by-design), but other projects that use This is also due to the fact the 'breaking' changes were not buggy code, but intentional changes meant for a major version bump, hence why tests still passed on Travis and thus why we were not notified of any breaking CI runs. As Sindre mentioned, we went ahead and still did a 2.2.1 patch release despite the unpublishing of 2.2.0, for those users that had already bumped their package.json dependencies to We acknowledge that the unpublishing of the broken version was unnecessary and a patch release would have been the preferable solution. The root issue has been addressed internally (unrelated: externally, too) and won't be happening again in the future. Maintainers that are facing errors related to
All other users experiencing errors related to We also recommend updating your dependencies to specify If you use shrinkwrap, refer to #17. For any other issues, you are definitely welcome to open a new ticket and I will personally see to it that it's handled in a timely manner. Sorry again to anyone affected by this! |
... and remember to do the same with jspm if you're using it, where this was hitting our project
|
Why was it necessary? Why not publish the version you intended at 2.2.1, and let everybody move forward? If they're referencing a broken 2.2.0 then they can update. I'm not trying to be a PITA about it, just curious where the necessity of unpublishing comes in as opposed to just publishing 2.2.1. |
@laurelnaiad you're correct. Sindre and I discussed it more after the fact and we agree that a patch would have been better. A lesson for the future 👍 |
Ok, it's not a big deal, just wanted to understand. Thanks! :) |
Thanks for the feedback 💃 |
the dancer emoticon should be bigger so we can see her fabulous red dress |
Seriously, unpublishing should be reserved for exceptional cases. I'm not sure how you thought this was a good idea less than a week after the |
PLEASE don't unpublish! This causes disasters downstream. NO, NO! |
@tlrobinson We didn't unpublish the module, just that particular version. Quite different. It also wasn't working just fine. Chalk was working, sure - but that was because the breaking changes were designed to be compatible with how chalk was already using it. However, some packages that used We now know a patch version would have been the better alternative. However, we didn't know the extent of breaking that had been released until after the fire was contained. At the time, unpublish seemed like the correct course of action. @bradvogel We're aware. 👍 Thank you! If you're still facing issues after referring to the post-mortem, please open a new ticket. I'll see to it personally that it is resolved quickly. |
npm ERR! fetch failed https://registry.npmjs.org/ansi-styles/-/ansi-styles-2.2.0.tgz npm WARN retry will retry, error on last attempt: Error: fetch failed with status code 404 chalk/ansi-styles#15 (comment)
Via comments on 7450295#commitcomment-16867578
No longer works, with the result of
Steps to reproduce:
npm install [email protected]
npm install [email protected]
The text was updated successfully, but these errors were encountered: