-
-
Notifications
You must be signed in to change notification settings - Fork 933
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
Incorrect dependency data on rspec old releases #1906
Comments
We seem to have created quite a few development dependencies on
Nothing out of ordinary seems to have happened as per rubygems.org git history and we were probably still in early development. Given the blip in Will verify for any more mismatches with a random set. |
But the example I gave has a different creation date, hasn't it? Not sure if it's related but 80 versions of RSpec were pushed on the 25th of July, 2009. Some of them (all?) having this problem. |
Versions has different creation date, yes. I am pointing to abnormal creation time of dependencies tho (versions and dependencies are separate models/tables). If no versions were created between
This info isn't as useful, we have other anomalous dependencies which weren't created in this period. I wrote a script and it gave
I doubt anyone is going to remember what went wrong in import or why data was massaged later. I will send in a rake task for updating dependencies table for versions before 100358. |
Thanks so much for the clarification @sonalkr132, I don't know rubygems.org data model well so I probably don't make much sense! 😅 So, how is the data going to get fixed? The redundant runtime dependencies will be deleted, right? |
Any news on this? It seems to create a lot of extra requests to RubyGems.org. |
I was going to regenerate the versions file but got never really got to it, sorry. There were other unrelated issues which cause mismatch, I had to verify that they are fixed. Hopefully will get to it this weekend. |
@sonalkr132 Just to be clear on the planned fix, it'd be to remove all the incorrect runtime dependencies from DB, and then regenerating the versions file to pick that up. Is that it? |
yes, download the gemspec (
I doubt @eregon issue has anything to do with this tho. Latest version affected by this issue was released on 11 September, 2009 (appengine-tools, 0.0.3) and I doubt any of gems used by eregon would depend on versions/gems released before this. EDIT: never mind this^. found the issue, will update when we fix. |
@sonalkr132 Just pinging you about this to make sure you still have it mind, and offering help if there's something I can do. |
Thanks for reminder. As I mentioned above, it would be best if we regenerate versions file after ensuring other issues with compact index are fixed. I have to ping bother others to get this task done as I don't have right access. |
Oh, I wasn't aware of that issue, thanks for pointing me to it 👍. I understand this requires time and coordination. I'm not in a hurry to fix this, but I'm excited to see this stuff moving forward :) Anyways, thanks for the transparency and for keeping me posted. If I can help in any way, just let me know. |
See #2343
I had missed disregarding a few cases which meant this number is higher than actual. |
🎉 Thanks so much for this amazing work! |
This has been fixed now, so I'm happily closing this ticket :) |
I noticed that the dependency data on old RSpec releases does not seem correct. For example,
curl https://rubygems.org/api/v2/rubygems/rspec/versions/1.2.3.json | jq
Rubygems.org considers
hoe
andcucumber
as runtime dependencies of RSpec.However, if I look for the gemspec using the
gem
client, I see they are only development dependenciesgem specification --remote -v 1.2.3 rspec
This is causing
bundler
to make unnecessary requests torubygems.org
when runningbundle install
. See https://github.com/bundler/bundler/issues/6914.If we find out the root cause of this we could probably run a script to fix all affected data. Any ideas welcome!
The text was updated successfully, but these errors were encountered: