Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.77%. Comparing base (0cebdb6) to head (29fca51).

@@            Coverage Diff             @@
##             main     #871      +/-   ##
==========================================
+ Coverage   97.76%   97.77%   +0.01%     
==========================================
  Files         155      155              
  Lines        7876     7906      +30     
  Branches     1921     1932      +11     
==========================================
+ Hits         7699     7729      +30     
  Misses        176      176              
  Partials        1        1              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

@StarpTech I'm not at all sure if that would fix the issue you brought in issue #701, so I'd like to have your opinion. I tried in dry-run mode by switching Lerna-Lite from fixed version to independent mode. I also did not try the Git commands either, so I have no idea if it would work or fail. I rarely type Git commands in the shell (I'm typically fine with VSCode so...)

So do you think that would work or is it going in the wrong direction? Because if that doesn't work then I have no idea on how else to address the problem. You can see below each tag are created 1 by 1 (that is nothing new, it was already like that) but then the next section "Pushing tags..." is where it's different, it's try to push the Git tags one at a time (if the Git command works, I'm not sure).

Here's 1 Git command (which again I haven't tried, so it might fail)

git push --no-verify --atomic origin @lerna-lite/cli@4.0.0 main

Actually 1 thing that I'm not sure is that the Git push is async and in Lerna it's awaiting the current one with await this.gitPushToRemote();. So I probably need to await each Git push individually with the new option.

EDIT

I modified the code to enqueue all Git push promises and also added p-limit with a concurrency of 1 to push to remote. I think the code will now better represent the 1 at a 1 execution. @StarpTech I still wish to get feedback from your side if you think this could potentially fix your problem. Also note again that I cannot run this on a live example at the moment because Lerna-Lite doesn't use independent mode, unless I create a full monorepo demo which would take more time.

Created a basic monorepo lernalite-basic-boiler for testing this option, it looks fine and pushed each git tags 1 by 1 as shown below in the verbose logs. So it looks fine to proceed

no feedback provided, but let's go ahead with it in a new release. @StarpTech if you have a chance to try it in the future and confirm that it works, that would be great. I think the PR will fix the GitHub limit. Cheers

Hi, in your repo, were you able to reproduce the issue?

@StarpTech no since the repo is quite basic and it only has 2 packages so I won't hit the limit, however I tested the git push 1 by 1 and it seems to be working as expected. Since it is really being sent 1 by 1 to the git remote, I'm assuming this will fix the issue (however please note that it will of course take longer to send all package tags since they are run 1 at a time and in series)