@onee-only overall the changes look good, but I haven't spent much time on this PR as the testing aren't passing. Let me know if you need any help debugging the issue.

@onee-only overall the changes look good, but I haven't spent much time on this PR as the testing aren't passing. Let me know if you need any help debugging the issue.

The changes only include exposing the tests for plumbing/transport/file . So the failures seem to be caused by existing code. I haven't tried to debug the issue since it didn't seem to belong to the scope of this PR.

About the failures, I suspect they were caused by data race as mentioned in this previous comment: #1450 (comment). And I think this issue points to the same problem #1526.

@onee-only thank you for taking a look into this. 🙇 We will need to get the tests fixed before merging, to avoid that branch failing post merge. Please note I added a commit to your PR to include the fixture information for the tests.

@aymanbagabas I took an initial look at this. The PipeWriter.Write is getting blocked in the io.Copy within SendPack, but only on the sendPackDeleteReference. Any ideas what the problem may be here?

I tried to log some of this today:
Could receive_pack be responsible for a hang? Specifically, this portion where needPackfile is determined and UpdateObjectStorage is conditionally called?

In the case of TestSendPackDeleteReference, the hash is zero so UpdateObjectStorage is skipped. I'm wondering if this leads to unconsumed data in the pipe, causing SendPack to wait for something to consume that data?

@onee-only do you mind merging this onee-only#1 please so that Ayman changes get into this PR as well?

@pjbgf I've merged the PR. Could you take a look at it?

@onee-only Could you merge v6-transport back into your branch?

EDIT: I've opened onee-only#2 to merge latest v6-transport and fix the issues present here (in the last two commits of that PR)

@pjbgf @aymanbagabas I've cleaned the history and fixed a typo in commit message to pass the check.