Three Issues with "exchange-files" example

[daviddias]

• Version:

Using latest master with the respective example

• Platform:

» uname -a
Darwin imp16.Home 20.1.0 Darwin Kernel Version 20.1.0: Sat Oct 31 00:07:11 PDT 2020; root:xnu-7195.50.7~2/RELEASE_X86_64 x86_64

• Subsystem:

Examples

Severity:

Medium

Description:

Found 3 issues with the exchange examples

1) The ipfs node returns every address, even if it failed to set up the listener (actual bug in IPFS)

✅ FIXED | 2) Typo in how addresses are formatted

Why does it attempt to show all addresses as /p2p/? Is there an issue if it displays /ipfs/ addresses?

Update: Fix at #3398

CLOSED | 3) Workspace functionality doesn't seem to work~~

Might have been a problem on my side. It didn't work for a few times, but now I can't repeat the failure.

Steps to reproduce the error:

Start the example with latest. The errors should be visible too

[daviddias]

🐞 Found an actual 3rd issue. 4) The PubSub messages don't seem to get transferred between peers.

Tried multiple combinations (chrome <-> firefox, chrome <-> chrome, firefox <-> firefox) and got not results.

I added some debug statements, the messageHandler never sees the message coming from the other peer.

Can someone run the example on their machine and see if it does work?

[achingbrain]

Hmm, weird - the tests pass and it works for me locally between FF and Chrome. The missing /p2p/ thing is a bug, though it's only a display thing, it shouldn't affect the running of the example.

[daviddias]

@achingbrain to distinguish, are you saying that both issue 1) and 4) do not present themselves to you?

[daviddias]

@achingbrain today my experience has degraded, I don't even see peers showing anymore

Additionally, I'm seeing the browser node connecting to a bunch of TCP peers, which as we know, it is not possible

[achingbrain]

I am not seeing 4). Does the test in the example pass for you? You can comment out this line to see the browsers do the steps in front of you.

.1) is a bit weird, and the tcp peers thing is definitely not right, I think #3400 is reporting a similar issue, how peer addresses are treated by the address book seems to have changed recently