V2.4 Testing Thread

[carlaKC]

We're close to being able to tag v2.4, so it's time to do some testing!

This is ✨ incredibly ✨ useful, because we don't cut RCs so if we find any bugs in the release we'll need to cut another one.

General testing is great, but in particular it would be helpful to test the following:

• sim-ln/bugfix: Set OpenSSL to be built from source  #249: does make build-docker work on your machine?
• Eclair node support #225: test different functionality against eclair
• sim-cli/bug: Interaction between simulation file path and results directory #228: check differnt ways to specify -s and -d all work
• simln-lib/feat: Parse address detecting if starts with https #212: sim files with and without http(s) work

If you've tested the release, please comment and specify:

• The commit you tested on
• Your OS/ARCH
• What you tested
• Bugs you hit or observations

The more details you can provided on what you tested, the better! It helps other people understand which areas are under-tested, and provides them with the ability to easily re-produce your results.

[zealsham]

OS-Tested: MAC OS / Linux

sim-file directly and data-dir directy

command: sim-cli -s sim.json -d results -t 120

Behaviour observed

worked as expected , the sim file was read and the data directory created to log the results

on second run the contents in the result dir wasnt deleted , rather. new log was appended to the directory .

• irrespective of the folder name passed to —d , sim-ln still creates a result directory , should we allow the user define directory name they would want ?.

sim-file not present root dir , but in data-dir

Command: sim-cli -d data-dir -t 120

• sim-cli successfully reads the simfile in the data dir
• however when the -d isnt provided and an existing results directory exist , sim-ln doesnt see the simulation file , was thinking it should since “results” directory is the default for sim-ln

I plan to set up a rust dev environment on my windows box to do some testing on windows , however for the main time , this was my observation on linux and mac .

[carlaKC]

Thanks for testing!

Should we allow the user define directory name they would want ?.

I'm not sure what you mean by this?

But I do agree that our behavior should be to overwrite not append!

however when the -d isn't provided and an existing results directory exist , sim-ln doesnt see the simulation file , was thinking it should since “results” directory is the default for sim-ln

Just so I understand this, you're running sim-cli with no other args and the following existing dirs/files:

/results/sim.json

We want to cater to:

• A default location (sim.json)
• A custom location (-d or -s)

I think it's okay to fail if the simulation file is in the results dir, if you want to have a non-default location then you should specify it with -s.