mod-neko travis-cl fails with "Server not ready"

[0b1kn00b]

>> Building container neko-mod
build-neko.40
> haxe -neko bin/neko/tests.n -lib travix -lib tink_http tests.hxml
> neko bin/neko/tests.n
Waiting for server to be ready...
Unable to find image 'codeurs/mod-neko:latest' locally
latest: Pulling from codeurs/mod-neko
f49cf87b52c1: Pulling fs layer
0658e5f39259: Pulling fs layer
b39e8e23f5cb: Pulling fs layer
24d0fefe342e: Pulling fs layer
1720d4f3584d: Pulling fs layer
24d0fefe342e: Waiting
1720d4f3584d: Waiting
0658e5f39259: Download complete
24d0fefe342e: Verifying Checksum
24d0fefe342e: Download complete
f49cf87b52c1: Verifying Checksum
f49cf87b52c1: Download complete
1720d4f3584d: Verifying Checksum
1720d4f3584d: Download complete
b39e8e23f5cb: Verifying Checksum
b39e8e23f5cb: Download complete
f49cf87b52c1: Pull complete
0658e5f39259: Pull complete
b39e8e23f5cb: Pull complete
24d0fefe342e: Pull complete
1720d4f3584d: Pull complete
Digest: sha256:ff022b417cb90fea16144006d125e3e026d2be530b462efbb0e669480b65fe27
Status: Downloaded newer image for codeurs/mod-neko:latest
9bd5478c640bc265c3c1bd437543091eda2580ded4ebdedf55a9168a09d1fc9c

>> Server not ready

Error: Command failed with error 1

The command "haxe master.hxml -D containers=node-tcp,neko-mod -D targets=node -D container_only" exited with 1.

[back2dos]

From what I've observed, it's always one of two builds that fails (and it's not necessarily the same). I wonder if it's possible that the servers are interfering with one another?

[benmerckx]

Not sure if we want to keep mod_neko testing going anyway at this point?

Since it's not really maintained at all: HaxeFoundation/neko#153

[0b1kn00b]

Is there not a hook where something is installed on the same IO as the waitForConnection in Master? Then it depends on the state of the network as to how long that operation takes.

[0b1kn00b]

100 more milliseconds, and five more retries, and it works. I'm not proud of it, but if it works, it works