-
Notifications
You must be signed in to change notification settings - Fork 1.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[new release] zmq (4 packages) (5.3.0) #25513
Conversation
CHANGES: * Add eio binding in zmq-eio (issuu/ocaml-zmq#126, @andersfugmann)
dea0f25
to
4275eae
Compare
I don't remember this issue on opensuse in the past, do you know what may cause it?
|
"dune" {>= "2.7"} | ||
"ocaml" {>= "4.04.1"} | ||
"zmq" {= version} | ||
"eio" {>= "0.10"} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should this be >= 0.10 < 1.0
? In case of breaking changes.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We try to avoid preventive upper bounds: if we see a breakage when eio 1.0 is released, we will add the upper bound manually.
In fact, this has already been built and tested by the CI with eio 1.0 without any problem
Other surprising failures on linux:
They seem similar to the opensuse one. I don't remember seeing these in the past, can you have a look? |
These errors seem like scheduling errors. I've seen those happen in other projects: some test scenario is written to check the behaviour of some small program, but the test is often overly specialised to the dev machine and/or the project's CI machine, and then it fails when running on the opam-repository CI because the machines are specced very differently. I'd suggest to try to run the tests on more exotic machines, on machines that are very busy with some high-CPU-usage process running concurrently, etc. to reproduce locally. Note that often, the issue is just with the test (it's too specific in its expectations) rather than the code. |
Ping @andersfugmann |
Sorry for the delay. |
I have not been able to track down why this happens. It seems that opensuse distributes a non-standard version of libzmq, as it delivered unknown (undocumented) events at times. I think the opensuse test failures are safe to ignore for now. |
They do indeed. I think the problem is that events sent to the zmq socket monitor changes with each release of zmq. I'm close to just rewrite the whole monitor test to verify that we get some events at all to better tolerate different versions of zmq. I think we can safely ignore the errors for now. The challenge is of course when testing and verifying updated to any dependencies. Is there a way to detect is tests are run as part of opam CI (e.g. an environment variable) we can depend on to disable the specific monitor test? I could create a new release with a much simpler version of the monitor test to have less flaky tests. |
The following is set in the environments:
|
Thanks! |
OCaml bindings for ZeroMQ 4.x
CHANGES: