Closed ElectreAAS closed 7 months ago
For no release yet: there is a failure case detected by fuzzing for ISO week date time pattern matching, and I just haven't found time to investigate that properly.
For the upper bound: I don't recall actually - I think it might have been just adding that constraint to old packages probably by me (could also be by opam maintainer).
What is your use case? If it's (somewhat) urgent I can organise time this week to get it onto opam. (Or just push a version without ISO week date pattern matching I guess.)
My use case is that I wanted to move a project which uses timere from ocaml 4.14 to 5.0.0
This isn't urgent in any way, so take all the time you need. For now I'm pinning timere on your repo and everything runs fine
I'm quite curious about what people use timere for - is the project public?
Yes the project is public, it's https://github.com/ocurrent/current-bench I don't even really know what we use timere for to be honest
All good, art-w replied to me about its use case in another issue.
Timedesc 1.1.0 has been published, so now it just remains to run timere tests for the last time and I'll be able to publish to opam.
Any update on this?
Sorry I'm working on PhD thesis and won't be looking at any major time sinks for next few months.
If use of timere is minimal then I think replacing it with something else might be better. I am not confident I will put that much effort in near future into what has been relatively fruitless/unused so far, sorry.
No worries Thanks for dropping by, and good luck with your PhD :)
Ah right current-bench uses timere to drive scheduling. @art-w how extensive is timere used in current-bench?
@ElectreAAS After some thinking, I think I'll pick the route to least resistance: I'll publish a version of Timere maybe next week and leave the bug fixing to future.
Pushed the experimental features to 0.9.0 release instead.
Submitted PR for 0.8.0 release: https://github.com/ocaml/opam-repository/pull/24858
Woh thanks a lot!
We use timere to benchmark the OCaml compiler on a weekly basis (for trunk
, 5.0
and 5.1
, to check for regressions in the time it takes to compile large OCaml projects and the size of the binary artifacts that are produced by the compiler). Those benchmarks take a lot of time to complete, and so the weekly runs save on energy compared to the more aggressive "benchmark on every commit" that we normally do. I believe we have @punchagan to thanks for setting this up, if you have more questions!
Thanks again for your time(re :D), and good luck with your PhD!
Thanks for the kind words from both of you! Fingers crossed haha.
Timere 0.8.0 PR at opam-repository has been accepted, closing this issue now.
Similarly to #59, but this time for the
timere
andtimere-parse
packages, I wonder why the latest available version for both of these is0.7.0
and not the latest on this repo. Specifically, on opamtimere
has an upper bound onocaml < 5.0
, which is not present on the latest.Was there a specific reason for that upper bound that is not yet resolved?
Thank you!