[new release] ansifmt (2.0.0) #28803
Conversation
|
A test is failing on some archs, probably still related to floating-point shenanigans (though I thought I had fixed it...), I'll try to see what is going on |
qexat
force-pushed
the
release-ansifmt-2.0.0
branch
from
7082e3a to
583aa5e
Compare
|
Okay, I'm honestly out of ideas for fixing it on these three archs. |
|
Hm. The test is comparing floating point values across widely different architectures IIUC: How about just comparing the results up to some 'tolerance'? 🤔 |
I expected |
CHANGES: ## Changes - `Fmt` has been rewritten from scratch to use my own library called `rich-string`. As such, a lot of tests have been dropped as they already exist upstream. The interface is mostly the same, besides the type itself and a few parameters that got renamed. - `Fmt.print`'s `ending` parameter now takes a `Fmt.t option` instead of a `string option`. ## Features - `Ansi` and `Color` now implement their own dedicated equality. No need to use the built-in polymorphic equality anymore. ## Removed - `Fmt`'s serialization feature is now dependent on `rich-string`. Since the latter does not provide it yet, serialization was dropped for now. - `Fmt.show` was also removed. On top of not being provided by `rich-string` and being the identity function, it was not all that useful.
qexat
force-pushed
the
release-ansifmt-2.0.0
branch
from
583aa5e to
4827b43
Compare
|
I expect classification to work reliably. It is the subtraction result you are passing to it that may differ... |
There was a problem hiding this comment.
Thanks - all green now, LGTM!
Would you consider adding an x-maintenance-intent entry?
https://github.com/ocaml/opam-repository/blob/master/governance/policies/archiving.md
|
I will add it once I'm back home |
|
If you want, we can add the intent manually here, without need to re-release. Then you can port it later to your main branch at your own convenience. What do you think? If you agree, what would be your preferred intent? |
|
@mseri your suggestion is fine by me, we can do that |
3 participants
A simple, lightweight library for ANSI styling
CHANGES:
Changes
Fmthas been rewritten from scratch to use my own library calledrich-string. As such, a lot of tests have been dropped as they already exist upstream. The interface is mostly the same, besides the type itself and a few parameters that got renamed.Fmt.print'sendingparameter now takes aFmt.t optioninstead of astring option.Features
AnsiandColornow implement their own dedicated equality. No need to use the built-in polymorphic equality anymore.Removed
Fmt's serialization feature is now dependent onrich-string. Since the latter does not provide it yet, serialization was dropped for now.Fmt.showwas also removed. On top of not being provided byrich-stringand being the identity function, it was not all that useful.