-
Notifications
You must be signed in to change notification settings - Fork 96
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
Building and testing with RFCs FS-1043 FS-1071 #261
Comments
I'm currently testing by setting these
I will also test with
|
Separately, I notice compilation is quite slow, I presume this is related to SRTP constraint processing (I'm also using the debug compiler). It would be wise to do a fulll compiler profiling run both before and after |
@dsyme, just FYI, in current compiler, there is an unreported compiler bug that makes buiding the F#+ test project in DEBUG not possible, you need to build that test project in RELEASE. |
OK, I'm just using Could we get that bug reported? Even if it'snot fixed that's the sort of corner case problem I'd really like to get pinned down with testing before we proceed with the RFCs |
OK turning on
|
It's actually a very long longstanding bug, I never managed to find the time to get mini-repro. |
That error you're getting regarding So it's not related to your work. |
I tried changing the spelling of Note there is definitely a difference in logic that could somehow explain this because the extension members like |
I managed to fix the bug found above, and FSHarpPlus now compiles with the preview of FS-1043 However
I will look at these tomorrow The test compilation failure is this:
|
Just to clarify, that's not really from the tests. It comes from the samples folder which has a project. |
|
It seems to me that it's trying to match the first trait-call for the one that should resolve is this one:
which contains another trait-call to the members defined on external classes, without tupling. Maybe this is a good case to test with @smoothdeveloper improvement on these error reports. |
Yes. It's a bit confusing that some Map instances take the extra arguments and some don't |
It is. Basically, all the overloads on the "Witnesses" type use dummy / tupled parameters while the ones defines in custom types have a clean signature. |
I've verified that the latest commit of RFC FS-1043 (dotnet/fsharp#6805) builds and compiles the latest master of FSharpPlus, with the exception of these two lines: type ParallelArray<'t> with
static member inline (+) (x: parray<'m>, y: parray<'m>) = lift2 plus x y : parray<'m>
type ZipList<'s> with
static member inline (+) (x: ZipList<'a>, y: ZipList<'a>) = lift2 plus x y : ZipList<'a> My conclusion from previous threads on this was that this can be considered an appropriate adjustment to the specification and a code change would be needed here. It is the only place where we've found a code change may be needed |
Could you possibly try it with #302 ? |
It's ok, I'm sure we can work it out once the feature lands. The case is under test both positively and negateively now in the RFC implementation. |
To test this I put
|
Testing with your feature/ext branch of the compiler I see lot of failures in the test projects. Here's a relative small repro:
I updated to the latest commit of the feature/ext branch but still fails. |
I'm opening this issue to cover steps to build and test this library with the implementations of RFCs FS-1043 FS-1071 activated
dotnet/fsharp#6805
dotnet/fsharp#6810
We can also look at adjusting the library to make use ot the features
The text was updated successfully, but these errors were encountered: