Luckily Hlint also support HLINT instead which removed the TH pipeline.

Where is this described/documented? All I can see here https://github.com/ndmitchell/hlint#readmeis

For {-# HLINT #-} pragmas GHC may give a warning about an unrecognised pragma, which can be suppressed with -Wno-unrecognised-pragmas.

which mentions HLINT pragmas but says nothing about what they do. Simon On Mon, 4 Dec 2023 at 09:05, Moritz Angermann wrote:

Any ANN annotation triggers the TH pipeline and makes them really painful to work with, in non-stage2 settings. Lots of Hlint annotations use ANN and then you have iserv be triggered for each module that has an ANN annotation.

Luckily Hlint also support HLINT instead which removed the TH pipeline.

That alone is enough for me personally to recommend against using ANN if there is an alternator option to anyone who asks me.

On Mon, 4 Dec 2023 at 5:01 PM, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...

The whole ANN mechanism https://ghc.gitlab.haskell.org/ghc/doc/users_guide/extending_ghc.html?highli...is, at root, a good idea. It is pretty generan, and allows annotations to be arbitrary expressions, provided they are in Typable and Data. And they are serialised across modules.

In practice though, I'm not sure how widely used they are. I'm not sure why. I'd love to hear of counter-examples.

Only top level binders can be annotated; but there is no reason in principle that you should not annotate instance declarations. I don't think it'd be too hard to implement.

Simon

On Sat, 2 Dec 2023 at 14:51, Jaro Reinders wrote:

...

Hi GHC devs,

I'm working on a GHC plugin which implements a custom instance resolution mechanism:

https://github.com/noughtmare/transitive-constraint-plugin

Currently, I need to place instances in a specific order in a specific file to recognize them and use them in my plugin. I think my life would be a lot easier if I could put annotations on instances. I imagine a syntax like this:

data MyInstanceTypes = Refl | Trans deriving Eq

class f <= g where inj :: f x -> g x

instance {-# ANN instance Refl #-} f <= f where inj = id

instance {-# ANN instance Trans #-} forall f g h. (f <= g, g <= h) => f <= h where inj = inj @g @h . inj @f @g

Using this information I should be able to find the right instances in a more reliable way.

One more thing I was thinking about is to make it possible to remove these instances from the normal resolution algorithm and only allow them to be used by my plugin.

Do you think this would be easy to implement and useful? Or are there other ways to achieve this?

Cheers,

Jaro _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

I don’t think they do anything specific.

Now I am truly baffled! If they don't do anything, why would they be a module at all! Surely they do something? Simon On Mon, 4 Dec 2023 at 11:58, Moritz Angermann wrote:

I don’t think they do anything specific. They just function as a marker to Hlint to find when parsing the source files. Here is one of the original issues we had: https://github.com/ndmitchell/hlint/issues/1251

Simply by not being ANN, it doesn’t trigger the Templar Haskell machinery and thus does not cause compilation slowdowns or iserv needs (e.g. render the module impossible to cross compiler for stage1 cross compilers with not TH support).

On Mon, 4 Dec 2023 at 7:45 PM, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...

Luckily Hlint also support HLINT instead which removed the TH pipeline.

...

Where is this described/documented? All I can see here https://github.com/ndmitchell/hlint#readmeis

...

For {-# HLINT #-} pragmas GHC may give a warning about an unrecognised pragma, which can be suppressed with -Wno-unrecognised-pragmas.

which mentions HLINT pragmas but says nothing about what they do.

Simon

On Mon, 4 Dec 2023 at 09:05, Moritz Angermann wrote:

...

Any ANN annotation triggers the TH pipeline and makes them really painful to work with, in non-stage2 settings. Lots of Hlint annotations use ANN and then you have iserv be triggered for each module that has an ANN annotation.

Luckily Hlint also support HLINT instead which removed the TH pipeline.

That alone is enough for me personally to recommend against using ANN if there is an alternator option to anyone who asks me.

On Mon, 4 Dec 2023 at 5:01 PM, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...

The whole ANN mechanism https://ghc.gitlab.haskell.org/ghc/doc/users_guide/extending_ghc.html?highli...is, at root, a good idea. It is pretty generan, and allows annotations to be arbitrary expressions, provided they are in Typable and Data. And they are serialised across modules.

In practice though, I'm not sure how widely used they are. I'm not sure why. I'd love to hear of counter-examples.

Only top level binders can be annotated; but there is no reason in principle that you should not annotate instance declarations. I don't think it'd be too hard to implement.

Simon

On Sat, 2 Dec 2023 at 14:51, Jaro Reinders wrote:

...

Hi GHC devs,

I'm working on a GHC plugin which implements a custom instance resolution mechanism:

https://github.com/noughtmare/transitive-constraint-plugin

Currently, I need to place instances in a specific order in a specific file to recognize them and use them in my plugin. I think my life would be a lot easier if I could put annotations on instances. I imagine a syntax like this:

data MyInstanceTypes = Refl | Trans deriving Eq

class f <= g where inj :: f x -> g x

instance {-# ANN instance Refl #-} f <= f where inj = id

instance {-# ANN instance Trans #-} forall f g h. (f <= g, g <= h) => f <= h where inj = inj @g @h . inj @f @g

Using this information I should be able to find the right instances in a more reliable way.

One more thing I was thinking about is to make it possible to remove these instances from the normal resolution algorithm and only allow them to be used by my plugin.

Do you think this would be easy to implement and useful? Or are there other ways to achieve this?

Cheers,

Jaro _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Ah, so returning to my original question: Where is this described/documented? All I can see here

https://github.com/ndmitchell/hlint#readmeis...

E.g. is there a HLint user guide? I'm interested in what the annotations can and cannot be. E.g. perhaps ANN could shortcircuit the TH stuff in some common cases? Simon On Mon, 4 Dec 2023 at 12:15, Moritz Angermann wrote:

I see. That’s where the confusion comes from. Hlint uses them to allow ignoring specific Hlint warnings:

{-# ANN module "HLint: ignore Use string literal" #-}

{- HLINT ignore "Use string literal" -}

and similar. One could maybe argue they should have never been ANN pragmas to begin with.

Examples taken from this SO question:

https://stackoverflow.com/questions/19237695/haskell-how-to-tell-hlint-not-t...

On Mon, 4 Dec 2023 at 8:07 PM, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...

...

I don’t think they do anything specific.

Now I am truly baffled! If they don't do anything, why would they be a module at all! Surely they do something?

Simon

On Mon, 4 Dec 2023 at 11:58, Moritz Angermann wrote:

...

I don’t think they do anything specific. They just function as a marker to Hlint to find when parsing the source files. Here is one of the original issues we had: https://github.com/ndmitchell/hlint/issues/1251

Simply by not being ANN, it doesn’t trigger the Templar Haskell machinery and thus does not cause compilation slowdowns or iserv needs (e.g. render the module impossible to cross compiler for stage1 cross compilers with not TH support).

On Mon, 4 Dec 2023 at 7:45 PM, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...

Luckily Hlint also support HLINT instead which removed the TH pipeline.

...

Where is this described/documented? All I can see here https://github.com/ndmitchell/hlint#readmeis

...

For {-# HLINT #-} pragmas GHC may give a warning about an unrecognised pragma, which can be suppressed with -Wno-unrecognised-pragmas.

which mentions HLINT pragmas but says nothing about what they do.

Simon

On Mon, 4 Dec 2023 at 09:05, Moritz Angermann < moritz.angermann@gmail.com> wrote:

...

Any ANN annotation triggers the TH pipeline and makes them really painful to work with, in non-stage2 settings. Lots of Hlint annotations use ANN and then you have iserv be triggered for each module that has an ANN annotation.

Luckily Hlint also support HLINT instead which removed the TH pipeline.

That alone is enough for me personally to recommend against using ANN if there is an alternator option to anyone who asks me.

On Mon, 4 Dec 2023 at 5:01 PM, Simon Peyton Jones < simon.peytonjones@gmail.com> wrote:

...

The whole ANN mechanism https://ghc.gitlab.haskell.org/ghc/doc/users_guide/extending_ghc.html?highli...is, at root, a good idea. It is pretty generan, and allows annotations to be arbitrary expressions, provided they are in Typable and Data. And they are serialised across modules.

In practice though, I'm not sure how widely used they are. I'm not sure why. I'd love to hear of counter-examples.

Only top level binders can be annotated; but there is no reason in principle that you should not annotate instance declarations. I don't think it'd be too hard to implement.

Simon

On Sat, 2 Dec 2023 at 14:51, Jaro Reinders wrote:

> Hi GHC devs, > > I'm working on a GHC plugin which implements a custom instance > resolution > mechanism: > > https://github.com/noughtmare/transitive-constraint-plugin > > Currently, I need to place instances in a specific order in a > specific file to > recognize them and use them in my plugin. I think my life would be a > lot easier > if I could put annotations on instances. I imagine a syntax like > this: > > data MyInstanceTypes = Refl | Trans deriving Eq > > class f <= g where > inj :: f x -> g x > > instance {-# ANN instance Refl #-} f <= f where > inj = id > > instance {-# ANN instance Trans #-} > forall f g h. (f <= g, g <= h) => f <= h > where > inj = inj @g @h . inj @f @g > > Using this information I should be able to find the right instances > in a more > reliable way. > > One more thing I was thinking about is to make it possible to remove > these > instances from the normal resolution algorithm and only allow them > to be used > by my plugin. > > Do you think this would be easy to implement and useful? Or are > there other > ways to achieve this? > > Cheers, > > Jaro > _______________________________________________ > ghc-devs mailing list > ghc-devs@haskell.org > http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs > _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs