[ANNOUNCE] GHC 8.2.1 release candidate 2

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

[ANNOUNCE] GHC 8.2.1 release candidate 2

Ben Gamari-3

Hello everyone,

The GHC team is very pleased to announce the second candidate of the
8.2.1 release of the Glasgow Haskell Compiler. Source and binary
distributions are available at

    https://downloads.haskell.org/~ghc/8.2.1-rc2/

This is the second of what will likely be either two or three release
candidates leading up the final 8.2.1 release. This release will
feature,

  * A new type-indexed Typeable implementation

  * The long awaited Backpack

  * Deriving strategies for disambiguating DeriveAnyClass,
    GeneralizedNewtypeDeriving, and stock mechanisms

  * Overloaded record fields

  * Improved compiler performance

  * Better code generation through more robust tracking of join points

  * Compact regions for more efficient garbage collection and serialization

  * Better support for machines with non-uniform memory architectures

  * More robust support for levity (e.g. RuntimeRep) polymorphism

  * A simple interface for streaming eventlog data from live processes

  * Further refinement of DWARF support

This candidate fixes most of the issues present in release candidate
one including,

 * #13233: typePrimRep panic while compiling GHC with profiling enabled
 * #13509: type error involving unboxed tuples
 * #13426: compile-time memory-usage regression
 * #13560: Windows binary distributions carry absolute paths to toolchain
 * #13585: Control.Lens.Wrapped.ala causes compiler panic
 * #13623: Join points produce bad code for stream fusion
 
As always, please let us know if you have difficulty. Thanks to everyone
who has contributed!

Happy testing,

- Ben

_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

signature.asc (497 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [ANNOUNCE] GHC 8.2.1 release candidate 2

Ben Gamari-3
George Colpitts <[hidden email]> writes:

> Hi Ben
>
> I built from source and ran the tests on my Mac and found some
> problems. I'm not sure if the failing tests have been ran successfully
> by others on this platform. I did "make slowtest". Maybe the problem
> only happens on my machine.
>
Currently Harbormaster only runs `make test`, not `make slowtest`.
Consequently, `slowtest` is generally rather broken, even on Linux.
Every once in a while I look at it and try to pare down the failures,
but it's an up-hill battle.

> I'm new to running the testsuite and not sure how the sleep settings on my
> computer affect long running computations.
>
>    - If I want to run a long running test such as "make slowtest" overnight
>    will my computer go to sleep preventing the test from running? i.e. should
>    I invoke it with something like "caffeinate -i make slowtest" ?
>
That sounds right to me.

> I almost didn't run the tests assuming they had been run as part of the
> release process but then I guessed that maybe slowtest had not been run. It
> would be a pain but would it be worth documenting which tests had been run
> on which platforms?
>
I currently don't validate the binary distribution tarballs. Instead I
judge validation state from Harbormaster's testing of the ghc-8.2
branch.

Over the summer we intend on revamping our CI infrastructure, which
should make it easier to do nightly runs of slowtest (and perhaps
provide nightly or even per-commit binary distributions).

> I assume I should file a bug for the following?
>
That would be great. I had a quick look at this and it looks quite
likely that the simplifier is looping: even -fsimpl-tick-factor=1000
doesn't succeed. This looks like a real regression.

Cheers,

- Ben


_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

signature.asc (497 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [ANNOUNCE] GHC 8.2.1 release candidate 2

George Colpitts
Yes, I agree, will file a bug this evening.

On Wed, May 17, 2017 at 10:26 AM Ben Gamari <[hidden email]> wrote:
George Colpitts <[hidden email]> writes:

> Hi Ben
>
> I built from source and ran the tests on my Mac and found some
> problems. I'm not sure if the failing tests have been ran successfully
> by others on this platform. I did "make slowtest". Maybe the problem
> only happens on my machine.
>
Currently Harbormaster only runs `make test`, not `make slowtest`.
Consequently, `slowtest` is generally rather broken, even on Linux.
Every once in a while I look at it and try to pare down the failures,
but it's an up-hill battle.

> I'm new to running the testsuite and not sure how the sleep settings on my
> computer affect long running computations.
>
>    - If I want to run a long running test such as "make slowtest" overnight
>    will my computer go to sleep preventing the test from running? i.e. should
>    I invoke it with something like "caffeinate -i make slowtest" ?
>
That sounds right to me.

> I almost didn't run the tests assuming they had been run as part of the
> release process but then I guessed that maybe slowtest had not been run. It
> would be a pain but would it be worth documenting which tests had been run
> on which platforms?
>
I currently don't validate the binary distribution tarballs. Instead I
judge validation state from Harbormaster's testing of the ghc-8.2
branch.

Over the summer we intend on revamping our CI infrastructure, which
should make it easier to do nightly runs of slowtest (and perhaps
provide nightly or even per-commit binary distributions).

> I assume I should file a bug for the following?
>
That would be great. I had a quick look at this and it looks quite
likely that the simplifier is looping: even -fsimpl-tick-factor=1000
doesn't succeed. This looks like a real regression.

Cheers,

- Ben


_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [ANNOUNCE] GHC 8.2.1 release candidate 2

Ryan Scott
In reply to this post by Ben Gamari-3
Hi Alberto. Thanks for the very detailed report!

> - A weird kind error when using ConstraintKinds in a propietary package
> which didn't manifest itself with ghc < 8.2:
> ...
> Is this expected behaviour?
> Should I try to isolate and open a ticket?

This looks like a proper bug to me. Can you minimize the example a submit a bug report at https://ghc.haskell.org/trac/ghc/newticket for this? Thanks!

> - I had to disable the tests for two packages since they seem to "hang"
>  (ie: they never finish running and don't seem to consume any CPU time).
> These packages are lens-4.15.1 and fingertree-0.1.1.0. Maybe it's a Nix
> environmental issue, I'm not sure. Can anyone reproduce this?

The fact that the lens tests run forever sounds unusual to me, as the lens repo has been running regression tests with GHC 8.2 for a while with no observed slowdowns. I'll double-check soon just to be sure, though.

However, I can confirm that the fingertree tests appear to loop forever at runtime with GHC 8.2 (as opposed to GHC 8.0, where they finish in about 7.5 seconds). This is certainly not a good thing, so I'll try to investigate this more. Thanks for noticing this.

> - I can't manage to install several packages which include executables
> (namely, update-nix-fetchgit and snap-server, for the moment) because Cabal
> says that it cannot find the source for the main module of the executables:
> "Setup: can't find source for Main in ."
> It seems that the "hs-source-dir" directive in the .cabal file is not being
> honored. Maybe a Nix-only issue? Can anyone reproduce this? Any ideas on
> how can I fix it?

I can't reproduce this issue, at least with update-nix-fetchgit-0.1.0.0 (by using `cabal install` to install it). Can you give more detailed instructions on how to trigger this error?

Ryan S.

_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [ANNOUNCE] GHC 8.2.1 release candidate 2

Ryan Scott
A follow-up:

* The fact that the fingertree test suite doesn't terminate is an occurrence of a known bug, GHC Trac #13429 [1].
* I just ran the lens test suite with GHC 8.2, and it terminated [2]. It does take a while, though, sound it's understandable that one would think it loops forever if you don't have the patience to let it finish :)

Again, I'm not able to reproduce the ConstraintKinds or Cabal regressions you reported, so it would be tremendously helpful if you could submit bug reports explaining how to trigger those issues. Thanks!

Ryan S.
-----

On Fri, May 19, 2017 at 9:16 AM, Ryan Scott <[hidden email]> wrote:
Hi Alberto. Thanks for the very detailed report!

> - A weird kind error when using ConstraintKinds in a propietary package
> which didn't manifest itself with ghc < 8.2:
> ...
> Is this expected behaviour?
> Should I try to isolate and open a ticket?

This looks like a proper bug to me. Can you minimize the example a submit a bug report at https://ghc.haskell.org/trac/ghc/newticket for this? Thanks!

> - I had to disable the tests for two packages since they seem to "hang"
>  (ie: they never finish running and don't seem to consume any CPU time).
> These packages are lens-4.15.1 and fingertree-0.1.1.0. Maybe it's a Nix
> environmental issue, I'm not sure. Can anyone reproduce this?

The fact that the lens tests run forever sounds unusual to me, as the lens repo has been running regression tests with GHC 8.2 for a while with no observed slowdowns. I'll double-check soon just to be sure, though.

However, I can confirm that the fingertree tests appear to loop forever at runtime with GHC 8.2 (as opposed to GHC 8.0, where they finish in about 7.5 seconds). This is certainly not a good thing, so I'll try to investigate this more. Thanks for noticing this.

> - I can't manage to install several packages which include executables
> (namely, update-nix-fetchgit and snap-server, for the moment) because Cabal
> says that it cannot find the source for the main module of the executables:
> "Setup: can't find source for Main in ."
> It seems that the "hs-source-dir" directive in the .cabal file is not being
> honored. Maybe a Nix-only issue? Can anyone reproduce this? Any ideas on
> how can I fix it?

I can't reproduce this issue, at least with update-nix-fetchgit-0.1.0.0 (by using `cabal install` to install it). Can you give more detailed instructions on how to trigger this error?

Ryan S.


_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [ANNOUNCE] GHC 8.2.1 release candidate 2

Alberto Valverde
Hi Ryan,

I hope to be able to reproduce the ConstraintKinds bug in a minimal example later today. I believe that the cabal bug might be something to do with my nix setup so will investigate more to see how nix installs Haskell packages under the hood. Will follow-up as soon as I can.

Thanks!

Alberto

On Fri, May 19, 2017 at 7:39 PM, Ryan Scott <[hidden email]> wrote:
A follow-up:

* The fact that the fingertree test suite doesn't terminate is an occurrence of a known bug, GHC Trac #13429 [1].
* I just ran the lens test suite with GHC 8.2, and it terminated [2]. It does take a while, though, sound it's understandable that one would think it loops forever if you don't have the patience to let it finish :)

Again, I'm not able to reproduce the ConstraintKinds or Cabal regressions you reported, so it would be tremendously helpful if you could submit bug reports explaining how to trigger those issues. Thanks!

Ryan S.
-----

On Fri, May 19, 2017 at 9:16 AM, Ryan Scott <[hidden email]> wrote:
Hi Alberto. Thanks for the very detailed report!

> - A weird kind error when using ConstraintKinds in a propietary package
> which didn't manifest itself with ghc < 8.2:
> ...
> Is this expected behaviour?
> Should I try to isolate and open a ticket?

This looks like a proper bug to me. Can you minimize the example a submit a bug report at https://ghc.haskell.org/trac/ghc/newticket for this? Thanks!

> - I had to disable the tests for two packages since they seem to "hang"
>  (ie: they never finish running and don't seem to consume any CPU time).
> These packages are lens-4.15.1 and fingertree-0.1.1.0. Maybe it's a Nix
> environmental issue, I'm not sure. Can anyone reproduce this?

The fact that the lens tests run forever sounds unusual to me, as the lens repo has been running regression tests with GHC 8.2 for a while with no observed slowdowns. I'll double-check soon just to be sure, though.

However, I can confirm that the fingertree tests appear to loop forever at runtime with GHC 8.2 (as opposed to GHC 8.0, where they finish in about 7.5 seconds). This is certainly not a good thing, so I'll try to investigate this more. Thanks for noticing this.

> - I can't manage to install several packages which include executables
> (namely, update-nix-fetchgit and snap-server, for the moment) because Cabal
> says that it cannot find the source for the main module of the executables:
> "Setup: can't find source for Main in ."
> It seems that the "hs-source-dir" directive in the .cabal file is not being
> honored. Maybe a Nix-only issue? Can anyone reproduce this? Any ideas on
> how can I fix it?

I can't reproduce this issue, at least with update-nix-fetchgit-0.1.0.0 (by using `cabal install` to install it). Can you give more detailed instructions on how to trigger this error?

Ryan S.


_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs



_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [ANNOUNCE] GHC 8.2.1 release candidate 2

Alberto Valverde
I've manage to reproduce the CK bug without any external dependencies and opened a ticket with an attached module that fails to compile with 8.2.1-rc2 (w/o explicit KindSignature) but compiles fine with 8.0.2.


The lens problem was a false alarm and a lack of patience from me... they indeed finish and pass if you give them enough time. Sorry for the noise.

Will hopefully follow up soon with the Cabal issues I'm experiencing

Alberto

On Mon, May 22, 2017 at 9:33 AM, Alberto Valverde <[hidden email]> wrote:
Hi Ryan,

I hope to be able to reproduce the ConstraintKinds bug in a minimal example later today. I believe that the cabal bug might be something to do with my nix setup so will investigate more to see how nix installs Haskell packages under the hood. Will follow-up as soon as I can.

Thanks!

Alberto

On Fri, May 19, 2017 at 7:39 PM, Ryan Scott <[hidden email]> wrote:
A follow-up:

* The fact that the fingertree test suite doesn't terminate is an occurrence of a known bug, GHC Trac #13429 [1].
* I just ran the lens test suite with GHC 8.2, and it terminated [2]. It does take a while, though, sound it's understandable that one would think it loops forever if you don't have the patience to let it finish :)

Again, I'm not able to reproduce the ConstraintKinds or Cabal regressions you reported, so it would be tremendously helpful if you could submit bug reports explaining how to trigger those issues. Thanks!

Ryan S.
-----

On Fri, May 19, 2017 at 9:16 AM, Ryan Scott <[hidden email]> wrote:
Hi Alberto. Thanks for the very detailed report!

> - A weird kind error when using ConstraintKinds in a propietary package
> which didn't manifest itself with ghc < 8.2:
> ...
> Is this expected behaviour?
> Should I try to isolate and open a ticket?

This looks like a proper bug to me. Can you minimize the example a submit a bug report at https://ghc.haskell.org/trac/ghc/newticket for this? Thanks!

> - I had to disable the tests for two packages since they seem to "hang"
>  (ie: they never finish running and don't seem to consume any CPU time).
> These packages are lens-4.15.1 and fingertree-0.1.1.0. Maybe it's a Nix
> environmental issue, I'm not sure. Can anyone reproduce this?

The fact that the lens tests run forever sounds unusual to me, as the lens repo has been running regression tests with GHC 8.2 for a while with no observed slowdowns. I'll double-check soon just to be sure, though.

However, I can confirm that the fingertree tests appear to loop forever at runtime with GHC 8.2 (as opposed to GHC 8.0, where they finish in about 7.5 seconds). This is certainly not a good thing, so I'll try to investigate this more. Thanks for noticing this.

> - I can't manage to install several packages which include executables
> (namely, update-nix-fetchgit and snap-server, for the moment) because Cabal
> says that it cannot find the source for the main module of the executables:
> "Setup: can't find source for Main in ."
> It seems that the "hs-source-dir" directive in the .cabal file is not being
> honored. Maybe a Nix-only issue? Can anyone reproduce this? Any ideas on
> how can I fix it?

I can't reproduce this issue, at least with update-nix-fetchgit-0.1.0.0 (by using `cabal install` to install it). Can you give more detailed instructions on how to trigger this error?

Ryan S.


_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs




_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [ANNOUNCE] GHC 8.2.1 release candidate 2

Jens Petersen-4
In reply to this post by Ben Gamari-3
Thanks for RC2!

I built it for Fedora in my Copr repo: https://copr.fedorainfracloud.org/coprs/petersen/ghc-8.2.1.

While I haven't tested it yet, I note that the builds take rather long:

https://copr.fedorainfracloud.org/coprs/petersen/ghc-8.2.1/build/555495/

A perf build with the testsuite takes up to 3 hours!  I am building with 8.2.1 if it matters...

I don't know if anything changed in the copr build system but for RC1 the same build take no more than 96 min!
So it is quite a noticeable change.

Jens

_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [ANNOUNCE] GHC 8.2.1 release candidate 2

Ryan Scott
In reply to this post by Ben Gamari-3
Alberto,

> I've manage to reproduce the CK bug without any external dependencies and
> opened a ticket with an attached module that fails to compile with
> 8.2.1-rc2 (w/o explicit KindSignature) but compiles fine with 8.0.2.

Thanks for doing this!

Will hopefully follow up soon with the Cabal issues I'm experiencing

One thing that might be useful to do is to invoke Cabal with high verbosity (e.g., -v3) and post the logs somewhere. I wouldn't be surprised if Nix was passing some uncommonly used flags that have regressed in recent versions of Cabal.

Ryan S.

_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [ANNOUNCE] GHC 8.2.1 release candidate 2

Alberto Valverde
Hi Ryan,

Some preogress... The cabal issues I reported seem more of a nix issue but I'm not sure what is causing them. FWIW, nix does not use cabal-install at all but instead compiles the Setup.hs script and runs that.

The weird thing is that i'm trying to reproduce what the nix haskell-package builder [1] does and the problem I described (no Main.hs could be found) does not occur but instead the Main.hs module fails to compile due to an unrelated error. So it seems that, somehow, the unrelated compile error is being mistaken by a lack of module.

Even weirder is that the snap-server.cabal says that the failing executables should only be built when flags are enabled (which are disabled by default) but they are being built anyway when nix calls Setup, even when I explicitly disable them via nix. The nix build log [2] even says that the flags (build-pong and build-testserver) are disabled! However, If I manually execute Setup the configure flags are honored.

I'll continue to investigate to try and make some sense of this but I think it's not cabal's fault so feel free to ignore me for the meantime. I'll chime back if/when I find it is an issue with the new Cabal.

Thanks,
Alberto


On Tue, May 23, 2017 at 6:28 PM, Ryan Scott <[hidden email]> wrote:
Alberto,

> I've manage to reproduce the CK bug without any external dependencies and
> opened a ticket with an attached module that fails to compile with
> 8.2.1-rc2 (w/o explicit KindSignature) but compiles fine with 8.0.2.

Thanks for doing this!

Will hopefully follow up soon with the Cabal issues I'm experiencing

One thing that might be useful to do is to invoke Cabal with high verbosity (e.g., -v3) and post the logs somewhere. I wouldn't be surprised if Nix was passing some uncommonly used flags that have regressed in recent versions of Cabal.

Ryan S.

_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs



_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [ANNOUNCE] GHC 8.2.1 release candidate 2

Alberto Valverde
It's most certainly not a Cabal issue. The problem only manifests itself on packages which I have applied Nix's "doJailbreak", which mangles the .cabal file in order to remove bounds. Maybe the format of the .cabal has changed slightly and the parser it uses trips up? Anyway, sorry for the noise

On Fri, May 26, 2017 at 11:11 AM, Alberto Valverde <[hidden email]> wrote:
Hi Ryan,

Some preogress... The cabal issues I reported seem more of a nix issue but I'm not sure what is causing them. FWIW, nix does not use cabal-install at all but instead compiles the Setup.hs script and runs that.

The weird thing is that i'm trying to reproduce what the nix haskell-package builder [1] does and the problem I described (no Main.hs could be found) does not occur but instead the Main.hs module fails to compile due to an unrelated error. So it seems that, somehow, the unrelated compile error is being mistaken by a lack of module.

Even weirder is that the snap-server.cabal says that the failing executables should only be built when flags are enabled (which are disabled by default) but they are being built anyway when nix calls Setup, even when I explicitly disable them via nix. The nix build log [2] even says that the flags (build-pong and build-testserver) are disabled! However, If I manually execute Setup the configure flags are honored.

I'll continue to investigate to try and make some sense of this but I think it's not cabal's fault so feel free to ignore me for the meantime. I'll chime back if/when I find it is an issue with the new Cabal.

Thanks,
Alberto


On Tue, May 23, 2017 at 6:28 PM, Ryan Scott <[hidden email]> wrote:
Alberto,

> I've manage to reproduce the CK bug without any external dependencies and
> opened a ticket with an attached module that fails to compile with
> 8.2.1-rc2 (w/o explicit KindSignature) but compiles fine with 8.0.2.

Thanks for doing this!

Will hopefully follow up soon with the Cabal issues I'm experiencing

One thing that might be useful to do is to invoke Cabal with high verbosity (e.g., -v3) and post the logs somewhere. I wouldn't be surprised if Nix was passing some uncommonly used flags that have regressed in recent versions of Cabal.

Ryan S.

_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs




_______________________________________________
ghc-devs mailing list
[hidden email]
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Loading...