No comments on this. Maybe I should raise a ticket? Anybody got an email address to ping David on? ---------- Forwarded message --------- From: Shayne Fletcher Date: Wed, May 22, 2019, 13:03 Subject: Re: [hadrian/windows] build broken To: GHC developers I guess, On Wed, May 22, 2019 at 12:27 PM Shayne Fletcher wrote:

Windows builds with hadrian are currently failing with "file does not exist and no build rule available" `_build/stage1/libffi/build/inst/bin/libffi-6.dll`. Anybody have an ideas on how to overcome that?

relates to 0af519ac583c3544b1c4b1315b38ba0174d3ccb1, "Refactor libffi and RTS rules" (@DavidEichmann). I get the feeling that the current state of affairs is known and there are MRs working their way through the system to restore things already? -- Shayne Fletcher Language Engineer c: +1 917 699 7763 e: shayne.fletcher@daml.com Digital Asset Holdings, LLC 4 World Trade Center 150 Greenwich Street, 47th Floor https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g New York, NY 10007, USA https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g digitalasset.com http://www.digitalasset.com/ -- This message, and any attachments, is for the intended recipient(s) only, may contain information that is privileged, confidential and/or proprietary and subject to important terms and conditions available at  http://www.digitalasset.com/emaildisclaimer.html http://www.digitalasset.com/emaildisclaimer.html. If you are not the intended recipient, please delete this message.

Yes. We actually had one until a few weeks ago but it became really unstable because of some cache messing about with cabal-install's ability to build Hadrian, see https://gitlab.haskell.org/ghc/ghc/issues/16574. I need to figure this out soon indeed, I'll do my best to get us a working Windows/Hadrian job ASAP. On 23/05/2019 14:09, David Eichmann wrote:

Ooops I've let this thread pass by me, but this is most likely due to my latest refactoring of libffi rules in Hadrian. No need to create a new issue, I'll just reopen #16304, and have a look at this now.

On that note, Alp, are we planning on adding a hadrian/windows CI job soon?

- David E

On 5/23/19 12:12 PM, Shayne Fletcher via ghc-devs wrote:

...

No comments on this. Maybe I should raise a ticket? Anybody got an email address to ping David on?

---------- Forwarded message --------- From: *Shayne Fletcher* mailto:shayne.fletcher@daml.com> Date: Wed, May 22, 2019, 13:03 Subject: Re: [hadrian/windows] build broken To: GHC developers mailto:ghc-devs@haskell.org>

I guess,

On Wed, May 22, 2019 at 12:27 PM Shayne Fletcher mailto:shayne.fletcher@daml.com> wrote:

Windows builds with hadrian are currently failing with "file does not exist and no build rule available" `_build/stage1/libffi/build/inst/bin/libffi-6.dll`. Anybody have an ideas on how to overcome that?

relates to 0af519ac583c3544b1c4b1315b38ba0174d3ccb1, "Refactor libffi and RTS rules" (@DavidEichmann).

I get the feeling that the current state of affairs is known and there are MRs working their way through the system to restore things already? -- Shayne Fletcher Language Engineer c: +1 917 699 7763 e: shayne.fletcher@daml.com mailto:shayne.fletcher@daml.com Digital Asset Holdings, LLC 4 World Trade Center 150 Greenwich Street, 47th Floor https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g New York, NY 10007, USA https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g digitalasset.com http://www.digitalasset.com/

This message, and any attachments, is for the intended recipient(s) only, may contain information that is privileged, confidential and/or proprietary and subject to important terms and conditions available at http://www.digitalasset.com/emaildisclaimer.html http://www.digitalasset.com/emaildisclaimer.html. If you are not the intended recipient, please delete this message.

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs -- David Eichmann, Haskell Consultant Well-Typed LLP,http://www.well-typed.com

Registered in England & Wales, OC335890 118 Wymering Mansions, Wymering Road, London W9 2NF, England

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

-- Alp Mestanogullari, Haskell Consultant Well-Typed LLP, https://www.well-typed.com/ Registered in England and Wales, OC335890 118 Wymering Mansions, Wymering Road, London, W9 2NF, England

On Thu, May 23, 2019 at 8:09 AM David Eichmann wrote:

Ooops I've let this thread pass by me, but this is most likely due to my latest refactoring of libffi rules in Hadrian. No need to create a new issue, I'll just reopen #16304, and have a look at this now.

Sounds good. Thanks David.

On that note, Alp, are we planning on adding a hadrian/windows CI job soon?

- David E On 5/23/19 12:12 PM, Shayne Fletcher via ghc-devs wrote:

No comments on this. Maybe I should raise a ticket? Anybody got an email address to ping David on?

---------- Forwarded message --------- From: Shayne Fletcher Date: Wed, May 22, 2019, 13:03 Subject: Re: [hadrian/windows] build broken To: GHC developers

I guess,

On Wed, May 22, 2019 at 12:27 PM Shayne Fletcher wrote:

...

Windows builds with hadrian are currently failing with "file does not exist and no build rule available" `_build/stage1/libffi/build/inst/bin/libffi-6.dll`. Anybody have an ideas on how to overcome that?

relates to 0af519ac583c3544b1c4b1315b38ba0174d3ccb1, "Refactor libffi and RTS rules" (@DavidEichmann).

I get the feeling that the current state of affairs is known and there are MRs working their way through the system to restore things already?

-- Shayne Fletcher Language Engineer c: +1 917 699 7763 e: shayne.fletcher@daml.com Digital Asset Holdings, LLC 4 World Trade Center 150 Greenwich Street, 47th Floor https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g New York, NY 10007, USA https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g digitalasset.com http://www.digitalasset.com/

This message, and any attachments, is for the intended recipient(s) only, may contain information that is privileged, confidential and/or proprietary and subject to important terms and conditions available at http://www.digitalasset.com/emaildisclaimer.html. If you are not the intended recipient, please delete this message.

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

-- David Eichmann, Haskell Consultant Well-Typed LLP, http://www.well-typed.com

Registered in England & Wales, OC335890 118 Wymering Mansions, Wymering Road, London W9 2NF, England

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

-- Shayne Fletcher Language Engineer c: +1 917 699 7763 e: shayne.fletcher@daml.com Digital Asset Holdings, LLC 4 World Trade Center 150 Greenwich Street, 47th Floor https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g New York, NY 10007, USA https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g digitalasset.com http://www.digitalasset.com/ -- This message, and any attachments, is for the intended recipient(s) only, may contain information that is privileged, confidential and/or proprietary and subject to important terms and conditions available at  http://www.digitalasset.com/emaildisclaimer.html http://www.digitalasset.com/emaildisclaimer.html. If you are not the intended recipient, please delete this message.

Hello Shayne, David and I figured out the cause of that problem, I am working on a patch, will put it up as a (WIP) MR as soon as it's ready. On 24/05/2019 13:10, Shayne Fletcher via ghc-devs wrote:

Hi David,

On Thu, May 23, 2019 at 8:09 AM David Eichmann mailto:davide@well-typed.com> wrote:

Ooops I've let this thread pass by me, but this is most likely due to my latest refactoring of libffi rules in Hadrian. No need to create a new issue, I'll just reopen #16304, and have a look at this now.

Hate to nag - any updates? Also, did you mean #16653? What ticket should I subscribe to

-- Shayne Fletcher Language Engineer c: +1 917 699 7763 e: shayne.fletcher@daml.com mailto:shayne.fletcher@daml.com Digital Asset Holdings, LLC 4 World Trade Center 150 Greenwich Street, 47th Floor https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g New York, NY 10007, USA https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g digitalasset.com http://www.digitalasset.com/

This message, and any attachments, is for the intended recipient(s) only, may contain information that is privileged, confidential and/or proprietary and subject to important terms and conditions available at http://www.digitalasset.com/emaildisclaimer.html http://www.digitalasset.com/emaildisclaimer.html. If you are not the intended recipient, please delete this message.

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

-- Alp Mestanogullari, Haskell Consultant Well-Typed LLP, https://www.well-typed.com/ Registered in England and Wales, OC335890 118 Wymering Mansions, Wymering Road, London, W9 2NF, England

Hi Alp, On Fri, May 24, 2019 at 11:34 AM Shayne Fletcher wrote:

On Fri, May 24, 2019 at 7:33 AM Alp Mestanogullari wrote:

...

Hello Shayne,

David and I figured out the cause of that problem, I am working on a patch, will put it up as a (WIP) MR as soon as it's ready.

Awesome!

Confused. I see 382dc918 ("Hadrian: always generate the libffi dynlibs manifest with globbing") has landed but I'm still getting, ``` 2019-05-30T10:47:57.5002328Z Error, file does not exist and no rule available: 2019-05-30T10:47:57.5002606Z _build/stage1/libffi/build/inst/bin/libffi-6.dll ``` What am I missing? -- Shayne Fletcher Language Engineer c: +1 917 699 7763 e: shayne.fletcher@daml.com Digital Asset Holdings, LLC 4 World Trade Center 150 Greenwich Street, 47th Floor https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g New York, NY 10007, USA https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g digitalasset.com http://www.digitalasset.com/ -- This message, and any attachments, is for the intended recipient(s) only, may contain information that is privileged, confidential and/or proprietary and subject to important terms and conditions available at  http://www.digitalasset.com/emaildisclaimer.html http://www.digitalasset.com/emaildisclaimer.html. If you are not the intended recipient, please delete this message.

Heh, it landed less than an hour ago. If the error persists, you'll probably want to do a clean build, by removing at the very least _build/stage1, even though I'd say it can't hurt to remove _build altogether. Along with that patch, I also brought back the Windows CI job and the error did not show up there, so I'm pretty confident you will not run into it again with a fresh, clean build. Do feel free to contact me or open a ticket if somehow the error still shows up, even against a clean tree. On 30/05/2019 13:01, Shayne Fletcher wrote:

On Thu, May 30, 2019 at 6:55 AM Shayne Fletcher mailto:shayne.fletcher@daml.com> wrote:

Hi Alp,

On Fri, May 24, 2019 at 11:34 AM Shayne Fletcher mailto:shayne.fletcher@daml.com> wrote:

On Fri, May 24, 2019 at 7:33 AM Alp Mestanogullari mailto:alp@well-typed.com> wrote:

Hello Shayne,

David and I figured out the cause of that problem, I am working on a patch, will put it up as a (WIP) MR as soon as it's ready.

Awesome!

Confused. I see 382dc918 ("Hadrian: always generate the libffi dynlibs manifest with globbing") has landed but I'm still getting, ``` 2019-05-30T10:47:57.5002328Z Error, file does not exist and no rule available: 2019-05-30T10:47:57.5002606Z _build/stage1/libffi/build/inst/bin/libffi-6.dll ``` What am I missing?

Golden rule of programming : when your program doesn't work and you've checked everything and are sure everything is right and your program still doesn't work then, one of the things you are sure of is wrong :)

My front-running guess is that the patch hasn't landed. So many confusing notifications!! Sorry for the noise and fingers crossed for its progress in the merge queue!

-- Shayne Fletcher Language Engineer c: +1 917 699 7763 e: shayne.fletcher@daml.com mailto:shayne.fletcher@daml.com Digital Asset Holdings, LLC 4 World Trade Center 150 Greenwich Street, 47th Floor https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g New York, NY 10007, USA https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g digitalasset.com http://www.digitalasset.com/

This message, and any attachments, is for the intended recipient(s) only, may contain information that is privileged, confidential and/or proprietary and subject to important terms and conditions available at http://www.digitalasset.com/emaildisclaimer.html http://www.digitalasset.com/emaildisclaimer.html. If you are not the intended recipient, please delete this message.

-- Alp Mestanogullari, Haskell Consultant Well-Typed LLP, https://www.well-typed.com/ Registered in England and Wales, OC335890 118 Wymering Mansions, Wymering Road, London, W9 2NF, England

I haven't (yet) worked on making an OS X job for Hadrian. I brought back the Windows/Hadrian CI job with that fix, but then tried running the testsuite, which didn't go well, see: https://gitlab.haskell.org/ghc/ghc/merge_requests/1035 https://gitlab.haskell.org/alp/ghc/-/jobs/87467 This seems to have brought back a problem that Ben and I started looking into and which led us to disabling that job in the first place... However, running the testsuite is one thing, but just making sure we can build GHC is another. I guess I can put together a patch to get us an OS X job that just builds GHC pretty quickly, and then later get both the Windows and OS X jobs to go as far as running the testsuite. That'd at least give us confidence that people can build GHC with Hadrian just fine on OS X. Regarding the resources, that certainly sounds like something that would help if we are to double the number of OS X builds, indeed. I'm sure Ben (cc'd), who handles those matters, would be happy to talk about this with you! In the meantime, I will prepare a patch to add an OSX/Hadrian job. On 30/05/2019 15:04, Shayne Fletcher wrote:

On Thu, May 30, 2019 at 8:32 AM Alp Mestanogullari mailto:alp@well-typed.com> wrote:

Heh, it landed less than an hour ago.

If the error persists, you'll probably want to do a clean build, by removing at the very least _build/stage1, even though I'd say it can't hurt to remove _build altogether. Along with that patch, I also brought back the Windows CI job and the error did not show up there, so I'm pretty confident you will not run into it again with a fresh, clean build.

It's cooking now. Looking promising :) I build fresh every time. Good news on the Windows CI!! On that note, how are we doing on the MacOS front ? (Let me know if we need resources to make that happen - might be able to help if so!).

Do feel free to contact me or open a ticket if somehow the error still shows up, even against a clean tree.

You bet. Thanks for your hard work!

-- Shayne Fletcher Language Engineer c: +1 917 699 7763 e: shayne.fletcher@daml.com mailto:shayne.fletcher@daml.com Digital Asset Holdings, LLC 4 World Trade Center 150 Greenwich Street, 47th Floor https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g New York, NY 10007, USA https://maps.google.com/?q=150+Greenwich+Street,+47th+Floor%C2%A0+%C2%A0+%C2%A0+%C2%A0+%C2%A0+New+York,+NY+10007,+USA&entry=gmail&source=g digitalasset.com http://www.digitalasset.com/

This message, and any attachments, is for the intended recipient(s) only, may contain information that is privileged, confidential and/or proprietary and subject to important terms and conditions available at http://www.digitalasset.com/emaildisclaimer.html http://www.digitalasset.com/emaildisclaimer.html. If you are not the intended recipient, please delete this message.

-- Alp Mestanogullari, Haskell Consultant Well-Typed LLP, https://www.well-typed.com/ Registered in England and Wales, OC335890 118 Wymering Mansions, Wymering Road, London, W9 2NF, England

Shayne Fletcher via ghc-devs writes:

On Thu, May 30, 2019 at 8:32 AM Alp Mestanogullari wrote:

...

Heh, it landed less than an hour ago.

If the error persists, you'll probably want to do a clean build, by removing at the very least _build/stage1, even though I'd say it can't hurt to remove _build altogether. Along with that patch, I also brought back the Windows CI job and the error did not show up there, so I'm pretty confident you will not run into it again with a fresh, clean build.

It's cooking now. Looking promising :) I build fresh every time. Good news on the Windows CI!! On that note, how are we doing on the MacOS front ? (Let me know if we need resources to make that happen - might be able to help if so!).

Hi Shayne, If you think you may be able to help with MacOS resources let's chat. This is certainly an area where we could use help. Currently we have two boxes which are generously provided to us by Davean Scies. These tend to struggle with the load and consequently jobs sometimes timeout before they are able to get a builder. It would be great to be able to do better here. Cheers, - Ben