Re: Duplicated typedef trips my gcc over

I have just committed a tentative fix I am using for a while, and keep watching the buildbots. Cheers, Gabor On 10/26/15, Christiaan Baaij wrote: > I cannot build HEAD for the same reason. Same GCC version. > > On 21 October 2015 at 13:08, Gabor Greif

Re: Duplicated typedef trips my gcc over

Gabor Greif writes: > Hi all, > > look: > > $ git grep "typedef struct LibDwSession_ " > rts/Libdw.c:typedef struct LibDwSession_ LibDwSession; > rts/Libdw.h:typedef struct LibDwSession_ LibDwSession; > > $ which gcc > /usr/bin/gcc > > $ gcc --version > gcc (GCC) 4.4.7 20120313

RE: Sphinx on Windows

Thanks. But it's a bit inefficient to debug by email. For now I've simply switched off the sphinx stuff in my build.mk. Might you work it through when you have access to a machine, and document the result? If the desired sequence is to * get pip * use pip to get sphinx then could the

RE: Sphinx on Windows

Simon Peyton Jones writes: > I have not tried that. But 'pip' gives "command not found". > > Maybe I need something in my path? I appear to have python 2.7.5 installed; > it does not appear to have pip.exe. > Pip ships with Python 2.7.9 and later. Alternatively, you

Re: Duplicated typedef trips my gcc over

I cannot build HEAD for the same reason. Same GCC version. On 21 October 2015 at 13:08, Gabor Greif wrote: > Hi all, > > look: > > $ git grep "typedef struct LibDwSession_ " > rts/Libdw.c:typedef struct LibDwSession_ LibDwSession; > rts/Libdw.h:typedef struct LibDwSession_

Re: Sphinx on Windows

Grab https://bootstrap.pypa.io/get-pip.py and do `python get-pip.py' from a command prompt (might need to be elevated, not sure). After that 'pip' should be available and you can do 'pip install sphinx'. On Mon, Oct 26, 2015 at 10:41 AM, Simon Peyton Jones wrote: > I have

RE: Sphinx on Windows

I have not tried that. But 'pip' gives "command not found". Maybe I need something in my path? I appear to have python 2.7.5 installed; it does not appear to have pip.exe. Simon | -Original Message- | From: David Kraeutmann [mailto:k...@kane.cx] | Sent: 26 October 2015 08:32 |

Re: Sphinx on Windows

'pip' is just a Python package manager. Following the steps at http://sphinx-doc.org/latest/install.html#windows-install-python-and-sphinx should give you a working sphinx install. I can test those on my Windows machine once I'm home, but that won't be for a few hours. On Mon, Oct 26, 2015 at

Sphinx on Windows

The GHC wiki says you need Sphinx to build docs, https://ghc.haskell.org/trac/ghc/wiki/Building/Preparation/Tools and refers to the Preparation doc for how to install Sphinx https://ghc.haskell.org/trac/ghc/wiki/Building/Preparation But the Windows sub-page of Preparation does not say

Re: Segfault in a CoreLinted program (and a GHC-generated Core question)

OK, thanks to people at IRC channel(especially @rwbarton) I realized that my lint calls were not actually running, simply because I wasn't using -dcore-lint.. I didn't know such a flag exists, and even with the absence of the flag I'd expect a core lint would work, because I'm explicitly calling

RE: can't get haddock patch

| Regardless, I can take care of it from here. great. But I could not validate because of this, so may not be finished. Would it make sense to push changes to the haddock repo (as I have done for my other inflight changes)? Anyway, if you'd like me to take this any further, let me know

Re: Warning suppression pragmas

Hi! I think SUPPRESS could be enough. Hope this helps On Mon, Oct 26, 2015 at 8:59 PM, Эдгар Жаворонков wrote: > Hi Ben! > Thanks for your feedback > I thought a little and fixed some in wiki page > > In my opinion SUPPRESS_WARNINGS is not a real long name for pragma,

Can't build on Windows

Gurgah. GHC is broken on my windows laptop. See below. Anyone have any ideas? Frustrating. Oddly the symbol is defined in the .exe bash$ nm inplace/bin/ghc-stage2.exe | grep stg_upd_frame 02c248e0 T _stg_upd_frame_info This is admittedly on a branch that I have modified; but I have only

RE: Can't build on Windows

But don't our windows buildbots do this for us? Surely! That's what they are for. Me bisecting on my little laptop is not a very happy situation. Don't we have better technology? Simon | -Original Message- | From: mad@gmail.com [mailto:mad@gmail.com] On Behalf Of Austin |

Re: recomp013

I couldn't repro on a devel2 build. Edward Excerpts from Simon Peyton Jones's message of 2015-10-26 08:24:00 -0700: > driver/recomp013 is failing, on Linux, on HEAD for me: > > Actual stdout output differs from expected: > > --- ./driver/recomp013/recomp013.stdout.normalised2015-10-26 >

Re: Can't build on Windows

(Resending to list properly) Hrm, I can't see anything off hand in the past few weeks that would have caused this at a glance. Can you try bisecting it perhaps? You can probably do something like: $ git bisect start $ git bisect bad $ git bisect good dcc342870b4d8a739ccbed3ae26e84dcc3579914 $

RE: Can't build on Windows

It happens with unmodified HEAD too. Sigh. Can anyone help? Simon From: ghc-devs [mailto:ghc-devs-boun...@haskell.org] On Behalf Of Simon Peyton Jones Sent: 26 October 2015 21:19 To: ghc-devs@haskell.org Subject: Can't build on Windows Gurgah. GHC is broken on my windows laptop. See below.

Re: Can't build on Windows

I can not reproduce this on 64bit Windows, a9c93bdd8b027d6de09a3eada7721e7fd2d3e050 builds succesfully with flavour devel2. * Which build flavour are you using? Any other mk/build.mk settings? * 32bit or 64bit? * Output of `git status` (i.e. did you 'git submodule update'? In case that fixes it,

Re: Can't build on Windows

Same here, can't reproduce on 64bit Windows 10. Edward Excerpts from Thomas Miedema's message of 2015-10-26 17:54:35 -0700: > I can not reproduce this on 64bit > Windows, a9c93bdd8b027d6de09a3eada7721e7fd2d3e050 builds succesfully with > flavour devel2. > > * Which build flavour are you using?

Re: can't get haddock patch

Simon Peyton Jones writes: > Ben > What does this mean? (below) Does it mean you have not pushed something to > the haddock repo> > This happens when I'm on branch wip/T9858-typeable-ben2, in the main repo > > Simon > > git submodule update > > fatal: reference is not a

Re: Warning suppression pragmas

Hi Ben! Thanks for your feedback I thought a little and fixed some in wiki page In my opinion SUPPRESS_WARNINGS is not a real long name for pragma, but i can't come up with whort name that clearly describes purpose of such pragma =( Your suggestions are really welcome) --- С уважением,

Re: Sphinx on Windows

The Sphinx + GHC HEAD on Windows story is somewhat complicated. Incidentally I've just managed to get it all working, thus I can share some of my knowledge: 1. Sphinx + GHC HEAD *doesn't work* on Windows if unmodified. The problem is in 'mkUserGuidePart' utility. It crashes when trying to

Re: Warning suppression pragmas

Hello Richard! Can you take a look at some sort of specification i wrote week ago? I placed it here: https://ghc.haskell.org/trac/ghc/wiki/Design/LocalWarningPragmas I can imagine only two use cases, and i'm not sure about correcteness of it --- С уважением, Жаворонков Эдгар Best regards, Edgar

RE: Better calling conventions for strict functions (bang patterns)?

Ryan Yes, I’m sure there is room for improvement here! Good observations. In particular, while I’d like DATA to support this, I believe that most ‘eval’s find that the thing being evaluated is already evaluated; so the fast-path should be that case. We should bet for evaluated. NB,

Re: [GHC] #11019: ApiAnnotations: Make all RdrName occurences Located

At the moment in HaRe we do use the wrapped location, but it results in code that is much more complex. This change is more about bringing predictability to the AST than anything else. If anyone feels strongly that it should not happen then we can live without it. On Mon, Oct 26, 2015 at 2:11

RE: Can't push to haddock repo

Simon Peyton Jones writes: > Well in utils/haddock/.git/config I see > [remote "origin"] > url = git://git.haskell.org/haddock.git > pushurl = ssh://g...@git.haskell.org/haddock.git > Note the difference in hostnames,

Can't push to haddock repo

I can't update the haddock repository! (I want to add a new branch for my in-flight work.) Can anyone help? Thanks Simon git push --set-upstream origin wip/spj-wildcard-refactor remote: W refs/heads/wip/spj-wildcard-refactor haddock simonpj DENIED by refs/.* remote: error: hook declined to

Re: Can't push to haddock repo

In the GHC tree: $ grep haddock packages utils/haddock- - ssh://g...@github.com/haskell/haddock.git says that the upstream repo for Haddock is ssh://... (there are some comments in the 'packages' file that explain what the last column

RE: Can't push to haddock repo

Well in utils/haddock/.git/config I see [remote "origin"] url = git://git.haskell.org/haddock.git pushurl = ssh://g...@git.haskell.org/haddock.git So that looks right. Why would git push not work? Simon | -Original Message- | From: Herbert Valerio Riedel

RE: Can't push to haddock repo

ah yes, my bad. Thanks | -Original Message- | From: Ben Gamari [mailto:b...@well-typed.com] | Sent: 26 October 2015 13:55 | To: Simon Peyton Jones; Herbert Valerio Riedel | Cc: ghc-devs@haskell.org | Subject: RE: Can't push to haddock repo | | Simon Peyton Jones

recomp013

driver/recomp013 is failing, on Linux, on HEAD for me: Actual stdout output differs from expected: --- ./driver/recomp013/recomp013.stdout.normalised2015-10-26 14:26:51.920153558 + +++ ./driver/recomp013/recomp013.run.stdout.normalised 2015-10-26 14:26:51.920153558 + @@ -3,5

RE: Sphinx on Windows

Thanks. I'll just disable it for now. When you have this sorted out, could you add something to the Windows instructions for GHC devs? That'd be super-helpful Simon | -Original Message- | From: ghc-devs [mailto:ghc-devs-boun...@haskell.org] On Behalf Of kyra | Sent: 26 October

Re: Sphinx on Windows

kyra writes: > The Sphinx + GHC HEAD on Windows story is somewhat complicated. > Incidentally I've just managed to get it all working, thus I can share > some of my knowledge: > Thanks for writing this down, Kyra! These notes are very helpful. Would you like to add a comment to

Re: Warning suppression pragmas

Hi Simon! It is a name of particular warning. By the way, you made me think of more general idea. So i think it would be enough to suppress all kinds of warnings that a function can throw(and a bit easier for me as a total newbee in GHC hacking =) ) --- С уважением, Жаворонков Эдгар Best

RE: Segfault in a CoreLinted program (and a GHC-generated Core question)

| So my questions are: Am I right in assuming that CoreLint accepted programs | should not segfault? Yes. Modulo unsafeCoerce, and FFI calls. | What about internal invariants? Should CoreLint check | for those? Is there any pass that checks for invariants and prints helpful | messages in

can't get haddock patch

Ben What does this mean? (below) Does it mean you have not pushed something to the haddock repo> This happens when I'm on branch wip/T9858-typeable-ben2, in the main repo Simon git submodule update fatal: reference is not a tree: 289ef817aad02c341beb6d4c28ba0495872f5a0f Unable to checkout

Re: Segfault in a CoreLinted program (and a GHC-generated Core question)

Ömer Sinan Ağacan writes: > I have a very simple Core plugin that generates some functions. After my > Core-to-Core pass is done, I'm running the linter to make sure the Core > generated by my plugin is well-formed and well-typed. However, even though > lint > checker

Segfault in a CoreLinted program (and a GHC-generated Core question)

I have a very simple Core plugin that generates some functions. After my Core-to-Core pass is done, I'm running the linter to make sure the Core generated by my plugin is well-formed and well-typed. However, even though lint checker passes, the code generated by my plugin is failing with a

Re: Warning suppression pragmas

Эдгар Жаворонков writes: > Hello Richard! > > Can you take a look at some sort of specification i wrote week ago? > I placed it here: > https://ghc.haskell.org/trac/ghc/wiki/Design/LocalWarningPragmas > I can imagine only two use cases, and i'm not sure about

Re: [DISARMED] RE: Better calling conventions for strict functions (bang patterns)?

Simon, I really enjoyed reading this paper... I was wondering if you could comment on the implementation of Strict Core? Was it ever implemented in GHC (even as a proof-of-concept)? If not, was it just due to a lack of time or some fundamental limitation or problem discovered after the paper? If