haddock test suite

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

haddock test suite

Richard Eisenberg-2
Hi all,

 

I've implemented the proposal at
http://hackage.haskell.org/trac/ghc/wiki/NewAxioms/Nonlinearity, which
involved a few changes to HsDecl datatypes (FamilyDecl, TyFamInstDecl,
etc.), thus requiring a change to Haddock. I tried to be as minimal as I
could there, but prudence tells me I should run the haddock test suite. How
do I do so? The instructions seem to say I should just `cabal test`, but
that requires `cabal build`, which fails, seemingly because of import path
issues:

 

cam-05-unx:10:59:54 /5playpen/eir/ghc/head/utils/haddock> cabal build

Building haddock-2.13.2...

Preprocessing library haddock-2.13.2...

 

src/Haddock/Convert.hs:29:8:

    Could not find module `CoAxiom'

    Use -v to see a list of the files searched for.

 

 

Any advice?

 

Separately, I see that Haddock does not generate any output for type family
instances. I'd be happy to bring someone up to speed about the new type
families and such so that this gap could be closed.

 

Thanks!

Richard

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.haskell.org/pipermail/ghc-devs/attachments/20130621/130986b0/attachment.htm>

Reply | Threaded
Open this post in threaded view
|

haddock test suite

Mateusz Kowalczyk
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 21/06/13 11:08, Richard Eisenberg wrote:

> Hi all,
>
>
>
> I've implemented the proposal at
> http://hackage.haskell.org/trac/ghc/wiki/NewAxioms/Nonlinearity,
> which involved a few changes to HsDecl datatypes (FamilyDecl,
> TyFamInstDecl, etc.), thus requiring a change to Haddock. I tried
> to be as minimal as I could there, but prudence tells me I should
> run the haddock test suite. How do I do so? The instructions seem
> to say I should just `cabal test`, but that requires `cabal build`,
> which fails, seemingly because of import path issues:
>
>
>
> cam-05-unx:10:59:54 /5playpen/eir/ghc/head/utils/haddock> cabal
> build
>
> Building haddock-2.13.2...
>
> Preprocessing library haddock-2.13.2...
>
>
>
> src/Haddock/Convert.hs:29:8:
>
> Could not find module `CoAxiom'
>
> Use -v to see a list of the files searched for.
>
>
>
>
>
> Any advice?
>
>
>
> Separately, I see that Haddock does not generate any output for
> type family instances. I'd be happy to bring someone up to speed
> about the new type families and such so that this gap could be
> closed.
>
>
>
> Thanks!
>
> Richard
>
>
>
>
> _______________________________________________ ghc-devs mailing
> list ghc-devs at haskell.org
> http://www.haskell.org/mailman/listinfo/ghc-devs
>
This means your GHC is too old. Compile from git and try again.

Out of interest, can you post a link to the changes you made to
Haddock if they are up online?

- --
Mateusz K.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
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=+izQ
-----END PGP SIGNATURE-----


Reply | Threaded
Open this post in threaded view
|

haddock test suite

Richard Eisenberg-2
> -----Original Message-----
> From: ghc-devs-bounces at haskell.org [mailto:ghc-devs-
> bounces at haskell.org] On Behalf Of Mateusz Kowalczyk
>
> This means your GHC is too old. Compile from git and try again.

I don't think that's it -- I merged with HEAD this morning.

>
> Out of interest, can you post a link to the changes you made to
> Haddock if they are up online?
>

I just pushed my commit to my github fork:
https://github.com/goldfirere/haddock/tree/overlapping-tyfams

The commit history may be a little mangled there (I think because I don't
bother to update master on my fork), but this link shows you the changes I
made today:
https://github.com/goldfirere/haddock/commit/66a142f3926d3be0cf9aab97bbc24ce
bfe6b3b16

(I realize it's redundant to have a fork and then a branch within that fork,
but it allows for easy integration when I do want to commit to the branch on
the main repo.)

Comments on the code are welcome.

Richard



Reply | Threaded
Open this post in threaded view
|

haddock test suite

Mateusz Kowalczyk
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 21/06/13 12:23, Richard Eisenberg wrote:
>> -----Original Message----- From: ghc-devs-bounces at haskell.org
>> [mailto:ghc-devs- bounces at haskell.org] On Behalf Of Mateusz
>> Kowalczyk
>>
>> This means your GHC is too old. Compile from git and try again.
>
> I don't think that's it -- I merged with HEAD this morning.
>
That's most strange. Whenever I encounter this CoAxiom error, it has
always just been the case of using a more recent GHC and this has been
consistent across multiple boxes. In fact, I get this error with 7.6.3
(and I don't with 7.7).

>>
>> Out of interest, can you post a link to the changes you made to
>> Haddock if they are up online?
>>
>
> I just pushed my commit to my github fork:
> https://github.com/goldfirere/haddock/tree/overlapping-tyfams
>
> The commit history may be a little mangled there (I think because I
> don't bother to update master on my fork), but this link shows you
> the changes I made today:
> https://github.com/goldfirere/haddock/commit/66a142f3926d3be0cf9aab97bbc24ce
>
>
bfe6b3b16
>
> (I realize it's redundant to have a fork and then a branch within
> that fork, but it allows for easy integration when I do want to
> commit to the branch on the main repo.)
Is it redundant? I do exactly this. It's easy to merge in changes from
upstream while developing on a separate branch.

>
> Comments on the code are welcome.
>
> Richard
>


- --
Mateusz K.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
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=5VsX
-----END PGP SIGNATURE-----