Inline imports

[stoeffel]

not sure if I'm happy with this yet, though. maybe it should have it's own prefix.
see example/Mock.elm

closes #7 #1

[stoeffel]

cc @rtfeldman since this was your idea.

[stoeffel]

not sure if I'm happy with this yet, though. maybe it should have it's own prefix.

I changed my mind. I think this is better than having a special prefix for imports. I also think it's better than the current implementation. It's more obvious what modules are used for the examples and it allows to use special dev-modules.

[stoeffel]

@eeue56 any thoughts about this?

[eeue56]

@stoeffel I think something that would be extra helpful would be the ability to define variables too in a let binding

[stoeffel]

Will think about that, but I would open a new pr for that. Will create an issue for let binding.

[stoeffel]

I think something that would be extra helpful would be the ability to define variables too in a let binding

I opened an issue for that #11

[stoeffel]

I wonder if we should allow:

{-|
    >>> import MyModule exposing
    ...     ( foo
    ...    , bar
    ...    )
-}

[rtfeldman]

I'd say default to no and see if people request it.

…

On Wed, Nov 30, 2016, 11:01 AM Christoph Hermann ***@***.***> wrote: I wonder if we should allow: {-| >>> import MyModule exposing ... ( foo ... , bar ... )-} — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#10 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABCxwMP2VsDUceNB-Qf11eY7idhbPuaSks5rDcgkgaJpZM4K-EjQ> .

[eeue56]

multiline for variables and the "test" is something I've needed - for modules, probably will too. see here -> https://github.com/eeue56/elm-lint/blob/master/src/Rules/Breakdown.elm#L14 On Wed, Nov 30, 2016 at 10:34 PM, Richard Feldman <notifications@github.com> wrote:

…

I'd say default to no and see if people request it. On Wed, Nov 30, 2016, 11:01 AM Christoph Hermann ***@***.*** > wrote: > I wonder if we should allow: > > {-| >>> import MyModule exposing ... ( foo ... , bar ... )-} > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <#10 (comment) >, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/ABCxwMP2VsDUceNB- Qf11eY7idhbPuaSks5rDcgkgaJpZM4K-EjQ> > . > — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#10 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABFh_pq58mGbLD_15HSuE4XZyisKMYo0ks5rDfn2gaJpZM4K-EjQ> .

-- Cheers, Noah noah@astraldynamics.co.uk Astral Dynamics Cyf Festri Cefn Y Waen, Deiniolen, Caernarfon, Gwynedd LL55 3NB Company number 07726862 Registered in England and Wales

[stoeffel]

@eeue56 you can already do:

{-|
    >>> indexOfLineWithEquals
    ...    <| Array.fromList
    ...    <| String.lines "a\nb\nc = d\nf = d\n"
    2
-}

see example here https://github.com/stoeffel/elm-doc-test/blob/master/example/Mock.elm#L59

[rtfeldman]

The ... is pretty noisy. 😬

I get the argument for >>> over > but this makes me want to re-advocate for >.

Compare:

{-|
    >>> indexOfLineWithEquals
    ...    <| Array.fromList
    ...    <| String.lines "a\nb\nc = d\nf = d\n"
    2
-}

{-|
    > indexOfLineWithEquals
    |      <| Array.fromList
    |      <| String.lines "a\nb\nc = d\nf = d\n"
    2
-}

[stoeffel]

hmmm... will think about it. maybe we should do an other poll.

[stoeffel]

it doesn't feel noisy to me though. it's how Python and elixir does it too

[stoeffel]

I actually think | is more distracting. It looks like a broken sumtype or |>. Meaning, it resembles actual syntax more than ...

[rtfeldman]

Other alternatives:

{-|
    > indexOfLineWithEquals
    #      <| Array.fromList
    #      <| String.lines "a\nb\nc = d\nf = d\n"
    2
-}

{-|
    > indexOfLineWithEquals
    ~      <| Array.fromList
    ~      <| String.lines "a\nb\nc = d\nf = d\n"
    2
-}

{-|
    > indexOfLineWithEquals
    \      <| Array.fromList
    \      <| String.lines "a\nb\nc = d\nf = d\n"
    2
-}

[rtfeldman]

Also reminded of this because I saw @ento use the > style in an unrelated PR and I ❤️ how it's such clear documentation, it's what people choose to write even if there's no intention to generate tests!

[stoeffel]

@rtfeldman thanks for the input. I created an issue to discuss syntax changes there, because it's unrelated to wether or not we inline imports. #12

@eeue56 @rtfeldman are you okay with inlining imports or is anyone opposed?

[eeue56]

I definitely think sticking with ... is the right syntax for multine things. when it comes to documentation, which doctests are all about, ... makes most visual sense.

seems good to me

[stoeffel]

reminder for @stoeffel: merge this tomorrow and publish new version