VBA Rubberducking

12:01 AM

RELOAD!

[Phrancis/TwitterTools] 6 commits. 100 additions. 63 deletions.

[rubberduck-vba/Rubberduck] 12 commits. 2 opened issues. 2 closed issues. 7 issue comments. 268 additions. 2692 deletions.

[Minesweeper] Games Played: 95, Bombs Used: 50, Moves Performed: 11842, New Users: 163

Mathieu Guindon

@M.Doerner as long as assignments don't point to the last resolved thing in the call chain, all is good I think.

Mathieu Guindon

12:46 AM

wow reddit, really?

this

1:06 AM

~sigh~ why, C#?

?input == "abc"
true
mocked.Drives[input]
null
mocked.Drives["abc"]
{Castle.Proxies.DriveProxy}
    AvailableSpace: null
    DriveLetter: null
    DriveType: UnknownType
    FileSystem: null
    FreeSpace: null
    IsReady: false
    Mock (Castle.Proxies.DriveProxy): {Moq.Mock<Scripting.Drive>}
    Mock (Castle.Proxies.DriveProxy): {Moq.Mock<Scripting.Drive>}
    Mock: {Moq.Mock<Scripting.Drive>}
    Path: "foobar"
    RootFolder: null
    SerialNumber: 0
    ShareName: null
    TotalSize: null

Mathieu Guindon

Huh

...encoding?

(grasping at straws)

this

input = "abc"
"abc"
ds[input]
null
ds["abc"]
{Castle.Proxies.DriveProxy}
    AvailableSpace: null
    DriveLetter: null
    DriveType: UnknownType
    FileSystem: null
    FreeSpace: null
    IsReady: false
    Mock (Castle.Proxies.DriveProxy): {Moq.Mock<Scripting.Drive>}
    Mock (Castle.Proxies.DriveProxy): {Moq.Mock<Scripting.Drive>}
    Mock: {Moq.Mock<Scripting.Drive>}
    Path: "foobar"
    RootFolder: null
    SerialNumber: 0
    ShareName: null
    TotalSize: null
    VolumeName: null

the assignment should have had handled the encoding.... right?

Mathieu Guindon

it can't be that. Is the indexer mock-setup?

this

yes

Mathieu Guindon

stupid idea: could it be taking the argument literally?

this

1:16 AM

var mockDrive = mockDrives.SetupChildMock("Item", "abc");

i originally had var mockDrive = mockDrives.SetupChildMock("Item", provider.It.Is("abc"));

Mathieu Guindon

no mug, that doesn't make sense either

this

you gave me an idea, though, it mgiht be my It taht's flawed

trying with Moq's It

var mockDrive = mockDrives.SetupChildMock("Item", It.Is<string>(x => x == "abc"));

Mathieu Guindon

You're building expression trees - could be that somehow the input argument is taken as a literal that doesn't equal the literal expression "abc". The distinction is rather subtle between the two

this

var mockDrive = mockDrives.SetupChildMock("Item", provider.It.IsAny()); passes

Mathieu Guindon

there's that =)

this

1:22 AM

so it is working --- just not with a specific value

Mathieu Guindon

are you evaluating the expression somewhere?

this

i'm just dispatching the expressions to Moq

Mathieu Guindon

like, input needs to resolve to "abc"

hmm

this

well, that's not accurate

I do have to set up the epxression tree for invoking Moq's stuff

e.g. the equivalent of mockDrives.Setup(x => x["abc"]).Returns(mockDrive.Object)

This is the actual expression built:

"(mock, value0) => Convert(mock.As().Setup(x => x.get_Item(Is(p00 => (p00 == Convert(\"abc\")))))).Returns(value0)"

(note that this isn't one-one mapping, so there's lot of weird stuff that you might not see in an actual C#. oh the joy)

hmm.

Mathieu Guindon

O__O

this

1:33 AM

IKR?

IvenBach

3:08 AM

@FreeMan I had that in mind when I made that statement. Xkcd has many references.

Peter M Taylor

3:23 AM

@this a mouthful if you ask me

@this that weird stuff could become aliases

Quack Exchange

6:01 AM

0

Q: Factory Class That Uses Reflection to Instantiate Objects by Class Name

In this SO post Dynamically choose class from string Alexander Platonov creates a Function on the fly and then uses Application.Run() to instantiate an Object by it's class name. Eureka Reflection in VBA!! How evil! How inspiring!! So that began my journey into the darkness. My quest for powe...

vba

IvenBach

6:15 AM

dealing with non technical people having computer problems

Make Excel give back the shrimp.

Bed time or I’ll take inundate with more Sheldon links.

M.Doerner

10:54 AM

I have to say, hunting down a stack overflow exception is no fun.

M.Doerner

11:18 AM

Arrgh, this is annoying: we get a SO because we use a recursive visitor on a structure that is one level too deep.

I just have no idea how that structure could get so deep.

And I do not really know which test case this is because the call stack is truncated.

M.Doerner

12:13 PM

Oh, this is really subtle.

We have a problem with the resolution failed expressions we return when with expressions are involved.

That can lead to cycles.

M.Doerner

1:14 PM

Ok, I have changed things to produce an acyclic graph. However, it is not a tree.

It actually does not look that much different than before, but it is set up such that we actually collect all unresolved context, but not those only generated to join several unresolved expressions.

Duga

2:15 PM

[rubberduck-vba/Rubberduck] MDoerner edited pull request #5089: Resolve recursive and parameterless default member accesses

Duga

2:40 PM

[rubberduck-vba/Rubberduck] MDoerner edited pull request #5089: Resolve recursive and parameterless default member accesses

Duga

4:20 PM

[rubberduck-vba/Rubberduck] MDoerner synchronized pull request #5089: Resolve recursive and parameterless default member accesses

[rubberduck-vba/Rubberduck] MDoerner removed label PR-Status: WIP from pull request #5089: Resolve recursive and parameterless default member accesses

[rubberduck-vba/Rubberduck] MDoerner added label PR-Status: Review Requested to pull request #5089: Resolve recursive and parameterless default member accesses

[rubberduck-vba/Rubberduck] MDoerner created comment on pull request #5089: Resolve recursive and parameterless default member accesses

> Now, I have fixed the two remaining resolver issues I wanted to deal with in this PR, the `MemberNotOnInterfaceInspection` not working for procedure arguments (and, as I had to realize, in a lot more places) and generating non-undeclared array declarations instead of undeclared non-array declarations for ReDim statements on arrays without a previous declaration.

I also took the liberty to strip the parentheses from the `AsTypeName` of functions and properties returning an array. This was in

[rubberduck-vba/Rubberduck] MDoerner edited comment on pull request #5089: Resolve recursive and parameterless default member accesses

> Now, I have fixed the two remaining resolver issues I wanted to deal with in this PR, the `MemberNotOnInterfaceInspection` not working for procedure arguments (and, as I had to realize, in a lot more places) and generating non-undeclared array declarations instead of undeclared non-array declarations for ReDim statements on arrays without a previous declaration.

I also took the liberty to strip the parentheses from the `AsTypeName` of functions and properties returning an array. This was in

[rubberduck-vba/Rubberduck] build for commit 3a915a48 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] MDoerner synchronized pull request #5089: Resolve recursive and parameterless default member accesses

M.Doerner

4:40 PM

@Duga Forgot to commit my last changes.

Mathieu Guindon

The parens were more of a visual cue in the RD cmdbar that the selected variable/member is/returns an array though. Does the cmdbar label indicate IsArray status without it?

Duga

[rubberduck-vba/Rubberduck] build for commit 617b27f0 on unknown branch: AppVeyor build succeeded

M.Doerner

5:17 PM

I don't know.

However, array variables never had parens in their AsTypeName.

Array variables show like bar() (variable:Long) and array returning functions showed as Foo (function:Long()).

That is really inconsitent.

Both should have Long() and none parentheses in the name.

Btw, there is no test specifying what the presentation should be.

Duga

[rubberduck-vba/Rubberduck] MDoerner synchronized pull request #5089: Resolve recursive and parameterless default member accesses

Duga

5:36 PM

[rubberduck-vba/Rubberduck] build for commit 6aa489bb on unknown branch: AppVeyor build succeeded

Mathieu Guindon

5:55 PM

@M.Doerner ah, all good then - the parens must be added by the statuslabel currentselection formatting code

M.Doerner

6:58 PM

I found that already, in the ContextFormatter.

Do we want to preserve the current behaviour that the parens for variables are at the variable and not at the type?

It would be easy to switch it around or the add the parens in the type name.

M.Doerner

7:13 PM

Both options would actually simplify the code.

I think I will add the parens to the type as not doing this would complicate my fix for the command bar text of array returning functions.

Oh, how I love convert to switch.

That makes thigs so much more obvious.

Mathieu Guindon

7:29 PM

@M.Doerner works for me! :)

M.Doerner

Should I keep adding the parens after the identifier name for array variables?

Personally, I do not really care much.

Mathieu Guindon

not if they're after the type too

M.Doerner

OK, then they are gone now.

I hope I did not mess up on the refactoring of the formatter.

Mathieu Guindon

7:52 PM

IIRC that formatter was quite a dreadful sight..

M.Doerner

Well, when it was written, string interpolation did not exist yet.

Duga

[rubberduck-vba/Rubberduck] MDoerner synchronized pull request #5089: Resolve recursive and parameterless default member accesses

Duga

8:16 PM

[rubberduck-vba/Rubberduck] build for commit 02bda4f7 on unknown branch: AppVeyor build succeeded

QHarr

9:10 PM

May be of interest:

0

Q: VBA inheritance

I heard a lot of that VBA is lacking inheritance. I did a little workaround and now it seems to me that it is exactly what inheritance does. I'm far from pro =) and may be missing something. So I would really appreciate your thoughts about possible downsides. I was very surprised when found that...

vba inheritance

Mathieu Guindon

10:15 PM

@QHarr wrote an answer, hope it helps

QHarr

10:32 PM

@MathieuGuindon It does. Thank you.

Transcript for