I'd still love to pass a VBA class to RD, and have RD execute a method the way @Comintern did by calculating the vtable (but without caveats)... It must be possible, as it can be done with MSXML's OnReadyStateChanged property, where you assign the property an instance of the class, and MSXML is able to call the default member. msdn.microsoft.com/en-us/library/ms757030(v=vs.85).aspx

RELOAD!

[Hosch250/CheckersUI] 1 commit. 6 additions. 1 deletion.

[retailcoder/Rubberduck] 129 commits. 92675 additions. 140069 deletions.

[rubberduck-vba/Rubberduck] 7 opened issues. 1 closed issue. 28 issue comments.

@ThunderFrame Huh. Now that I think about it, we only really need an IUnknown or IDispatch. We could just discover the interface the old-school COM way.

@Comintern What is that XMLHttpRequest.OnReadyStateChange = MyOnReadyStateWrapper statement actually assigning, if it's not using Set? The address of the default procedure?

I'm not sure. What library was that again?

MSXML

Microsoft XMl V6.0

Oh, I meant the dll name

msxml6.dll in System32

[id(0x0000000e), propput, helpstring("Register a complete event handler")]
HRESULT onreadystatechange([in] IDispatch* rhs);

WTH is the rhs annotation?

@Comintern good question... ask @Vogel612? I suppose it made it easier to return multiple values

@Comintern that's the name of the argument?

Since we only use the actual filename in Application.Run, the length of the string should usually not be the problem. Say 40 characters for the filename, 40 for the project, and 31 for the module leave 144 charcters for the name of the test method.

@Mat'sMug Figured it out - it changes the input parameters to allow passing through to the base method. That's actually a bug.

@ThunderFrame LOL. So yes, it takes an Object.

so WTH isn't there a Set?

Yeah, I'm kind of perplexed by that.

@Comintern or a Variant?

@Comintern Oh wait, it's MS Documentation... They're deliberately misleading us?

lol

Nope, that's a hard IDispatch.

Oh f%ck. PermissiveAssertClass has another bug.

AssertClass has the same bug.

String comparisons need to take the Option Compare into account.

hmm IIRC that was deliberately left out

@Comintern Does COM not enforce the difference between a Let and a Set?

Nope. It could care less. A Set takes an interface pointer - a Let takes a value type.

'HelperClass
Private dispObject As Object

Public Property Let PropLet(o As Object)
  Set dispObject = o
End Property

Public Property Set PropSet(o As Object)
  Set dispObject = o
End Property

Public Sub Execute(Arg1 As String)
  dispObject = Arg1
End Sub

Public Function CallAMethod(MethodName As String, Arg1 As String) As String
  CallAMethod = CallByName(dispObject, MethodName, VbMethod, Arg1)
End Function

'CallBackClass
'Can't seem to pass arguments to the default member if it is a procedure, but can pass an argument if it is a property
'@DefaultMember
Public Property Let FooBar(Arg1 As String)
  Debug.Print "Default Member", Arg1
End Property

Public Function DoSomething(Arg1 As String) As String
  DoSomething = "DoSomething did " & Arg1
End Function

'Module1
Sub test()

  Dim arrange As HelperClass
  Set arrange = New HelperClass
  Dim callBack As CallBackClass
  Set callBack = New CallBackClass
  Dim result As String

  arrange.PropLet = callBack
  arrange.Execute "PropLetArgument"
  result = arrange.CallAMethod("DoSomething", "Foo")

  Set arrange.PropSet = callBack
  arrange.Execute "PropSetArgument"
  result = arrange.CallAMethod("DoSomething", "Foo")

End Sub

so CallByName works (as you'd expect), but so does calling the default member.

[rubberduck-vba/Rubberduck] MDoerner created pull request #2858: Exception handling for unit tests and more to merge MDoerner/Rubberduck/stuff into rubberduck-vba/Rubberduck/next

> This and that: 1) Slightly more safety for long procedure names in Excel. (If > 255 chars, drop the filename.) 2) Catching and logging of COMExceptions thrown during running the unit tests. (Application.Run might not like what we are feeding it.) Exception gets logged and test gets set to Inconclusive - Unexpected COM error. 3) Small refactoring of SyncCOMReferences in the ParseCoordinator. Now, the setup for parallelism gets only performed if there is actually soemthing to load.

@Comintern but both Let and Set produce the same outcome

So Let with an object as a parameter is just the pointer?

I'm on a work PC - Does RD get an inspection result for assigning an Object to a property Let?

[rubberduck-vba/Rubberduck] build for commit 29c403af on unknown branch: AppVeyor build succeeded

I would doubt it. I'll check as soon as my test run completes.

WTH? Why is GivenField_NamedUnambiguously_InStatementFieldCallResolvesToFieldDeclaration failing?

@ThunderFrame you mean defining a Property Let member with an object return type? that would be a good one to have!

@Comintern lemme guess - name too long?

@Mat'sMug Defining a property let with an Object parameter.

LOL. No, that's one of the new tests @Hosch250 put in a couple days ago.

 Public Property Let Foo(ByVal value As Object)
 End Property

^ definitely deserves an inspection result

^

^

^

look, a pine tree

Now I want to see if I can use it to break something.

GivenField_NamedUnambiguously_InStatementFieldCallResolvesToFieldDeclaration is doing the weird intermittent failure thing.

related inspection for usage of a reference declaration that accepts an Object as its Parameter... i.e.onreadystatechange

It ambiguously passes.

huh, think I just setup CI for my fork on rubberduckvba.visualstudio

stops clicking new buttons in VS2015

I'd turn it off.

It'll really eat up system resources.

so that's like MS-hub

@Comintern ? it's remote

I guess that assignment ambiguity is the same thing we discussed around assigning a Connection to a Command Object without using Set

[rubberduck-vba/Rubberduck] comintern synchronized pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

@Mat'sMug Maybe I'm thinking of the CodeLens one.

looks like MS wired their own "hub" to VS 2015

it's literally GitHub, MS-flavored

reeks of TFS

Sh!tHub?

[rubberduck-vba/Rubberduck] build for commit 194aa1c2 on unknown branch: AppVeyor build succeeded

> Deleting project "Rubberduck"

[rubberduck-vba/Rubberduck] comintern edited pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

ok. at least rubberduckvba.visualstudio.com is reserved lol

lol

You want to take a quick scan over the permissive SequenceEquals really quick? It's also untestable.

@Comintern h3.dion.ne.jp/~sakatsu/…

That's basically what RD does.

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/Inspections/MoveFieldCloserToUsageInspection.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UnitTesting/IAssert.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

@Comintern calling class members by MemID?

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UnitTesting/PermissiveObjectComparer.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UnitTesting/PermissiveObjectComparer.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

[rubberduck-vba/Rubberduck] comintern opened issue #2859: RD needs an inspection for Property Lets with Object parameters.

[rubberduck-vba/Rubberduck] comintern added label difficulty-02-ducky to issue #2859: RD needs an inspection for Property Lets with Object parameters.

[rubberduck-vba/Rubberduck] comintern added label feature-inspections to issue #2859: RD needs an inspection for Property Lets with Object parameters.

[rubberduck-vba/Rubberduck] comintern added label up-for-grabs to issue #2859: RD needs an inspection for Property Lets with Object parameters.

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UnitTesting/PermissiveObjectComparer.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

@ThunderFrame Well, not explicitly - the RCW marshals the delegate as a function pointer and handles the dispatch requests.

[rubberduck-vba/Rubberduck] comintern commented on RetailCoder.VBE/UnitTesting/IAssert.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

hmm, the example looked like the author was executing instance members by ID as opposed to by name.

Is it good to put in a test for an assumption?

@Comintern just change the NotImplementedException to a NotSupportedException and that PR is good to go

@IvenBach it's good to put in tests that satisfy and that don't satisfy the assumption

@IvenBach in a dream world, tests document the entire program's specifications - so yes, absolutely!

So if I'm assuming that a goalseek is going to have both cells on the same sheet I should test for that?

[rubberduck-vba/Rubberduck] comintern commented on RetailCoder.VBE/UnitTesting/PermissiveObjectComparer.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

@IvenBach depends what you consider part of the specification (your requirements) vs what you consider part of implementation details

If it's part of my classes implementation, that both cells will be on the same sheet, then I should test for it?

[rubberduck-vba/Rubberduck] comintern commented on RetailCoder.VBE/UnitTesting/PermissiveObjectComparer.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

[rubberduck-vba/Rubberduck] comintern commented on RetailCoder.VBE/UnitTesting/PermissiveObjectComparer.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

wtf my internet just died out of nowhere

@Mat'sMug I was reconsidering whether that should throw.

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UnitTesting/PermissiveObjectComparer.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

if obj.GetHashCode() isn't reliable, I'd make it throw

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UnitTesting/PermissiveObjectComparer.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

Yeah, you're probably right - it should return different hashes than the Equals method.

[rubberduck-vba/Rubberduck] comintern synchronized pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

[rubberduck-vba/Rubberduck] comintern commented on RetailCoder.VBE/UnitTesting/PermissiveObjectComparer.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.Parsing/VBA/ParseCoordinator.cs of pull request #2858: Exception handling for unit tests and more

@ThunderFrame Not angry so much as distressed

[rubberduck-vba/Rubberduck] build for commit 64e23b9c on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.Parsing/VBA/ParseCoordinator.cs of pull request #2858: Exception handling for unit tests and more

[rubberduck-vba/Rubberduck] comintern commented on RetailCoder.VBE/UnitTesting/PermissiveObjectComparer.cs of pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.VBEEditor/Application/ExcelApp.cs of pull request #2858: Exception handling for unit tests and more

[rubberduck-vba/Rubberduck] retailcoder closed issue #2847: MoveFieldCloserToUsageInspection should ignore RD's Assert class variable in test modules.

[rubberduck-vba/Rubberduck] retailcoder merged pull request #2850: Add type checking for Asserts, implement {Not}SequenceEquals from comintern/Rubberduck/next into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] retailcoder pushed 11 commits to next (only showing some of them below)

[rubberduck-vba/Rubberduck] comintern pushed commit 044c2c45 to next: Add reference\value type checking for Asserts, implement Sequence Asserts.

[rubberduck-vba/Rubberduck] retailcoder closed issue #908: AreEqual to support Arrays

[retailcoder/Rubberduck] retailcoder pushed 12 commits to rd-next (only showing some of them below)

[rubberduck-vba/Rubberduck] build for commit e67c044d on next: AppVeyor build succeeded

Can you, by default, with VBA test if firstCell.Parent is secondCell.Parent?

why not?

Debug.Print firstCell.Parent = secondCell.Parent

Debug.Print ObjPtr(firstCell.Parent) = ObjPtr(secondCell.Parent)

Doesn't the parent property default down to .Name?

yeah. use ObjPtr to compare the actual object pointer

I've never had to do this kind of thinking let alone testing.

Debug.print firstCell.Parent Is secondcell.Parent

it's the difference between "ugh. vba." and "huh, it's vba!"

@ThunderFrame that's actually more idiomatic I think (and more elegant too)

I always forget about Is and resort to undocumented pointer-level checks lol

Why look in the manual when you can have much more fun under the hood?

[rubberduck-vba/Rubberduck] comintern added label discussion to issue #2860: Unit tests should be aware of Option Compare settings.

=)

[rubberduck-vba/Rubberduck] comintern added label feature-unit-testing to issue #2860: Unit tests should be aware of Option Compare settings.

[rubberduck-vba/Rubberduck] comintern opened issue #2860: Unit tests should be aware of Option Compare settings.

@Mat'sMug That can be dangerous. ObjPtr returns the interface pointer, so if you have the same object with variables pointing to different interfaces, the comparison fails.

[rubberduck-vba/Rubberduck] retailcoder commented on issue #2860: Unit tests should be aware of Option Compare settings.

> Determining the setting for the module shouldn't be too hard actually. I think the most elegant solution would be to honor the Option Compare statement - the big question mark is what do we do with Option Compare Database?

@Comintern TIL

Is compares the reference pointers.

I need to remember that

[rubberduck-vba/Rubberduck] comintern commented on issue #2860: Unit tests should be aware of Option Compare settings.

> Hmmm... Option Compare Database is only supported in Access. If we're honoring the setting, we'd need to ask the host for the DBE's internationalization settings.

[rubberduck-vba/Rubberduck] retailcoder added label up-for-grabs to issue #2861: Kill DeclarationType.ModuleOption

[rubberduck-vba/Rubberduck] retailcoder added label difficulty-01-duckling to issue #2861: Kill DeclarationType.ModuleOption

[rubberduck-vba/Rubberduck] retailcoder added label technical-debt to issue #2861: Kill DeclarationType.ModuleOption

[rubberduck-vba/Rubberduck] retailcoder added label enhancement to issue #2861: Kill DeclarationType.ModuleOption

[rubberduck-vba/Rubberduck] retailcoder opened issue #2861: Kill DeclarationType.ModuleOption

[rubberduck-vba/Rubberduck] ThunderFrame commented on issue #2860: Unit tests should be aware of Option Compare settings.

> We're going to need the DAO Interop anyway, but once we have that, the collation is available with CurrentDB.CollatingOrder

[rubberduck-vba/Rubberduck] ThunderFrame commented on issue #2860: Unit tests should be aware of Option Compare settings.

@Duga I never removed the Emitter hack - we could emit that function and retrieve its result

Just make a new member to get that function body, and call Emitter.ExecuteWithResult<int>(Emitter.GetOptionCompareSetting())

Could be a last-resort option

I know it works. Not sure it helps make RD stable though

@Mat'sMug - I forgot to reply to your comment about being able to trigger a compile, but not being able to determine if it did compile....

I'm pretty sure you can check the state of the CommandBarButton to see if it was successful or is currently compiled

OK. Crash time.

public long MocksageBox(string prompt, long buttons, string title, string helpfile, long context)
{
    Debug.WriteLine("MsgBox called from VBA.");
    return 0;
}

MocksageBox lol!

DuckedMsgBox

MsgBoxDuck

oh there's something

mock duck

QuackBox has a ring, too

Heading home to test and see if I can reproduce the error I was having earlier.

Thanks again for answering all my questions.

make sure you sync! :)

gah - I'm writing classes (and enumerable collections) for entities I am delivered in Excel files.... Open the file, read the range into an array, create a class instance for each row, and add the class to my custom collection. Works well. But the project owner keeps giving me new files (that reflect new type of objects). It's almost like I need a class-builder given a table with headers.

Mmmm... Mock duck. It's really tasty.

LOL!

Rubber MockDuckFace

Fried up in a coconut curry, Massaman style with some chopped peanuts...

@Comintern That looks truly disgusting.

^

It's kind of like tofu, but with more oil and a chewier.

pictures fatcat in a 44-gallon hat saying "Let's can this Sh!t and see if somebody buys it"

which was also the M&A strategy of Yahoo

lol

Hmmm... dll not loaded by current process...

^

time for AccessViolationException-party?

I'm just moving up another level and hooking vbe7.dll directly instead. The runtime apparently doesn't load unless code is running.

Oh, crap. Wrong signature.

huh, this is going to work!?

It should. I just need to figure out the calling convention for rtcMsgBox.

It apparently isn't delegate int MessageBoxDelegate(string prompt, int buttons, string title, string helpfile, int context);

[rubberduck-vba/Rubberduck] retailcoder commented on issue #2860: Unit tests should be aware of Option Compare settings.

[rubberduck-vba/Rubberduck] comintern commented on issue #2860: Unit tests should be aware of Option Compare settings.

> Something wrong with just calling Application.CurrentDb.CollatingOrder?

@Ducka, I meant for helper usage in the VBA test. Wouldn't RD already know the compare method having parsed the file?

Ah - inside the test.

Access defaults to Binary, doesn't it?

[rubberduck-vba/Rubberduck] retailcoder commented on issue #2860: Unit tests should be aware of Option Compare settings.

You think rtcMsgBox takes the arguments as string pointers?

[rubberduck-vba/Rubberduck] retailcoder commented on issue #2860: Unit tests should be aware of Option Compare settings.

> Oh.

@Duga @ThunderFrame see edit :)

@Comintern straight-up strings won't do?

(I've no idea)

Either that or I'm missing a step in initializing the hook.

4th hit for "rtcMsgBox" is "CRACKiNG TUTORiAL"

Could hook the kernel call to MessageBoxIndirectA instead...

@Comintern Modules default to binary, although Access defaults to including Option Compare Database at the top of a module, so I guess you could say Access Modules want to be Binary, but the VBE gets in the way

the Immediate Window has its own setting

All your MsgBox are belong to us!

ok what's going on here?

what's the reach of that?

OK, so I just sink a hook in vbe7.dll and intercept all calls to _rtcMsgBox.

how hard would it be to INTERCEPT ALL THE THINGS!! ? (profiler?)

I returned a vbMsgBoxResultWTF and didn't pass the call to the runtime.

Pretty much any function that the vbe7.dll supports would be easy.

The host might be a little more dangerous - I'd guess that most of the functions aren't in the export table.

Could probably figure out UserForm loads.

@Comintern so... we have an actual mock engine?!!?

wait no

Yup. msvbvm60 might be a bit more work - we can't set the hook until the code is executing.

well

that's sick

We'd need to explicitly support functionality for the most part.

This is literally all the code that happened there:

        public void OnStartupComplete(ref Array custom)
        {
            //CreateTestMenu();

            var address = EasyHook.LocalHook.GetProcAddress("vbe7.dll", "rtcMsgBox");

            _hook = EasyHook.LocalHook.Create(address, new MessageBoxDelegate(MocksageBox), null);
            _hook.ThreadACL.SetInclusiveACL(new int[] { 0 });
        }

        [UnmanagedFunctionPointer(CallingConvention.StdCall, SetLastError = true)]
        delegate int MessageBoxDelegate(IntPtr prompt, int buttons, IntPtr title, IntPtr helpfile, int context);

    RDTest.Mocking.MsgBox.Returns vbCancel
    sut.Run

    RDTest.Mocking.MsgBox.Verify It.IsAny("String"), vbOkCancel).Once

Yep.

InputBox too.

Environ

Dir

Open

Close

?

Beep <--LOL

rtcFileAttributes	0x101bee23	0x001bee23	555 (0x22b)	VBE7.DLL	C:\Program Files (x86)\Common Files\microsoft shared\VBA\VBA7.1\VBE7.DLL	Exported Function
rtcFileCopy	0x100db204	0x000db204	576 (0x240)	VBE7.DLL	C:\Program Files (x86)\Common Files\microsoft shared\VBA\VBA7.1\VBE7.DLL	Exported Function
rtcFileDateTime	0x100db379	0x000db379	577 (0x241)	VBE7.DLL	C:\Program Files (x86)\Common Files\microsoft shared\VBA\VBA7.1\VBE7.DLL	Exported Function
rtcFileLen	0x100db453	0x000db453	578 (0x242)	VBE7.DLL	C:\Program Files (x86)\Common Files\microsoft shared\VBA\VBA7.1\VBE7.DLL	Exported Function

[rubberduck-vba/Rubberduck] bclothier commented on issue #1184: Early to Late Binding Refactoring

Date and Time.

Timer

Input

Wow.

@Duga not crazy at all

Rnd and Randomize

SendKeys

Shell

oh wow

mmm, Shell

^^

DoEvents

^ Shell - My favorite breakout of a Citrix session

[rubberduck-vba/Rubberduck] bclothier commented on issue #1184: Early to Late Binding Refactoring

Huh. _ErrObject::Raise. How does Assert.ErrorRaised(6, "Overflow not raised") sound?

On Error Resume Rubberduck

OK, I need to make a list.

Just tested github.com/rubberduck-vba/Rubberduck/issues/2854 and I get identical results.

Oh hai, rtcCreateObject...

^^That's how you mock host objects.

holy crap

@Comintern you just cracked open Pandora's box!

:-D

so how can we mock a Range?

Application.Calculation

uh, doesn't that ultimately amount to... re-implementing VBA?

Uh yeah, more or less.

These would be more like Fakes instead of Mocks though.

right

Heh. DllVbeTerm.

huh

wait that's The Terminator?

Oh hai, DllGetClassObject. What do you do.

OK, time to try newing me up some VBA object...

Oh wait. DllGetClassObject should only return the VBE instance. That's part of the OLE harness.

something tells me RD might eventually be running its tests off VBA class modules...

huh?

we have that already

:(

The one I really need is VBEGlobal::Load.

Let's see if it will find the entry point...

I wonder if some of these addresses match up with the pointers in the treeview lParams.

I can restrict the hook to specific threads.

Holy sh!t -> easyhook.github.io/api/html/T_EasyHook_COMClassInfo.htm

Apparently EasyHook supports hooking in COM vtables.

wah

[rubberduck-vba/Rubberduck] IvenBach commented on issue #2854: Rubberduck>Unit Test>Test Method

> Just tested it at home and results are identical. I can reproduce it every time.

[rubberduck-vba/Rubberduck] retailcoder commented on issue #1184: Early to Late Binding Refactoring

[rubberduck-vba/Rubberduck] retailcoder commented on issue #1184: Early to Late Binding Refactoring

[rubberduck-vba/Rubberduck] retailcoder commented on issue #2854: Rubberduck>Unit Test>Test Method

> Is there another project in the VBE with a test module named the same as the one you're in?

    Private Sub cmdButtonClick() '@Handles cmdButtonCancel.Click, cmdButtonCancel.Click
        Me.Hide
    End Sub

'Rubberduck generated code:
    Private Sub cmdButtonCancel_Click()
        cmdButtonClick
    End Sub

    Private Sub cmdOkCancel_Click()
        cmdButtonClick
    End Sub

#idea

with annotation intellisense, of course

when the methods exist, the annotation is dimmed; when they don't, it's squiggly-underlined and you bring up a little in-place drop-down to add the missing methods

That would come in handy. When does the code building happen?

ON-THE-FLY!

^

Woot!

You know that user class base object that declares the Initialize and Terminate events?

ClassClass?

classy

^ I'm trying to hook that interface.

you found it?

how about Activate, Deactivate, QueryClose?

Apparently the method that EasyHook uses to find the vtable entries requires newing up an instance via the native APIs.

but that information alone doesn't tell us the type being created, does it?

No, but we get the object instance I think.

huh

jaw drops

The thing that sucks about requiring it to new something up is that only public creatable GUIDs will yield a vtable.

[rubberduck-vba/Rubberduck] IvenBach commented on issue #2854: Rubberduck>Unit Test>Test Method

> No. I tested it on a new workbook with only TestModule1 and the code from my original post. I get the same results.

I'm gonna break this thing.

@Duga ok I've no idea wtf is going on there

I thought I'd replicated it earlier, but I realized that I'd copied a test, and hit the add method and pasted it manually into the middle of something with ctrl-v. Ooops.

Is there any code you want me to look it in VS?

@IvenBach I'm stumped

^

I'm going to open a feature issue for RD fakes.

[rubberduck-vba/Rubberduck] retailcoder added label feature-request to issue #2862: @Handles annotation for auto-implementing event handlers

[rubberduck-vba/Rubberduck] retailcoder opened issue #2862: @Handles annotation for auto-implementing event handlers

[rubberduck-vba/Rubberduck] retailcoder added label feature-annotations to issue #2862: @Handles annotation for auto-implementing event handlers

[rubberduck-vba/Rubberduck] retailcoder added label feature-inspections to issue #2862: @Handles annotation for auto-implementing event handlers

Ok.

Early bed time. Sleep! That's where I'm a viking!

'night!

[rubberduck-vba/Rubberduck] comintern added label feature-request to issue #2863: RD Unit Testing needs a Fakes architecture

[rubberduck-vba/Rubberduck] comintern added label discussion to issue #2863: RD Unit Testing needs a Fakes architecture

[rubberduck-vba/Rubberduck] comintern added label feature-unit-testing to issue #2863: RD Unit Testing needs a Fakes architecture

[rubberduck-vba/Rubberduck] comintern opened issue #2863: RD Unit Testing needs a Fakes architecture

[rubberduck-vba/Rubberduck] retailcoder commented on issue #2863: RD Unit Testing needs a Fakes architecture

> Linking #1550

[rubberduck-vba/Rubberduck] retailcoder closed issue #1550: How could we approach mocking of dependencies?

[rubberduck-vba/Rubberduck] bclothier commented on issue #1184: Early to Late Binding Refactoring

[rubberduck-vba/Rubberduck] retailcoder commented on issue #1184: Early to Late Binding Refactoring

[rubberduck-vba/Rubberduck] retailcoder commented on issue #2863: RD Unit Testing needs a Fakes architecture

> I like the Private Fakes As New Rubberduck.FakesProvider approach - it's in line with the spirit of the current API. Although... I like how Private Ducks As New Rubberduck.DuckProvider is similar to Private Mocks As New Rubberduck.MockProvider 🙄

[rubberduck-vba/Rubberduck] comintern commented on issue #1184: Early to Late Binding Refactoring

> Picking up change events from the Properties dialog is trivial with the WinEvents APIs - the only issue is that it only receives update events when it's visible. It's also really easy to subclass now (too easy as it turned out - I think at one point it was accidentally subclassed) so tracking its state is also trivial. Just thought I'd throw that out there...

@Duga so there's another solution then!

@Duga @Mat'sMug - wouldn't fake ducks be something like a Geese?

ooooh lol

Or UglyDuck - it's really a Swan.

Duck, duck, mock duck runs

ducks

lol

people simply won't believe 3.0

> wait, you guys coded an IDE?

nope. that is the VBE.

...under there somewhere...

a mock framework. that hijacks the runtime and intercepts everything. that's.... wah

It's OK, you can still end a Property with End Function - we didn't move your cheese

LOL

[rubberduck-vba/Rubberduck] retailcoder synchronized pull request #2858: Exception handling for unit tests and more

[rubberduck-vba/Rubberduck] build for commit 9c8ba72a on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] retailcoder merged pull request #2858: Exception handling for unit tests and more from MDoerner/Rubberduck/stuff into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] MDoerner pushed commit c4ee5cda to next: Additional safety for long procedure names in Excel.

[rubberduck-vba/Rubberduck] bclothier commented on issue #1184: Early to Late Binding Refactoring

> The flicker is necessary in the case where you have an existing VBA project with project-level compilation constants already blank. Perhaps when RD is first started on parsing a new VBA project, and it detects it has no project-level annotations, it should flicker to inspect the current setting. That would be one-time only and subsequent parsing can just refer to the project-level annotations. That would reduce the need to define the #Const constant (which I really dislike using - it's more

[retailcoder/Rubberduck] MDoerner pushed commit c4ee5cda to rd-next: Additional safety for long procedure names in Excel.

[retailcoder/Rubberduck] MDoerner pushed commit 88545576 to rd-next: Wrapped test Execution in try catch to catch COMExceptions and log them.

[rubberduck-vba/Rubberduck] bclothier commented on issue #1184: Early to Late Binding Refactoring

> The flicker is necessary in the case where you have an existing VBA project with project-level compilation constants already blank. Perhaps when RD is first started on parsing a new VBA project, and it detects it has no project-level annotations, it should flicker to inspect the current setting. That would be one-time only and subsequent parsing can just refer to the project-level annotations. That would reduce the need to define the #Const constant (which I really dislike using - it's more

[rubberduck-vba/Rubberduck] build for commit 4c5de3d4 on next: AppVeyor build succeeded

preference for x = -y, or x = -1 * y

x = y * -1

me too

inspection for readability?

  Debug.Print - _
  x

FFS

NUKE ALL THE LINE CONTINUATIONS!!!

  Debug.Print -- _
  -- _
  x

why do tests take half a century to run in VS2015?

[retailcoder/Rubberduck] retailcoder pushed commit ccac5f32 to rd-next: implemented rewriter methods. need an easier way to insert in the declarations section and appending members. Some failing tests fail because of extra whitespace/lines left blank,

[retailcoder/Rubberduck] retailcoder pushed commit e748ede0 to rd-next: Merge branch 'rd-next' of github.com/retailcoder/Rubberduck.git

TTGTB

</mug>

rewriter is starting to get tests passing again

[rubberduck-vba/Rubberduck] retailcoder commented on issue #2863: RD Unit Testing needs a Fakes architecture

> Each member of FakeProvider could be a factory method that returns an object that's responsible for setting up & tearing down its own hooks... could we use the destructor for this or that would be too late?

[rubberduck-vba/Rubberduck] SystemsModelling opened issue #2864: Indenting in 2.0.13.32288