older PC systems (even those sold as 64-bit) were often 32-bit word systems

16-bit words are a historical inconvenience that doesn't reflect the state of hardware per se

RELOAD!

[Cardshifter/Cardshifter] 1 opened issue.

not that programming languages should bother with the actual word-width of the HW they run on ...

[Cardshifter/HTML-Client] 1 opened issue. 4 closed issues. 6 issue comments.

[rubberduck-vba/Rubberduck] 67 commits. 4 opened issues. 2 closed issues. 33 issue comments. 167880 additions. 109145 deletions.

[Zomis/Server2] 2 opened issues.

@robodude666 the abstract factory pattern is perhaps the single most useful tool to use with DI/IoC =)

So no, not a smell

A factory factory might be though

The use case is whenever you need an instance of a class that has dependencies that aren't known at resolution time, i.e. at startup.

[rubberduck-vba/Rubberduck] retailcoder merged pull request #3876: Tidbits for this and that from Vogel612/Rubberduck/refactor-carefully into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] retailcoder closed issue #3751: Refactor.Rename undoes previous edits

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

[rubberduck-vba/Rubberduck] Vogel612 pushed commit 6fbffb14 to next: Simplify targetClass retrieval in ImplementInterfaceCommand

[rubberduck-vba/Rubberduck] Vogel612 pushed commit 4022c062 to next: Stop prematurely disposing of Project in AddTestModuleCommand, see #3877

[rubberduck-vba/Rubberduck] Vogel612 pushed commit 8f12a16f to next

[rubberduck-vba/Rubberduck] Vogel612 pushed commit 059acce4 to next

[rubberduck-vba/Rubberduck] Vogel612 pushed commit 05fbf33f to next

[rubberduck-vba/Rubberduck] Vogel612 pushed commit 2cf851b6 to next: Set up MessageBox Mock without verifying it to fix MoveCloser test fluking

[rubberduck-vba/Rubberduck] web-flow pushed commit 87d7fd73 to next: Merge branch 'next' into refactor-carefully

[rubberduck-vba/Rubberduck] web-flow pushed commit db9d3793 to next: redo message-box setup

[rubberduck-vba/Rubberduck] web-flow pushed commit e8a405d1 to next: Update Font for filtering label

[rubberduck-vba/Rubberduck] web-flow pushed commit 53fc9dd0 to next

@Vogel612 @IvenBach just for clarification - the word word has double meaning - Vogel referred to the hardware concept but we also have word as a data type in Windows data type, which is really unfortunate because the word in Windows was conceived back in 16-bit days...but then they had to keep backward compatiblity when we switched to 32-bit so instead of redefining word to mean 32-bit, they kept it at 16-bit and invented dword (aka "double word") to represent 32-bit.

Therefore, context matters - if you're reading documentation on a Win32 API, word is fixed. If you're reading documentation on some kind of CPU instructions (?), word may be based on the numbers of registers that the CPU has.

[rubberduck-vba/Rubberduck] retailcoder published prerelease Rubberduck-v2.1.2.3046

[rubberduck-vba/Rubberduck] retailcoder created tag Rubberduck-v2.1.2.3046

[rubberduck-vba/Rubberduck] build for commit 53fc9dd0 on next: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit 53fc9dd0 on next: 57.45% (target 0%)

@MathieuGuindon / @Vogel612 friendly reminder --- given the release, you might wanna to get i18n done for the installer....

Yup

You know, with that nifty resx manager add-in, I'd be much less reluctant to split up RubberduckUI.resx now.

Before, it was easy to miss one, so I'd sooner have as much as possible under 1 file

Now... I'd be willing to split it into namespaces

[rubberduck-vba/Rubberduck] bclothier created pull request #3887: Remove user settings deletion to merge bclothier/Rubberduck/BuildRegistryWithWix into rubberduck-vba/Rubberduck/next

@MathieuGuindon I've yet to even open VS. Too much work.

[rubberduck-vba/Rubberduck] build for commit f652a2fe on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit f652a2fe on unknown branch: 57.45% (target 0%)

[rubberduck-vba/Rubberduck] bclothier commented on issue #3862: Modularization strikes again - Let's move some stuff around

[rubberduck-vba/Rubberduck] retailcoder commented on issue #3862: Modularization strikes again - Let's move some stuff around

@MathieuGuindon Not till next week, after 4/13, can I think about any RD contributions.

@IvenBach no pressure! =)

who'd thunk expanding an assembly in debug mode => corrupted state

welp i feel dumb. i didn't need the assembly after all. Tree, meet forest.

[rubberduck-vba/Rubberduck] bclothier edited wiki page Installing

@robodude666 Thanks.

@this Referring to the link @robodude666 sent me.

> For arrays the supporting COM type is the safe array, which comes with a large group of utility functions.

Are these the things that give a notice when you hit a Subscript out of range RTE 9?

@bruglesco Welcome to the pond.

yes, pretty much.

in C++, going outside the array was undefined

maybe you got gibberish. maybe you get a nice GPF. maybe you send an email to president and tell him he's a OK dude after all.

but in VBA-landia, you don't have that problem; it keeps you all safe and secure.

Hmm that gets me thinking --- is there such thing as an undefined behavior in VBA?

an obvious one would be API calls but that's technically going outside the VBA...

@this GPF?

The program has stopped working and must be shut down. We apologize for convenience. We are searching for a solution....

Reading SafeArrayGetLBound function has the the signature of ...

HRESULT SafeArrayGetLBound(
  _In_  SAFEARRAY *psa,
  _In_  UINT      nDim,
  _Out_ LONG      *plLbound
);

@IvenBach General Protection Fault - an access violation, reading memory you shouldn't be reading.

In c++ is there the concept of a function returning a value?

it does.

I see the _Out_ being 'returned'

HRESULT is the return

that's basically like ByRef

here's a reason why....

HRESULT is handle?

in C++ there is no On Error GoTo <label> equivalent*

in a large majority of C++ code, it's very common to check the result because otherwise, you wouldn't know you have an error and need to stop doing other things.

It's as if you have On Error Resume Next perpetually on

Hrm... So you do some function but before that function returns the value you have to check the value to make sure it's valid?

Raymond's article is a good example: blogs.msdn.microsoft.com/oldnewthing/20050114-00/?p=36693

yes, that's the purpose of the HRESULT

so a large majority, if not all COM functions always return some kind of HRESULT

HRNGH! I read that article earlier and didn't understand it.

but to pass the interesting bits (e.g. the data you really want), you usually read it via one of hte ref parameter, hence the out parameter).

I've the feeling that I'm going to need more time to fully digestify this stuff.

'night duckies

Night mug.

night

Some new contributor will one day ask why he's 'Mug'.

@this How much knowledge is needed to understand the c++ stuff? I'm understanding small parts of what I read but most of it's flying way above me.

SAFEARRAY structure states about thread safety

> ... For example, consider an application that uses the SafeArrayLock and SafeArrayUnlock functions. If these functions are called concurrently from different threads on the same SAFEARRAY data type instance, an inconsistent lock count may be created. This will eventually cause the SafeArrayUnlock function to return E_UNEXPECTED. You can prevent this by providing your own synchronization code.

Is this the undefined you mentioned earlier?

not quite

becaus eint his case, you get an error code, E_UNEXPECTED

so you know something's not right and can take action

@IvenBach TBH i really never did that much C++. I could make few ultra simple C++ but that's it. More important that you at least understand the pointer works and what it means for it to be "undefined"

More than a little frustrating reading and not understanding it.

fwiw, i didn't get raymond for a while

but kept coming back and reading.

then at one point, click.

bytecomb.com/… has at the bottom the explanation.

Is the part with *VarPtrArray being shown big endian.

ok what about it?

Is that an example of big endian or little?

> Here’s how it all maps out in detail. As always, the byte order and pointer size depends on the architecture; in my case it’s 32-bit office on a little-endian Intel processor:

ok.

u know what endian is all about?

I'll have to remember that big endian reads like it's shown.

Not really. I've read about it and identify, now, which one is which.

yeah we use big endian in our numeric system... 1, 10, 100... the number on the left is "bigger" than to right

a little endian system would be as if we had a number system where we where 1, 01, 001

e.g. the little number goes left

It's just swapped.

but that's an oversimplification because the endian is dependent on the word

for example, cDim is 01 00

The grouping of 80 DB 46 00 is what messes with me.

which means only one

I'm not well practiced with base16

right, it's the word, not the bit that goes together

I kinda sorta see what it's doing now with those table, and why. I'll have to return to this till I git it.

[rubberduck-vba/Rubberduck] bclothier review_requested pull request #3888: Fix unit testing

[rubberduck-vba/Rubberduck] bclothier created pull request #3888: Fix unit testing to merge bclothier/Rubberduck/FixUnitTesting into rubberduck-vba/Rubberduck/next

> Closes #3809 The unit tests now will run after a parse, rather than during which helps avoid errors due to running tests and parsing concurrently. Updated the test engine to listen to the event. I had to add a flag to indicate whether to run the tests. I also fixed the issue with assembly because Assembly.GetEntryAssembly was returning null, and the code made wrong assumption about the name of the library. We avoid this by doing a typelib registry lookup instead. That neatly solves the...

problem without needing to discover the Main at all and more importantly, give us an opportunity to warn of a problem with early binding --- if we can't locate the typelib, then early binding is not going to work at all.

[rubberduck-vba/Rubberduck] build for commit 0305aae4 on unknown branch: AppVeyor build failed

BUILD FAILURE!

Home time, me so tired.

[rubberduck-vba/Rubberduck] PeterMTaylor commented on issue #3885: Branding Rubberduck

> Please indicate as per chat earlier if using the issues generic term “branding” would be open for review on Stack Exchange UI/UX or not thanks.

@MathieuGuindon DefVar easter-egg: "Explicit Language Warning: This code may contain explicit language directives that are too explicit.

by the same token, is there any point in the CVar function?

[rubberduck-vba/Rubberduck] MDoerner commented on issue #3862: Modularization strikes again - Let's move some stuff around

@Duga Welp, that's interesting. I reproduced this once locally....

[rubberduck-vba/Rubberduck] bclothier synchronized pull request #3888: Fix unit testing

[rubberduck-vba/Rubberduck] build for commit 585cd6a3 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) commented on pull request #3888: Fix unit testing

[rubberduck-vba/Rubberduck] build for commit 585cd6a3 on unknown branch: 57.41% (target 0%)

@Duga We have a Heisenbug.

ugh...

it's something that only happens on 2nd or 3rd run.

but it does happen locally.

I also don't htink it'll happen if you run it isolated

[rubberduck-vba/Rubberduck] Vogel612 commented on Rubberduck.Core/UnitTesting/TestEngine.cs of pull request #3888: Fix unit testing

In my case, i have to Run All, and only then I can get teh bug.

yeap. similar experience here..

I thought setting up the MessageBox mock fixed the problem, but apparently it didn't ...

given that it only occurs when we Run All, i'm thinking we're leaking state somehow.

since the tests run in parallel, right?

well, no. I just got a repro with it being the only test.

yaaay...

have to put in breakpoint since it won't let me unwind. :\

only took me 5 tries...

ok possibly a dumb question but --- there's no selection logic

as a result, the code pane's selection is null

how does the code know what to select if we don't specify a selection?

@MathieuGuindon @IvenBach Back in the Win 9x days, it used to be an instant blue-screen-of-death. Since Win XP, it just means instant death to the process that caused it.

@this first user-defined declaration

right.

so. - i'm in InsertNewDeclartion

on this line. var newVariable = $"Dim {_target.IdentifierName} As {_target.AsTypeName}{Environment.NewLine}";

_target.IdentifierName == CreateObject

I think that's the bug. it all depends on gets into the AllUserDelcarations first -- the Foo or CreateObject

ohhh

confirmed.

CreateObject is a Variable

ugh...

@this oh shiiiiiiii.... That one should probably be in the grammar itself, no?

I bet GetObject has the same problem

I have no idea, honestly. I'm just telling what I see. I don't know why CreateObject is given a declaration type of Variable - is that appropriate?

@MathieuGuindon pretty sure we do not want to special case GetObject and CreateObject

at least not in the grammar... maybe in the resolver,...

Then we'll have to fake them

so this whole thing is a unit test artifact?

in the production code CreateObject is something else?

possibly,yes

no. It's just that the unit-tests may not be correctly faking CreateObject

@Vogel612 but is it in the custom declarations loaders somewhere?

searching

@MathieuGuindon .... that's either clever or scary. IDK.

according to Excel's code, CreateObject is a (function:Variant)

wut.

shouldn't it return an Object?

so that means it's OK in productin? that unit test is wrong in saying it's a variable?

[rubberduck-vba/Rubberduck] retailcoder closed issue #3884: Settings are removed when uninstalling.

[rubberduck-vba/Rubberduck] retailcoder merged pull request #3887: Remove user settings deletion from bclothier/Rubberduck/BuildRegistryWithWix into rubberduck-vba/Rubberduck/next

nope, it does return Variant

[rubberduck-vba/Rubberduck] bclothier pushed commit f652a2fe to next: Remove user settings deletion

[rubberduck-vba/Rubberduck] web-flow pushed commit 4e9c732f to next

Remember how they love dem variants....

@Duga @FreeMan now you can rejoice.

In any case it being resolved as a variable means you're prolly getting a result for undeclared variable

if we were running all inspections in UTs, sure. But we aren't, right?

nope, we're not

We aren't, and there is no CreateObject custom declaration. Can you find it in the OB?

yes I can VBA.Interaction module

Function CreateObject(Class As String, [ServerName As String])
    Member of VBA.Interaction

no As Object so Variant is correct.

Ok.. so it's just a UT setup issue then

For right now, I'm going to fix the test by explicitly selecting the Foo

instead of whatever comes out of the AllUserDeclarations first.

which is.... non-deterministric. ;)

You can't have CreateObject in UT code without loading the stdlib xml

hm.

Ditto with MsgBox or anything else defined in a type library

resolver otherwise treats it as an undeclared variable

which is what you're seeing there

setting it up w/ standard librar....

Or, don't use the function in the test vba code ;-)

i think that was a repro from a previous code

so not touching that

K

now the CreateObject's no longer in the list

+1

Good bye, heisenbug. It was not nice knowing you.

Yay!

[rubberduck-vba/Rubberduck] retailcoder published prerelease Rubberduck-v2.1.2.3050

[rubberduck-vba/Rubberduck] retailcoder created tag Rubberduck-v2.1.2.3050

[rubberduck-vba/Rubberduck] build for commit 4e9c732f on next: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit 4e9c732f on next: 57.45% (target 0%)

[rubberduck-vba/RubberduckWeb] tdalon opened issue #52: Dead link SourceControl page

@this Rejoice I shall!! thanks!

I notice that the new installer forces me to close Excel & Access, yet ignores Outlook. also status-by-design?

[rubberduck-vba/RubberduckWeb] bclothier commented on issue #52: Dead link SourceControl page

> @tdalon we need to update the web page. The links are dead because we have discontinued the source integration. See this closed issue in our Rubberduck repository. To use Rubberduck with your source control, you should use the Export functionality which is still available for easily importing / exporting files en masse and you would use the source control externally.

What's changed that they now have to be closed?

I didn't update that method.

I think there is a method in the code that check if there's a process running

It also ignores an open instance of Word

yeah. not very smart. ;)

[rubberduck-vba/RubberduckWeb] retailcoder commented on issue #52: Dead link SourceControl page

> The website will be updated over the weekend with the release of v2.2 (needs the 2.2 build for the inspections page).

wow, it's like Win3.1!!

and it's not going to be perfect if we look for those process name

@Duga thanks, deleted half of my reply =)

Abort, Retry, Fail?

need to do some thing to look if hte DLL is in use so it'll work for any host

it was quite happy with a Retry, btw...

it's written in Pascal.

impressive, it even re-opened the two workbooks I had open!

@this eh????

yeah, that's why it's... quirky. ;)

@FreeMan if you have Outlook open and never loaded the Outlook VBE, it wouldn't interfere with the installer AFAICT

not so impressive ^

It killed the processes, it didn't shut them down gracefully... :(

oooo. fun.

so... don't install while it's in use. Otherwise our installer will go into a murderous rage.

:p

@MathieuGuindon seems reasonable. I don't remember having the vbe open in this particular session of Excel. But, I do tend to launch it and have it open for days...

@FreeMan makes sense, it only knows about processes, not their running apps, even less their shutdown api

@this Yeah, might be worth a warning in the installer.

i may look at the code there. AS I said, I really didn't update that section of code.

Caution: Save your work and shut it down yourself, else ahead be a murderous rage.

bonus: make the dialog drip with blood

Yes!

add "Here's johnny!" soundwave

@this I'm of the age where my instant thought was "Carson". I'm sure there's another reference in there that refers to "Chucky" or something similar

No. Jack Nicholson in The Shining

(the one by Stanley Kubrick)

I do think I watched that (or parts of it) once. /Not a horror movie fan

I think this clip but not sure: youtube.com/watch?v=WDpipB4yehk

thanks for all the installer updated, @this, @MathieuGuindon, and whoever else contributed! Time to get back to what they pay me for... :(

[rubberduck-vba/Rubberduck] bclothier synchronized pull request #3888: Fix unit testing

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) commented on pull request #3888: Fix unit testing

[rubberduck-vba/Rubberduck] bclothier commented on Rubberduck.Core/UnitTesting/TestEngine.cs of pull request #3888: Fix unit testing

[rubberduck-vba/Rubberduck] build for commit 44f78870 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit 44f78870 on unknown branch: 57.41% (target 0%)

[rubberduck-vba/RubberduckWeb] retailcoder commented on issue #52: Dead link SourceControl page

> Essentially the feature provided a dockable panel similar to Visual Studio's Team Explorer, but it was hard-wired to work with git, and had a number of bugs that made it rather unstable and annoying. Since providing source control tooling is only marginally in-scope for Rubberduck, we chose to remove the buggy feature instead of devoting our limited time & resources to it. The feature might re-emerge at one point in the future, but for now it was interfering with the core stuff, so we're pul

[rubberduck-vba/Rubberduck] eteichm created pull request #3889: Localization fr to merge eteichm/Rubberduck/localization-fr into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] build for commit cf62d764 on unknown branch: AppVeyor was unable to build non-mergeable pull request

BUILD FAILURE!

I've managed to screw up zoom by somehow touching the touchpad on my Dell whilst doing pressing maybe Crtl. I now have stupidly enlarged stackoverflow page. I tried pulling fingers together from opposite corners of a diagonal on touchpad which has shrunk page but it is not correct. Can anyone remember what the keyboard shortcut is to restore normal zoom view pleasE?

My google-fu is failing on this!

Ctrl+, Ctrl-

[rubberduck-vba/Rubberduck] bclothier commented on pull request #3889: Localization fr

[rubberduck-vba/Rubberduck] bclothier commented on pull request #3889: Localization fr

[rubberduck-vba/Rubberduck] retailcoder closed issue #3809: Unit testing issues

[rubberduck-vba/Rubberduck] retailcoder merged pull request #3888: Fix unit testing from bclothier/Rubberduck/FixUnitTesting into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] bclothier pushed commit 6ef9abb6 to next

[rubberduck-vba/Rubberduck] bclothier pushed commit d0352f3b to next: Revert the changes made - we don't need the assembly at all....

[rubberduck-vba/Rubberduck] bclothier pushed commit 17273164 to next: Remove a unneeded using statement

[rubberduck-vba/Rubberduck] bclothier pushed commit 0305aae4 to next: Refit the EnsureReferenceToAddInLibrary to use registry lookup instead of playing Assembly circus

[rubberduck-vba/Rubberduck] bclothier pushed commit 585cd6a3 to next: Add abort logic

[rubberduck-vba/Rubberduck] bclothier pushed commit 00472ab8 to next: Fix the sily bug due to CreateObject being treated as a variable because we didn't include the standard library.... Heisenbug!

[rubberduck-vba/Rubberduck] bclothier pushed commit 44f78870 to next: Add a new IsError function to ParseProgressEventArgs to simplify the check for error state.

[rubberduck-vba/Rubberduck] web-flow pushed commit 0462e658 to next

A bit belated but I realized now that a better solution would have been to never use First() but rather Select()

That way it'll throw there, rather failing the assertion.

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UI/RubberduckUI.fr.resx of pull request #3889: Localization fr

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UI/RubberduckUI.fr.resx of pull request #3889: Localization fr

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UI/RubberduckUI.fr.resx of pull request #3889: Localization fr

#TIL: I think I miscategorized it as a heisenbug - it's more properly a mandelbug.

@MathieuGuindon Thanks. I tried that. It still leave a lot of white space either side. Either is is very large or when I shrink I get a whole of white space on the sides :-( I remember a super user telling me how to fix this when I screwed up before but I can't remember :-(((((((((((((((((((( wahhhhhhhh

isn't it typically ctrl-0?

@this *bug

@this TIL

That is what I thought @Vogel612

??

Sorry! That was for @this My bad

Ctrl + O is doing the windows open dialog

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UI/RubberduckUI.fr.resx of pull request #3889: Localization fr

[rubberduck-vba/Rubberduck] build for commit 0462e658 on next: AppVeyor build failed

BUILD FAILURE!

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UI/RubberduckUI.fr.resx of pull request #3889: Localization fr

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UI/RubberduckUI.fr.resx of pull request #3889: Localization fr

@QHarr no, the number zero, not the letter o

Ahhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh

@Duga WTF

I think for now I am gonna have to live with it wonky. Ctrl + 0 made it large again (though i am pretty sure that fixed it last time). I think I have overfiddled. <Sighs>

@QHarr CTRL-0 (zero)

Thank you all

CTRL +/- is larger/smaller, but not 100%.

@QHarr copy the URL to a new tab, should work too :) nevermind, doesn't.

I actually do Ctrl+Mousewheel to get to 100%

will do Ctrl+0 now

Oh, this beat me to it.

but you were clearer than me, however.

too easy to misread 0 as O

it's only obivous when you have them side by side but sheesh.

I wish the defautl font would just use the zero w/ a slash.

and not rely on us to eyeball the subtle difference in the width. ;)

I tried Ctrl + 0 (zero) and also pasting url new window.....

I have f***** it for now it seems and it is really annoying me!

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UI/RubberduckUI.fr.resx of pull request #3889: Localization fr

[rubberduck-vba/Rubberduck] retailcoder commented on pull request #3889: Localization fr

> Just sync the PR / resolve the conflicts and we're good to go!

Ahhh........Houston we have lift off.... it is a delicate balance to get it to just the right place so it doesn't look wrong. Thanks again.

@PeterMTaylor I'm afraid I don't know how to answer your question RE: branding & UI/UX review....

@WaynePhillipsEA I was wondering if you would be willing to do one more favor and run the latest version via that wrapper thingy of your and see if we're leaking anything. I know that we are leaking something when we unload/reload and that probagbly won't get fully fixed for 2.2 but would be good to know we aren't in other places.

TBH I think we should refrain from making changes to COM management stuff until Sunday

not suggesting that we do this; was thinking more about noting what is still known to not work, besides unloading/reloading

would be good to have that in the release notes, no?

yeah

@this probably won't get chance before sunday tbh. playing catchup at the moment with work

no worries!

it was just to update the documentation on this whole thing since there's been a lot of things going on. I plan to close various issues as stale with a note asking them to open a new issue if htey have any specific problem with crashing or whatnots.

last time I checked it was still clear. that was a couple of months back. but I'm not up on what changes have been made since.

sure, makes sense

g2g good luck with the v2.2 release guys :)

@WaynePhillipsEA later! and thanks!

[rubberduck-vba/Rubberduck] bclothier created pull request #3890: Fix move closer UT again to merge bclothier/Rubberduck/FixUnitTesting into rubberduck-vba/Rubberduck/next

> ... using Single instead of First to ensure it fails fast. Maybe that'll make AppVeyor happy?

[rubberduck-vba/Rubberduck] eteichm commented on RetailCoder.VBE/UI/RubberduckUI.fr.resx of pull request #3889: Localization fr

[rubberduck-vba/Rubberduck] eteichm commented on pull request #3889: Localization fr

[rubberduck-vba/Rubberduck] build for commit 3958893c on unknown branch: AppVeyor build failed

BUILD FAILURE!

@this Confirming my knowledge. If there is no As [Type] at the end it's implicitly a variant?

@Duga AppVeyor, you're such a bastard, you know?

@this Just found it.

@IvenBach Yes it is

AppVeyor, I ran this like 5 times and you have the gall to fail.

Trying to catch up on chat and pick up my daily breadcrumbs.

on the semi-bright side, the error is more explicit - we see that there's more than one matching element. Too bad it doesn't say what else matched.

Why is it that some functions are explicit in including As Variant while others omit it?

That's one thing that confused me for the longest time.

whose functions? Ours or VBE?

TBH, I never ever saw VBE's functions be explicit w/ Variants.

I’ll remember to post any that I find. Might be I’m half imagining it.

[rubberduck-vba/Rubberduck] retailcoder commented on pull request #3889: Localization fr

> Don't worry about the stale keys, but the conflicts need to be resolved so I can merge the PR.

@IvenBach same annoyance with ByRef parameters :)

I'm getting less ignorant with every passing day it seems.

@MathieuGuindon Or fully qualified Member calls Excel.Application.InputBox vs VBA.Interaction.InputBox

nah that's different

It's different but, yes. Bit it still was one of those things that confused me greatly.