VBA Rubberducking

@this did you fill up the survey they send about exactly that in, uh, January I think?

yeah and voted on the UV. That thread on UV is borderline spam.

MSFT taking their sweet time

"but what about that one single stupid legit use case we're going to break?!"

and it's doubly annoying when you've taken pain to pre-format the target area and it goes on to ignore it

^

7:18 PM

because we know you don't need or want it.

We know what's best for you!

I'd love to hear about a actual legit use case.....

nope, can't find one

you know, weed being legal in Seattle does explain a bunch of things...

2

not really because it's recent. Excel has that behavior since what?

inception?

...well to be legal, the idea had to have supporters in the first place

and yeah, as far back as I can remember. PITA since at least 1994

betcha that ol'Billy had a token or two.

(and everyone else at the place)

@this Social Security Numbers can have leading zeros! Oh, wait, I need to keep those, don't I?

7:31 PM

:barf: #TIL Excel addins need a unique name for the onLoad event to properly get wired up.

I should have thought of that beforehand

[rubberduck-vba/Rubberduck] Vogel612 commented on Rubberduck.Core/App.cs of pull request #4006: Splice Off Rubberduck.Refactorings from Rubberduck.Core

Of course, now, when you finally manage to force it to display the leading zeros, it very helpfully displays a little green triangle to warn you of the fact that you're breaking Uncle Bill's rule.

I mean, why would they use Excel to do word processing. It's for number crunching, philistines!!

text, Word, numbers, Excel. Is that hard to understand?

There are people that I work with that take meeting notes in Excel.

No, I don't know why.

[rubberduck-vba/Rubberduck] Vogel612 [commented on Rubberduck.Core/UI/Refactorings/RemoveParameters/RemoveParametersPresenter.cs]‌(github.com/rubberduck-vba/Rubberduck/pull/…) of pull request #4006: Splice Off Rubberduck.Refactorings from Rubberduck.Core

7:37 PM

@FreeMan #BecauseTheyCan

[rubberduck-vba/Rubberduck] Vogel612 commented on Rubberduck.Core/UI/Refactorings/ReorderParameters/ReorderParametersPresenter.cs‌ of pull request #4006: Splice Off Rubberduck.Refactorings from Rubberduck.Core

I guess...

@IvenBach by that reasoning, it's A-OK to take notes on command prompt

heck, write on the monitor w/ a sharpie works, too!

It's rumored that only the older professors at universities get to write on the overhead pulldown canvas. Only the brave ones do it though.

gee whiz... do they still use those....

7:40 PM

They did in 2015 when I was at college.

~grumble grumble

VS get your effing build process in order

@Vogel612 earlier Son had a problem

is it a problem w/ Rubberduck.Deployment ?

don't think so ...

just making sure

Just Rubberduck.Refactorings not properly dealing with the references it should get ...

actually let me look at the Framework level :/

yea ... that makes sense....

7:49 PM

@this especially now that copy+paste works much better than it used to!

sorry. would be so nice to have "I don't wanna per-project, just kneel down before solution already" setting...

nah. the major gripe is with the compiler not actually seeming to be smart enough to figure out the framework needs to be retargeted...

now it's the deployment project...

~sets execution policy

mansellan

@Duga what does it conflict with?

@Vogel612 do you really have to reset policy? I never had to do this more than once?

dunno. just tried for the heck of it

didn't help

7:56 PM

yeah I don't think it should

what's the error in the Output window?

> Rubberduck.Core/Navigation/CodeExplorer/CodeExplorerComponentViewModel.cs
Rubberduck.Resources/CodeExplorer/CodeExplorerUI.resx

@this none yet ...

I also cannot build next.

oh dear.

@M.Doerner what's the error?

Severity	Code	Description	Project	File	Line	Suppression State
Error		The command "C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe -command "N:\Repositories\Rubberduck\Rubberduck.Deployment\BuildRegistryScript.ps1 -config 'Debug' -builderAssemblyPath 'N:\Repositories\Rubberduck\Rubberduck.Deployment\bin\Debug\Rubberduck.Deployment.dll' -netToolsDir 'C:\Program Files (x86)\Microsoft SDKs\Windows\v10.0A\bin\NETFX 4.6.1 Tools\' -wixToolsDir 'N:\Repositories\Rubberduck\Rubberduck.Deployment\WixToolset\' -sourceDir 'N:\Repositories\Rubberduck\Rubberduck.Deployment\bin\Debug\' -targetDir…

come to think of it... why can I and AV can build...

no, need to look at the Output tab

to see the error from powershell

7:58 PM

1>------ Build started: Project: RubberduckCodeAnalysis, Configuration: Release Any CPU ------
2>------ Build started: Project: Rubberduck.Resources, Configuration: Debug Any CPU ------
2>  Rubberduck.Resources -> N:\Repositories\Rubberduck\Rubberduck.Resources\bin\Debug\Rubberduck.Resources.dll
1>  RubberduckCodeAnalysis -> N:\Repositories\Rubberduck\RubberduckCodeAnalysis\bin\Release\RubberduckCodeAnalysis.dll
3>------ Build started: Project: Rubberduck.VBEditor, Configuration: Debug Any CPU ------

^^ same error as M.Doerner for me

I see it correctly falling back to tlbexp.exe since it didn't find the C++ build tool, good.

but cannot find the type?

that doesn't make sense.

that'd be referencing the BuildRegistryScript.cs in the Builders, which should be embedded in the Rubberduck.Deployment.dll assembly.

there is no "BuildRegistryScript.cs" there

there's only RegistryEntryBuilder

Hm, just a thought: VS is telling me that the class is never used. That might make the optimizer simply not compile it at all.

why now and not before?

@Vogel612, sorry, conflated the names there

8:06 PM

Have you ever tested without the C++ build tool?

i did but by aliasing out

I guess, AV has it.

yes AV does

[rubberduck-vba/Rubberduck] Vogel612 synchronized pull request #4006: Splice Off Rubberduck.Refactorings from Rubberduck.Core

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4006: Splice Off Rubberduck.Refactorings from Rubberduck.Core

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4006?src=pr&el=h1) Report
> Merging [#4006](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4006?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/d533a968cc883cc2347e16dc039913cfde9b0e80?src=pr&el=desc) will **decrease** coverage by `0.12%`.
> The diff coverage is `43.14%`.

```diff
@@ Coverage Diff @@
## next #4006 +/- ##
=======================

I might've used to have the C++ build tools as well...

just nuked them because I needed the space on my boot drive

8:08 PM

I think i have another VM that doesn't have C++ build tools - i'll test there to see if there's a difference.

assumed that aliasing the path was enough of a test to cause it to fall back

[rubberduck-vba/Rubberduck] Vogel612 removed label PR-Status: Conflicting from pull request #4006: Splice Off Rubberduck.Refactorings from Rubberduck.Core

Another thought, does the tool load the Rubberduck.Deplaoyment assembly into the CLR?

not sure if powershell loading it qualifies.

it does load it into the powershell script during build event and that was how it has been working prior to the PR 3975, as well

github.com/rubberduck-vba/Rubberduck/blob/next/…

ooooooohhhhhhhhh

i see the problem

that line is in totally wrong place

should not be even in the if block at all, I think.

if move that line to line 102, it should succeed, I think

which line?

from line 105 (linked above) to line 102

(e.g. outside the if block)

e.g. [System.Reflection.Assembly]::LoadFrom($builderAssemblyPath);

without that line, no types can be loaded. That'd be why.

8:19 PM

That did the trick.

Will PR in the hotfix.

[rubberduck-vba/Rubberduck] build for commit 3fedf3d4 on unknown branch: AppVeyor build succeeded

sorry about that, guys.

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4006: Splice Off Rubberduck.Refactorings from Rubberduck.Core

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4006?src=pr&el=h1) Report
> Merging [#4006](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4006?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/d533a968cc883cc2347e16dc039913cfde9b0e80?src=pr&el=desc) will **decrease** coverage by `0.26%`.
> The diff coverage is `44.76%`.

```diff
@@ Coverage Diff @@
## next #4006 +/- ##
=========================

[rubberduck-vba/Rubberduck] build for commit 3fedf3d4 on unknown branch: 53.4% (target 0%)

@this let he who never broke a build cast the first stone

3

[rubberduck-vba/Rubberduck] MDoerner created pull request #4035: Fixed assembly loading on fall back in deployment. to merge MDoerner/Rubberduck/HotfixForDeployment into rubberduck-vba/Rubberduck/next

> Fixes build error when building without the C++ build toll installed.

8:26 PM

Argh, typo.

I like the "build toll"

Is this another false positive inspection?

and c'mon, C++ isn't exactly toll-free.

LogManager.Register FileLogger.Create(Strings.Join(Array("DBLog", Format$(Now, "yyyy.mm.dd hh.mm"), ".txt"), vbNullString), loggingLevel, TempVars.Item("dbPath").Value, 10)

@FreeMan depends on the call sites

and yes

not quite an inspection issue though

8:27 PM

well, I'll not quite report it, then. ;)

I mean it's a resolver issue :)

[rubberduck-vba/Rubberduck] daFreeMan opened issue #4036: False positive 'Return value of fuction 'x' is never used'.

> The function:

```VBA
Implements ILogger

Public Function Create(ByVal loggerName As String, ByVal loggerMinLevel As logLevel) As ILogger

Dim result As New DebugLogger
result.Name = loggerName
result.MinLevel = loggerMinLevel
Set Create = result

End Function
```
The call site:

` LogManager.Register FileLogger.Create(Strings.Join(Array("DBLog", Format$(Now, "yyyy.mm.dd hh.mm"), ".txt"), vbNullString), loggingLevel, TempVars.Item("dbPath").Value, 10)`

And t

Also, note the Procedure is not referred to (165 results) immediately below it.

it is referred to, and there aren't 165 procedures in the class...

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> At the call site, is Create showing up as a member of FileLogger in the RD toolbar?

more like 8

8:33 PM

wtf

[rubberduck-vba/Rubberduck] daFreeMan created comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> Quarterly Report.DebugLogger.Create (function:ILogger) it says.

@Duga more wtf

[rubberduck-vba/Rubberduck] build for commit 5a3c5640 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit 5a3c5640 on unknown branch: 53.66% (target 0%)

I can create interesting code - you may not want me contributing to RD... ;)

@FreeMan interesting as sticking in a critical function call inside a if block? :p

8:43 PM

sure! ?

well, then you've met the low bar. ;)

@FreeMan looks like your FileLogger.Create method is yielding a DebugLogger instance... intended? or wrong logger?

or do you have both?

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> Ok so the good news is that the resolver is doing its job and correctly identifying member calls in an argument list.

The bad news is that it's not clear whether your `FileLogger` is actually creating a `DebugLogger`? Code says one thing, TD toolbar says another? I'm confused...

Hrm.... no, I'm actually creating a FileLogger, not a DebugLogger. So, it's not so much that the return value isn't used, it's that the function is never called.

ah, exactly!

not wrong per se, but definitely confusing

8:49 PM

[rubberduck-vba/Rubberduck] retailcoder added label feature-inspections to issue #4036: False positive 'Return value of fuction 'x' is never used'.

[rubberduck-vba/Rubberduck] retailcoder added label up-for-grabs to issue #4036: False positive 'Return value of fuction 'x' is never used'.

[rubberduck-vba/Rubberduck] retailcoder added label difficulty-01-duckling to issue #4036: False positive 'Return value of fuction 'x' is never used'.

[rubberduck-vba/Rubberduck] retailcoder added label bug to issue #4036: False positive 'Return value of fuction 'x' is never used'.

[rubberduck-vba/Rubberduck] retailcoder added label good first issue to issue #4036: False positive 'Return value of fuction 'x' is never used'.

[rubberduck-vba/Rubberduck] daFreeMan created comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> No, I'm confused. I went to the first `*Logger.Create` that the VBE's search found and forgot that there are both a `FileLogger` and a `DebugLogger`.

The `DebugLogger` isn't being used by my code, while the `FileLogger` is. The inspection takes me to `DebugLogger.Create`. So far, I haven't gotten to any references to `FileLogger.Create` not being used.

(I've got a fair few inspections to go through, but I'll hold out hope...)

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> Got a repro: the inspection yields a result for a function that's simply never invoked. While technically accurate, that is rather confusing.

The inspection needs to filter out functions that have `!References.Any()` - should be a trivial fix.

I guess the fact that Find all references takes me to the Set Create = result at the end of the proc, instead of listing a variety of locations shoulda been my clue bat.

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> ```csharp
var functions = State.DeclarationFinder
.UserDeclarations(DeclarationType.Function)
.Where(item => !IsIgnoringInspectionResultFor(item, AnnotationName))
.ToList();
```

If that's the only problem with it, then adding `&& item.References.Any()` to that `Where` predicate should do it.

@MathieuGuindon (or whoever else can merge) just so you did not miss it, PR 4035 should be merged (it built).

[rubberduck-vba/Rubberduck] retailcoder edited comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> ```csharp
var functions = State.DeclarationFinder
.UserDeclarations(DeclarationType.Function)
.Where(item => !IsIgnoringInspectionResultFor(item, AnnotationName))
.ToList();
```

<s>If that's the only problem with it, then adding `&& item.References.Any()` to that `Where` predicate should do it.</s>

Okay, a little twist: needs to exclude references *that are within the scope of the function itself*, since the "return value a

OMG edit history is FINALLY available on GH

8:54 PM

[rubberduck-vba/Rubberduck] retailcoder merged pull request #4035: Fixed assembly loading on fall back in deployment. from MDoerner/Rubberduck/HotfixForDeployment into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] MDoerner pushed commit 5a3c5640 to next: Fixed assembly loading on fall back in deployment.

[rubberduck-vba/Rubberduck] web-flow pushed commit 5db0eab0 to next

Merge pull request #4035 from MDoerner/HotfixForDeployment

Fixed assembly loading on fall back in deployment.

TTGH.

Saves Access.accdb

OK, home.

Work from home day - I get those every now and then...

lol

soo ... I'm kinda done with the pr-comments

[rubberduck-vba/Rubberduck] retailcoder published prerelease Rubberduck-v2.2.0.3288

[rubberduck-vba/Rubberduck] retailcoder created tag Rubberduck-v2.2.0.3288

[rubberduck-vba/Rubberduck] build for commit 5db0eab0 on next: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit 5db0eab0 on next: 53.66% (target 0%)

9:11 PM

Exercise: Find a flat, circular field 100km in radius. Build a tall, white, circular wall enclosing the field. Mexico will pay for it. Get a dreml rotary tool and a battery powered laser. Attach the laser to the business end of the dreml, get it spinning at 2000 revolutions per second. The spot that the laser makes on the wall is moving at 628000 km per second, twice the speed of light. How is this possible? — Eric Lippert 2 mins ago

"Mexico will pay for it"

lol

It fits in there so perfectly.

9:39 PM

TTGH

mansellan

@EricLippert Impossible, ridiculous and easily refuted: Mexico won't pay the wall. — Thorsten S. 10 mins ago

Some of the questions...

The wall would have to be really high or you will have to shave off quite a good amount off the earth.

Fun fact, if you are standing up, the horizon is about 14km away.

10:03 PM

@M.Doerner well you gotta build the wall with something, right?

puzzlepiece87

poofs in from the politics room lol

still open for reviews on the Refactorings split

10:36 PM

[rubberduck-vba/Rubberduck] MDoerner created comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> Why not simply check for the existence of references that are not assignments?

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> @MDoerner because we're looking for functions whose return value is never used; it's a declaration result. That said, a reference result that flags all places where a function is invoked like a sub would be rather useful too!

@MathieuGuindon AFAIK, only the lhs reference of an assignment gets flagged as assignment.

Having the function anywhere but in the function itself on the lhs of an assignment would be rather strange, I think.

10:55 PM

[rubberduck-vba/Rubberduck] mansellan commented on Rubberduck.Interaction/IMessageBox.cs of pull request #4006: Splice Off Rubberduck.Refactorings from Rubberduck.Core

[rubberduck-vba/Rubberduck] MDoerner created comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> I think you misunderstood my comment. I was commenting on enhancement of the condition in the current inspection and referring to the preceding comment.

11:25 PM

@M.Doerner sorry, I think that would work too, but then what the inspection projects into results needs to be the declaration for the function, otherwise we need to reword/rename it =)

11:39 PM

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #4036: False positive 'Return value of fuction 'x' is never used'.

> Right. Unless the function returns an object that has a default member.. although at this point IIRC the resolver sets the assignment flag on the default member and not the function, right?