VBA Rubberducking

12:01 AM

RELOAD!

[Hosch250/Rubberduck] 25 commits. 2902 additions. 1073 deletions.

[rubberduck-vba/Rubberduck] 28 commits. 9 opened issues. 2 closed issues. 18 issue comments. 3398 additions. 10975 deletions.

[SimonForsberg/minesweeper-api] 1 commit.

[SimonForsberg/minesweeper-commons] 1 commit.

[SimonForsberg/minesweeper-plugin1] 2 commits.

[SimonForsberg/minesweeperflagsgdx] 9 commits.

Duga

12:40 AM

[rubberduck-vba/Rubberduck] Hosch250 synchronized pull request #2149: Add a bunch of tests

[Hosch250/Rubberduck] Hosch250 pushed commit 332994fb to Issue2135: A bunch more tests

[Hosch250/Rubberduck] Hosch250 pushed commit 61a43eb7 to Issue2135: Merge branch 'Issue2135' of github.com/Hosch250/Rubberduck

Hosch250

We have 1604 tests now.

Duga

[rubberduck-vba/Rubberduck] build for commit 61a43eb7 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] Hosch250 merged pull request #2149: Add a bunch of tests from Hosch250/Rubberduck/Issue2135 into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] Hosch250 pushed commit 6938cd2c to next: Add a bunch of tests

[rubberduck-vba/Rubberduck] Hosch250 pushed commit 836d459f to next: Merge branch 'Issue2135' of github.com/Hosch250/Rubberduck

[rubberduck-vba/Rubberduck] web-flow pushed commit 6a3f2c0e to next: Merge branch 'next' into Issue2135

[rubberduck-vba/Rubberduck] Hosch250 pushed commit 332994fb to next: A bunch more tests

[rubberduck-vba/Rubberduck] Hosch250 pushed commit 61a43eb7 to next: Merge branch 'Issue2135' of github.com/Hosch250/Rubberduck

[rubberduck-vba/Rubberduck] web-flow pushed commit 15b9aead to next

Hosch250

It turns out those CanExecute tests took just enough time to cover the hour lost yesterday.

Duga

Merge pull request #2149 from Hosch250/Issue2135

Add a bunch of tests

Hosch250

And no, I didn't rig it.

Duga

12:43 AM

[Hosch250/Rubberduck] build for commit 61a43eb7 on Issue2135: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit 15b9aead on next: AppVeyor build succeeded

SlowLearner

1:44 AM

@Hosch250 honestly, I think we need another 444 tests ;)

ThunderFrame

1:56 AM

@SlowLearner - you owe MS an apology - the Print behavior is the same in Word and Excel.

'In a Class/Document Module
Sub foo()
  Print 1 '"Object doesn't support this property or method"
End Sub

'In a Standard Module
Sub foo()
  Print 1 '"Method not valid without suitable object"
End Sub

the MS docs had the clue:

> Not all methods can be performed by all objects

SlowLearner

in word it won't even compile: Print "a"

in excel it compiles....

ThunderFrame

> methods that need an explicit object qualifier when used in a standard module include [Line], Circle, Print, and PSet

SlowLearner

just read that

but - it's totally different. MS owes me an appology ;0

ThunderFrame

doesn't compile in a standard module in Excel or Word

does compile in a document module in Excel and Word

SlowLearner

Sub aTest()

print "a"

End Sub

won't compile in word

at least not on my PC

ThunderFrame

2:01 AM

is that in ThisDocument or Module1?

SlowLearner

in Word I tested in Module1

ThunderFrame

In 2013 32-bit, that's a run-time error in ThisDocument, but a comple error in Module1

The behavior is the same in Excel

(I could say it a 4th time, but it won't change)

;-)

SlowLearner

no... I failed to listen the first 3 times. Say it again. I know you're right

dafuggendifferencebetweenthisdocandmodule1

ThunderFrame

document module is a special kind of class module

standard module is not a class module

i.e ThisDocument is an Object

Module1 is not an Object

SlowLearner

erm... Sorry Microsoft. I've falsely accused you of not loving MS Word as much as Excel. I was wrong to think ? should behave in code the same way it behaves in the immediate window and even more wrong to assume that a class module and standard module would not amplify the differences. I am but a humbled SlowLearner

5

ThunderFrame

2:06 AM

try using Me in a standard module

SlowLearner

@ThunderFrame thanks for pointing this out to me; seriously, learning small things like this make a big difference :-) appreciate the education

@ThunderFrame yeah, I tried that once

walked away disappointed

Hosch250

2:34 AM

LOL.

Duga

6:19 AM

[Vogel612/TranslationHelper]‌ Vogel612 pushed commit 0050ccfc to TranslationProjects: Built skeleton structure for Project-Based working

[Vogel612/TranslationHelper]‌ Vogel612 pushed commit ff330835 to TranslationProjects: Rewrote FilesetOverviewModel to work with ResourceFiles and allow loading ResourceSets. Added unit-tests where relevant

[Vogel612/TranslationHelper]‌ Vogel612 pushed commit 4098c8df to TranslationProjects: Cleaned up unused (and leaky) Translation constructor

[Vogel612/TranslationHelper]‌ Vogel612 pushed commit 844833c8 to TranslationProjects: Fixed intermittent Test Failures due to dialogs being clingy

[Vogel612/TranslationHelper]‌ Vogel612 pushed commit d5a05d5b to TranslationProjects: Minor Argument Validation and some Unit-Tests

[Vogel612/TranslationHelper]‌ Vogel612 created pull request #47: Translation projects to merge TranslationProjects into master

> This will close #44 #45 and #46 when it's finished

[Vogel612/TranslationHelper] build for commit d5a05d5b on TranslationProjects: The Travis CI build failed

BUILD FAILURE!

[Vogel612/TranslationHelper] build for commit d5a05d5b on TranslationProjects: The Travis CI build failed

BUILD FAILURE!

[Vogel612/TranslationHelper]‌ Vogel612 synchronized pull request #47: Translation projects

[Vogel612/TranslationHelper]‌ Vogel612 pushed commit d9f1b137 to TranslationProjects: fixed initialization logic for ProjectSerializerTest

[Vogel612/TranslationHelper] build for commit d9f1b137 on TranslationProjects: The Travis CI build failed

BUILD FAILURE!

Vogel612

6:40 AM

@Duga what the heck, travis? This runs fine on my local machine...

mocks should work...

Duga

[Vogel612/TranslationHelper] build for commit d9f1b137 on TranslationProjects: The Travis CI build failed

BUILD FAILURE!

ThunderFrame

8:49 AM

axtools.com/products-codesmart-vb6.php

^ wut???

It's like they stole Cominterns branch!

Duga

10:05 AM

[rubberduck-vba/Rubberduck] Henque starred us

Quack Exchange

11:01 AM

0

Q: Partial anagram generator in VBA Excel

I'm trying to make a VBA code to arrange k-element subset of an n-set into some sequence. In other word, I'm trying to list all of k-permutations of n member set. For example, let's try to list all of 2-permutations of set {A,B,C} where each characters are located in the cells of Range("A1:C1"). ...

Zak

2:41 PM

Can I just say, the "Excel-Crashes-Upon-Exit" bug really needs to be fixed. It's causing me enough grief that I'm giving serious thought to uninstalling RD until it is.

Hosch250

3:43 PM

What?? That looks like a duplicate of RD with a different UI and a few different features!

Comintern

@Hosch250 - And it costs $249 more per seat...

Hosch250

BTW, based on SE's pay calculator, I'm worth 93K if I have 0 professional years and 1 skill.

If I have 2 skills, I'm worth 102K.

SlowLearner

@Hosch250 well, that would mean you have to sell 409 seats of this fake RD stuff....

and on that note, it's 1:45 am and bed awaits. Night all

Hosch250

However, on a more realistic note, I only have about half a skill based on their skill calculator.

Night.

And that means I'm worth about 86k.

Comintern

On a less realistic note, a developer with 0 years and 0 skills is apparently worth $86K.

Do you want fries with that?

Hosch250

3:51 PM

I assume you are insinuating that is too high?

Comintern

Around here, it's off by about $40K.

Hosch250

I've heard entry level positions pay pretty low.

I think they are implying they wouldn't hire anyone who wasn't worth the 86k.

Comintern

Yeah, the skill ratings only go down to 'B'.

Hosch250

Well, based on my grades, I'm pretty much a steady A student.

Does that count?

(Meant as a joke comment.)

Comintern

F - Could be less destructive. This is a good thing to avoid ever working on.

Duga

4:48 PM

[Hosch250/Rubberduck] Hosch250 pushed commit 09d024ea to Issue2135: Give Write Only Property a custom quick fix

[rubberduck-vba/Rubberduck] Hosch250 created pull request #2150: Give Write Only Property a custom quick fix to merge Hosch250/Rubberduck/Issue2135 into rubberduck-vba/Rubberduck/next

Hosch250

@Mat'sMug You might want to review that.

Duga

[rubberduck-vba/Rubberduck] Hosch250 synchronized pull request #2150: Give Write Only Property a custom quick fix

[Hosch250/Rubberduck] web-flow pushed commit 15b9aead to Issue2135

Merge pull request #2149 from Hosch250/Issue2135

Add a bunch of tests

[Hosch250/Rubberduck] web-flow pushed commit d67f66a8 to Issue2135: Merge branch 'next' into Issue2135

[Hosch250/Rubberduck] build for commit 09d024ea on Issue2135: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit d67f66a8 on unknown branch: AppVeyor build succeeded

[Hosch250/Rubberduck] build for commit d67f66a8 on Issue2135: AppVeyor build succeeded

Hosch250

6:06 PM

Working on fixing memory leaks. I've fixed a couple so far.

We are down to 49 survived objects.

And 774 new objects after killing it and running a couple GC's...

Hosch250

6:29 PM

Fixed more memory leaks.

I'm modifying some of the bindings quite heavily, so this will definitely need to be approved before merging.

Hosch250

6:42 PM

I have the UI being properly disposed now.

The biggest leak is the configs.

Comintern

@Hosch250 - What object is it leaking in the configs?

Hosch250

All of them.

It isn't properly cleaning up after itself.

Comintern

Is it leaking settings objects, or something from inside the PersistanceServices? IIR pretty much everything is singleton.

Duga

[Hosch250/Rubberduck] Hosch250 created branch ninjectMemoryLeaks

[Hosch250/Rubberduck] Hosch250 pushed commit 2d01027f to ninjectMemoryLeaks: Clean up memory leaks from the dockable presenters

[Hosch250/Rubberduck] Hosch250 pushed commit 64117cdf to ninjectMemoryLeaks: Fix messed-up config.

Hosch250

I'm checking something.

TBH, I'm kind of working half-blind here.

I just find what looks like a potential leak and clean it up and see if was the leak.

So far, I've been pretty successful, but that's mostly because I've done this a few times.

Comintern

6:49 PM

Ninject binding and only half-blind? You must have infra-vision to see into that black box.

Hosch250

LOL.

Take a look at that commit.

Notice how I specified exactly what type we need to get rid of the constructor args.

Duga

[Hosch250/Rubberduck] build for commit 64117cdf on ninjectMemoryLeaks: AppVeyor build succeeded

Hosch250

I've read that explicit ctor args in Ninject can be a cause of memory leaks, and that is one of the only places we aren't explicitly specifying a scope on the object.

I feel a little dizzy. Going to get a drink.

Comintern

Interesting. If that's the case, generic functions taking a T argument that are compiled into a non-interface concrete would leak like mad.

I.e. public T Load(T toDeserialize)...

Hosch250

Depends how T was injected, and its scope.

I've been trying to get a snapshot after my latest change, but it isn't working quite.

First Excel crashed when I unloaded RD to get the snapshot.

Next, I clicked Kill Process instead of Force GC after unloading RD.

Got it.

Comintern

6:59 PM

Oh wait. Settings objects aren't singleton bound - only the providers.

Hosch250

The providers are leaking.

I wonder... Should/could they be call bound?

Bind<IPersistanceService<CodeInspectionSettings>>().To<XmlPersistanceService<CodeInspectionSettings>>().InSingletonScope();

OK, so we bind that.

Where do we bind GeneralConfigProvider?

Bind<IIndenterConfigProvider>().To<IndenterConfigProvider>().InSingletonScope();
Bind<ISourceControlConfigProvider>().To<SourceControlConfigProvider>().InSingletonScope();

Those are the only config provider bindings we have, AFAICT.

Interestingly enough, those two aren't being leaked.

Comintern

I really don't see why not they can't be in call scope.

@Hosch250 Those two are the only ones where the Settings class is in a different assembly.

Hosch250

They still need to be bound, apparently.

Comintern

Doesn't that get caught with:
Kernel.Bind(t => t.From(assemblies)
.SelectAllClasses()
.InNamespaceOf<Configuration>()
.BindAllInterfaces()
.Configure(binding => binding.InSingletonScope()));

Hosch250

Should.

But, I changed that to Call Scope.

I'll just make it rebind them anyway.

Comintern

7:07 PM

The end goal of pulling the providers out was to get rid of Configuration and UserSettings classes. There are probably a bunch of places where an IPersistanceService<T> can be directly injected instead.

Would probably want to add a dirty flag to the settings objects to eliminate repetitive IO calls, but it was a first step to reducing coupling.

Mat's Mug

@Hosch250 best avoid ReBind as much as possible. That setup is already confusing enough as it is

Hosch250

OK, these are being retained by ConfigurationLoader.

Rebinding did absolutely nothing, so I removed it.

Comintern

7:26 PM

@Hosch250 - IIR, the only thing that really needs an IGeneralConfigService is the UI.

Hosch250

I got rid of all survived objects by making the menu items be created in call scope instead of singleton scope.

It doesn't seem to break anything.

Well, that wasn't a strict truth. It seems the module still references itself.

Down from about 3k retained Ninject items to about 300.

The absolute biggest leak right now is the settings.

Hosch250

7:59 PM

Looks like whoever did the settings didn't know about generics.

I just genericized the interface and was able to remove all but the indenter and SC settings interfaces (those are in different projects).

Actually, I could move that interface to the persistance project, and everything could access it.

Comintern

@Hosch250 I wonder who that could have been... :whistles:

Hosch250

;)

ThunderFrame

that VB6 add-in seems to have more UI tweaks than it does inspection/refactoring abilities.

Comintern

All eye candy, no substance.

ThunderFrame

hmmm, candy.

Comintern

8:10 PM

@Hosch250 - While you're poking around in there (and before I forget) - since you changed the provider scope to caller, don't forget to make the _emptyNamespace and _outputXmlSettings static in XmlPersistanceService<T>.

ThunderFrame

my guess is they're so dependent on the VB6 VBE that it was too hard to port it to VBA.

Comintern

They probably should have been static before - I just missed it.

@ThunderFrame Could be. When was it updated last?

ThunderFrame

well given the name includes '2013`?

May 9, 2014

Hosch250

I just PR'ed my memory leak fixes.

Comintern

From the screenshots it looks like they just do custom painting on top of the original codepanes.

Hosch250

8:15 PM

It isn't perfect, but it is a lot, and I'm not sure where to go next.

Comintern

Is @Duga asleep again?

Hosch250

Yeah.

Comintern

Understandable. She's been really busy the last couple of months.

Hosch250

in C# on Stack Overflow Chat, 11 mins ago, by Ryan Ternier

Question: I need a hammer, how do I get one.
Ultimate Answer: Make a factory to make hammers. Then you can make many types of hammers based on paramaters you give the factory
?But... I need to just hammer in a tiny nail to hang a picture
!GREAT! You just build the factory and then spend 1 hour picking out the hammer
?but... tiny... hammer?

OK, what's next?

Comintern

@Hosch250 Looking at that PR makes my change look really lazy.

ThunderFrame

8:21 PM

@Comintern Oh really? @Duga seems like all talk to me. I can't even remember her last commit.

3

Hosch250

What change?

Comintern

The interfaces on the settings.

Hosch250

Oh, yeah.

Don't worry, I've done the same.

Well, even in my last week here, I've fixed a lot of bugs.

Comintern

No worries at all. IIR I was in a hurry because of all the potential merge conflicts.

Hosch250

I essentially re-wrote the code generation for remove/reorder parameters. I've tested and debugged a bunch of commands and stuff...

Duga

8:23 PM

[Hosch250/Rubberduck] Hosch250 pushed commit 0b99ef92 to ninjectMemoryLeaks: Fix more memory leaks.

[Hosch250/Rubberduck] build for commit 0b99ef92 on ninjectMemoryLeaks: AppVeyor build succeeded

Hosch250

And now I just slaughtered a bunch of memory leaks.

Duga

[Hosch250/Rubberduck] Hosch250 pushed commit 968632dc to ninjectMemoryLeaks: Make EmptyNamespace and OutputXmlSettings static

[rubberduck-vba/Rubberduck] Hosch250 created pull request #2151: Ninject memory leaks to merge Hosch250/Rubberduck/ninjectMemoryLeaks into rubberduck-vba/Rubberduck/next

Hosch250

No walks today, it is storming out.

Duga

[rubberduck-vba/Rubberduck] Hosch250 synchronized pull request #2151: Ninject memory leaks

[Hosch250/Rubberduck] Hosch250 pushed 21 commits to ninjectMemoryLeaks

Hosch250

Not a nice, wild, thrilling storm; just one of those lazy city storms.

Duga

8:24 PM

[Hosch250/Rubberduck] web-flow pushed commit c1bbbe2b to ninjectMemoryLeaks: Merge branch 'next' into ninjectMemoryLeaks

[Hosch250/Rubberduck] build for commit 968632dc on ninjectMemoryLeaks: AppVeyor build failed

BUILD FAILURE!

Hosch250

Mostly rain with a couple flashes of lightning and a few far-off growls of thunder.

Duga

[rubberduck-vba/Rubberduck] build for commit 968632dc on unknown branch: AppVeyor build succeeded

[Hosch250/Rubberduck] build for commit c1bbbe2b on ninjectMemoryLeaks: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit c1bbbe2b on unknown branch: AppVeyor build succeeded

Hosch250

@Duga sigh

I wish I knew why some of these tests keep failing sporadically...

It might be because it is running multiple tests all at once, and there is something going on that way...

I'm not really convinced it is that, though.

ThunderFrame

@Comintern I watched some of the Local Window message flow when using the debugger. Seems like there's enough info (in conjunction with the key/mouse hooks) to determine stops, steps, stack-inspections, and continues.

I don't think there's anything else we'd need

Comintern

Taking a look at it. Is it sending custom messages?

ThunderFrame

8:31 PM

@Duga just PM'ed me and asked when my last commit was. :-/

Hosch250

LOL.

Comintern

LOL

Hosch250

I didn't know ordinary users could PM?

Comintern

Gah... I always forget not to mouse over windows when I'm watching the message pump.

ThunderFrame

well, it was more me reading my harsh words about @Duga, and then feeling remorse.

@ComIntern, yeah the SETTEXT, IIRC

Duga

8:36 PM

[Hosch250/Rubberduck] build for commit c1bbbe2b on ninjectMemoryLeaks: AppVeyor build succeeded

Comintern

OK, so it gets a WM_COMMAND EN_UPDATE every time the debugging context changes.

So... RD can verify that the project is in break mode, and then just snap to the current selection.

ThunderFrame

yes, but the SETTEXT also updates the name of the current procedure, and also when you inspect the stack.

Comintern

Wouldn't it be easier to just use the selection? It should always be in the correct procedure\module.

One sec. I'm just looking at the parent window.

ThunderFrame

yes, but I mean for detecting that the break has occurred, or the call stack view has changed

Comintern

I need longer test code to run.

:thumbsup: That looks perfect. "<Ready>", "Foo.Whereever", "<Running>"

Too bad the Call Stack window is modal.

Hosch250

8:55 PM

I wonder if @FreeMan is around.

I could do an "internal" build for him to see if the crashes are mitigated at all.

Comintern

The access violations?

Hosch250

Apparently, Excel was just going down a bunch

I don't know if it was all access violations or what.

ThunderFrame

@Comintern yeah, but we can tell the outcome of a Call Stack window because the user hasn't continued the code, but the SETTEXT message has fired.

Comintern

Yeah, they were all access violations in #2147. Line 285 of DeclarationExtensions in FindBuiltInEventHandlers.

@ThunderFrame I was thinking of the possibility of an RD error object. :-)

Duga

[rubberduck-vba/Rubberduck] Hosch250 commented on issue #2056: Test explorer is freezing when trying to run tests with early bound Assert without Rubberduck reference

> So, basically check whether any non-builtin declarations have the AsTypeName of Rubberduck.PermissiveAssertClass or Rubberduck.AssertClass and automatically reference RD's library if there are any?

Duga

9:17 PM

[Hosch250/Rubberduck] Hosch250 created branch Issue2056

[Hosch250/Rubberduck] Hosch250 pushed commit acd50380 to Issue2056: Close #2056

[rubberduck-vba/Rubberduck] Hosch250 created pull request #2152: Ensure projects reference the RD library before running tests to merge Hosch250/Rubberduck/Issue2056 into rubberduck-vba/Rubberduck/next

> If any tests have the `AsTypeName` of `Rubberduck.PermissiveAssertClass` or `Rubberduck.AssertClass`, that is.

Close #2056

[rubberduck-vba/Rubberduck] Hosch250 assigned Hosch250 to issue #2056: Test explorer is freezing when trying to run tests with early bound Assert without Rubberduck reference

[Hosch250/Rubberduck] build for commit acd50380 on Issue2056: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit acd50380 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] Hosch250 merged pull request #2152: Ensure projects reference the RD library before running tests from Hosch250/Rubberduck/Issue2056 into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] Hosch250 closed issue #2056: Test explorer is freezing when trying to run tests with early bound Assert without Rubberduck reference

[rubberduck-vba/Rubberduck] Hosch250 pushed commit acd50380 to next: Close #2056

[rubberduck-vba/Rubberduck] web-flow pushed commit bbeaffb7 to next

Merge pull request #2152 from Hosch250/Issue2056

Ensure projects reference the RD library before running tests

[rubberduck-vba/Rubberduck] Hosch250 synchronized pull request #2150: Give Write Only Property a custom quick fix

[Hosch250/Rubberduck] Hosch250 pushed 1 commit to Issue2135

[Hosch250/Rubberduck] web-flow pushed commit bbeaffb7 to Issue2135

Merge pull request #2152 from Hosch250/Issue2056

Ensure projects reference the RD library before running tests

[Hosch250/Rubberduck] web-flow pushed commit 23201492 to Issue2135: Merge branch 'next' into Issue2135

[rubberduck-vba/Rubberduck] build for commit bbeaffb7 on next: AppVeyor build succeeded

[Hosch250/Rubberduck] build for commit 23201492 on Issue2135: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit 23201492 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] Hosch250 merged pull request #2150: Give Write Only Property a custom quick fix from Hosch250/Rubberduck/Issue2135 into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] Hosch250 pushed commit 09d024ea to next: Give Write Only Property a custom quick fix

[rubberduck-vba/Rubberduck] web-flow pushed commit d67f66a8 to next: Merge branch 'next' into Issue2135

[rubberduck-vba/Rubberduck] web-flow pushed commit 23201492 to next: Merge branch 'next' into Issue2135

[rubberduck-vba/Rubberduck] web-flow pushed commit c2d06963 to next

Merge pull request #2150 from Hosch250/Issue2135

Give Write Only Property a custom quick fix

[rubberduck-vba/Rubberduck] Hosch250 synchronized pull request #2151: Ninject memory leaks

[Hosch250/Rubberduck] Hosch250 pushed 5 commits to ninjectMemoryLeaks

[Hosch250/Rubberduck] web-flow pushed commit c2d06963 to ninjectMemoryLeaks

Merge pull request #2150 from Hosch250/Issue2135

Give Write Only Property a custom quick fix

[Hosch250/Rubberduck] web-flow pushed commit 458ec20a to ninjectMemoryLeaks: Merge branch 'next' into ninjectMemoryLeaks

[rubberduck-vba/Rubberduck] build for commit c2d06963 on next: AppVeyor build succeeded

[Hosch250/Rubberduck] build for commit 458ec20a on ninjectMemoryLeaks: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit 458ec20a on unknown branch: AppVeyor build succeeded

ThunderFrame

@Comintern an RD Error object? You're 2 steps ahead of me. What are you thinking?

Comintern

9:35 PM

It was just a passing thought. I envisioned this:

Private Foo As Rubberduck.Callstack

Private Sub Bar()
    On Error GoTo Label
    'Codez
Label:
    Debug.Print Foo.Stack
End Sub

Hosch250

You mean the poor VBA programmers don't have call stacks?

Comintern

@Hosch250 - Only available through the VBE in break mode - there's no way to use them in code.

ThunderFrame

oh, so just like using a collection as a stack to push and pop the proc names?

Comintern

Yeah - like the class Mat put on CR the other day.

But more magical.

With rainbows and unicorns and stuff.

SlowLearner

Morning all

Duga

9:40 PM

[rubberduck-vba/Rubberduck] Hosch250 commented on issue #166: ImplicitVariantReturnTypeInspection quick-fix removes all comments

> Very much obsolete.
```
' test test test test test
Function foo( _
val1 As Variant, _
val2) ' test test test test test

End Function
```
Is turned into:
```
' test test test test test
Function foo( _
val1 As Variant, _
val2) As Variant ' test test test test test

End Function
```

[rubberduck-vba/Rubberduck] Hosch250 closed issue #166: ImplicitVariantReturnTypeInspection quick-fix removes all comments

SlowLearner

@Comintern NSFW - but Unicorns are cool: youtube.com/watch?v=YbYWhdLO43Q

Duga

[rubberduck-vba/Rubberduck] Hosch250 removed label status-deferred from issue #166: ImplicitVariantReturnTypeInspection quick-fix removes all comments

SlowLearner

warning - you can't unsee that

Hosch250

I hate unicorns.

Comintern

I've never met one.

SlowLearner

9:42 PM

@Hosch250 WTF how can you hate VBA and unicorns? They're totally opposite!?

Hosch250

They are even more stupid than dragons. At least dragons are useful for something, even if it is just burning up 17'th century enemy foot soldiers.

@SlowLearner Because I don't like extremes.

Comintern

@SlowLearner The enemy of my enemy is not necessarily my friend.

Hosch250

BTW, look at this real-life unicorn:

SlowLearner

I never knew that unicorns were the enemy of VBA... I did hear that the reason they're now extinct ... cdn-images.threadless.com/threadless-shop/products/368/…

Comintern

I see two. They just lack bilateral symmetry.

SlowLearner

9:48 PM

@Hosch250 what about Narwhals? You can't hate them

ThunderFrame

The friend of my friend is an Access violation.

Comintern

@SlowLearner It was a horrible misunderstanding. They never should have coded that whole damn thing in VBA.

SlowLearner

@ThunderFrame fuckimslow

I just don't know enough

In case you need a distraction.... this song is stupid, but oddly cool youtube.com/watch?v=anM1N5oN-OM

Vogel612

weebl warning

better than badgers though..

Hosch250

I'm sick of sitting here.

Someone give me something to do.

Comintern

9:56 PM

@ThunderFrame I thought that in VBA that a Friend of a Friend was a compiler error because there is only PublicNotCreatable and Private instancing.

2

"I would like to provide a more complete answer, but not sure if you can write a function ..." SNAP!

SlowLearner

@Comintern it was a while ago, it was the most advanced IDE they had

Duga

[rubberduck-vba/Rubberduck] Hosch250 commented on issue #1258: App.cs does too many things

> This is pretty old now. Is it cleaned up enough to suit you?

[rubberduck-vba/Rubberduck] Hosch250 commented on issue #341: Add Search & Navigation to the Rubberduck commandbar

> All of these are done an integrated except #327. What's the plan for this?

Hosch250

Hi, @RubberDuck.

RubberDuck

10:12 PM

Hi @Hosch250!

Duga

[rubberduck-vba/Rubberduck] Hosch250 commented on issue #1043: Update and clean up localized strings for release

> Swedish and Japanese will not be in the 2.0 release. The French and German translations still need a couple tweaks, but this is essentially done.

Peter M Taylor

A brief morning to all

Duga

[rubberduck-vba/Rubberduck] Hosch250 closed issue #1043: Update and clean up localized strings for release

Hosch250

2.0 hit 95% done: github.com/rubberduck-vba/Rubberduck/milestone/20

Comintern

Morning @PeterMTaylor (UTC+/-16).

Peter M Taylor

10:14 PM

:)

I found something interesting about machine learning and leant it better in 5 mins. r2d3.us/visual-intro-to-machine-learning-part-1

Duga

[Hosch250/Rubberduck] Hosch250 pushed commit 596f1f09 to Issue2056: Close #984 (delete French translation to keep us reminded)

[rubberduck-vba/Rubberduck] Hosch250 created pull request #2153: Close #984 (delete French translation to keep us reminded) to merge Hosch250/Rubberduck/Issue2056 into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] Hosch250 synchronized pull request #2153: Close #984 (delete French translation to keep us reminded)

[Hosch250/Rubberduck] Hosch250 pushed 5 commits to Issue2056

[Hosch250/Rubberduck] web-flow pushed commit cd9d5750 to Issue2056: Merge branch 'next' into Issue2056

Peter M Taylor

I love how they displayed the decision tree vertically. Made me wonder how RD tried to parse code when they showed the flow of data...

Well done @Hosch250 to reach that target

Duga

[Hosch250/Rubberduck] build for commit 596f1f09 on Issue2056: AppVeyor build succeeded

Hosch250

You know, we released 1.4.3 at around this point.

Duga

[rubberduck-vba/Rubberduck] build for commit cd9d5750 on unknown branch: AppVeyor build succeeded

Hosch250

10:19 PM

Mat just marked everything else as the next release, or as status-declined.

Duga

[rubberduck-vba/Rubberduck] Hosch250 merged pull request #2153: Close #984 (delete French translation to keep us reminded) from Hosch250/Rubberduck/Issue2056 into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] Hosch250 closed issue #984: UntypedFunctionUsageInspection quickfix can introduce runtime errors

Hosch250

Just 24 measly issues marked as 2.0.

Duga

[rubberduck-vba/Rubberduck] Hosch250 pushed commit 596f1f09 to next: Close #984 (delete French translation to keep us reminded)

[rubberduck-vba/Rubberduck] web-flow pushed commit cd9d5750 to next: Merge branch 'next' into Issue2056

[rubberduck-vba/Rubberduck] web-flow pushed commit a9da1687 to next

Merge pull request #2153 from Hosch250/Issue2056

Close #984 (delete French translation to keep us reminded)

[rubberduck-vba/Rubberduck] build for commit 596f1f09 on next: AppVeyor build cancelled

BUILD FAILURE!

[Hosch250/Rubberduck] build for commit cd9d5750 on Issue2056: AppVeyor build succeeded

Hosch250

Of course, I need @Vogel612 to translate one new inspection value, and @Mat'sMug to translate a bunch of new values in the main, inspection, and SC resx's.

Duga

[rubberduck-vba/Rubberduck] build for commit a9da1687 on next: AppVeyor build succeeded

Peter M Taylor

10:24 PM

@Comintern I have put some thought into the auditing process to help RD which I am happy to do. Would you have any objections as my how question if I Learnt more on Python so I can use scrapy to make a spider to grab the list of about 22,000 articles of C#

@Comintern and cross reference this against the latest fork of RD any code snips.

Comintern

@PeterMTaylor No problem at all. And no rush. It's too bad the code project dev dropped off the map.

Hosch250

What exactly is this all about?

Peter M Taylor

cool

Comintern

@Hosch250 #1893. I suspect it's subtly broken actually.

Hosch250

Oh.

Peter M Taylor

10:28 PM

@Hosch250 Comintern just asked if I could help RD to just check our code if anyone in CodeProject has published their to certify our compliance to their CPOL licence

In regards to RD code.

Hosch250

Oh.

Could be tricky.

Peter M Taylor

So I felt I would want a automate this process and found a chance to learn a new language this way and checked it's ok

Hosch250

Could be extremely tricky.

Comintern

May be the root cause of the escape key thing in #2087.

Peter M Taylor

I know at least I get a new soft skill in this process.

Hosch250

10:29 PM

Does it count if we just reformat/rename things?

Peter M Taylor

Using Powershell to compare files.

Hosch250

I don't think so, but I can guarantee that almost all copy/pasta code will have been...

Hi, @chris.1997.

Peter M Taylor

Anyway at work now and have a good day.

Comintern

@Hosch250 No, needs to be white-room or else it's considered a derived work.

Hosch250

That's what I thought.

A simple compare will most likely not catch anything.

Comintern

10:34 PM

Although I'm guessing that a bunch of it isn't subject to copy-write because it's using publicly documented APIs. The structure is though.

Hosch250

10:48 PM

Hmm, found another potential memory leak.

SlowLearner

@Comintern @PeterMTaylor @Hosch250 So my question, based on a position of total ignorance, is how would someone verify whether or not the code in question (with the restrictive license) was not originally based on some other code, using some similar structure which was available under a less restrictive license and thus throwing into question the validity of the assumed licence restrictions?

Transcript for