VBA Rubberducking

Duga

00:01

RELOAD!

[retailcoder/Rubberduck] 8 commits. 135 additions. 118 deletions.

[rubberduck-vba/Rubberduck] 9 commits. 241 additions. 205 deletions.

[Zomis/Duga] 7 commits. 144 additions. 15 deletions.

[Zomis/test] 4 commits. 6 additions.

ThunderFrame

00:16

@Mat'sMug that VB6 code adviser supposedly has snap-in inspections, but I can't find how to define one.

Mat's Mug

00:48

Now I'm really tempted to try a VB6 Rubberduck build

> Rubberduck.Setup.1.4.3.0.exe (4.32MB) - Downloaded 2712 times.
Last updated on 2015-07-08

Like, 2.0-tempted

ThunderFrame

@Mat'sMug you need to custom install the code adviser, and check the developer options. Beyond that, I can't find the definitions.

Duga

05:08

[retailcoder/Rubberduck] retailcoder pushed commit e4ebaca3 to next: restricted SelfAssignedDeclarationInspection to non-ValueType variables with a specified type

[retailcoder/Rubberduck] retailcoder pushed commit c84e8bc4 to next: implemented ObjectVariableNotSetInspection; closes #383

Duga

05:32

[retailcoder/Rubberduck] retailcoder pushed commit d7da07de to next: fixed bug with SelfAssignedDeclarationInspection

Duga

06:32

[rubberduck-vba/Rubberduck] retailcoder created pull request #1344: ObjectVariableNotSetInspection to merge retailcoder/Rubberduck/next into rubberduck-vba/Rubberduck/next

> Private Sub Workbook_Open()

Dim target As Range
target = Range("A1")

target.Value = "forgot something?"

End Sub

Also includes other fixes and tweaks

[rubberduck-vba/Rubberduck] retailcoder merged pull request #1344: ObjectVariableNotSetInspection from retailcoder/Rubberduck/next into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] retailcoder closed issue #383: Object variable not set

[rubberduck-vba/Rubberduck] build for commit d7da07de on next: AppVeyor build cancelled

BUILD FAILURE!

[rubberduck-vba/Rubberduck] retailcoder pushed commit 6791992e to next: remove built-in declarations for removed references

[rubberduck-vba/Rubberduck] retailcoder pushed commit d30906f9 to next: random tweaks to mousehook and codepanerefactorrenamecommand

[rubberduck-vba/Rubberduck] retailcoder pushed commit e4ebaca3 to next: restricted SelfAssignedDeclarationInspection to non-ValueType variables with a specified type

[rubberduck-vba/Rubberduck] retailcoder pushed commit c84e8bc4 to next: implemented ObjectVariableNotSetInspection; closes #383

[rubberduck-vba/Rubberduck] retailcoder pushed commit d7da07de to next: fixed bug with SelfAssignedDeclarationInspection

[rubberduck-vba/Rubberduck] retailcoder pushed commit b62455a1 to next

Merge pull request #1344 from retailcoder/next

ObjectVariableNotSetInspection

[rubberduck-vba/Rubberduck] build for commit b62455a1 on next: AppVeyor build succeeded

Mat's Mug

wow, I just noticed, we have more followers than @MZTooolsSoftware on Twitter!

I think this is it

TTGTB

2.0a-pre release notes secret gist

2

Chris

07:00

@Mat'sMug: several things are already covered in RD, such as source control support, I just need some time to get used to it. I'll get back here once additional wishes come up, but I think I'll wait until release of 2.0 and have a look at that first. Other beefs that I do have are currently more about the language itself.

Duga

09:43

[rubberduck-vba/Rubberduck] autoboosh added label vba-grammar to issue #1345: SHORTLITERAL/INTEGERLITERAL chaos

[rubberduck-vba/Rubberduck] autoboosh assigned autoboosh to issue #1345: SHORTLITERAL/INTEGERLITERAL chaos

[rubberduck-vba/Rubberduck] autoboosh added label bug to issue #1345: SHORTLITERAL/INTEGERLITERAL chaos

[rubberduck-vba/Rubberduck] autoboosh opened issue #1345: SHORTLITERAL/INTEGERLITERAL chaos

> The lexer shouldn't actually produce a short literal because that's not part of the spec. Also, non-floating numbers with an exponent should actually be classified as floating literals. Here are the correct lexer rules:

```
FLOATLITERAL :
FLOATINGPOINTLITERAL FLOATINGPOINTTYPESUFFIX?
| DECIMALLITERAL FLOATINGPOINTTYPESUFFIX;
fragment FLOATINGPOINTLITERAL :
DECIMALLITERAL EXPONENT
| DECIMALLITERAL '.' DECIMALLITERAL? EXPONENT?
| '.' DECIMALLITERAL EXPONENT?;
INTE

[rubberduck-vba/Rubberduck] autoboosh added label vba-grammar to issue #1346: Alternative NEQ/GEQ/LEQ are missing in the grammar

[rubberduck-vba/Rubberduck] autoboosh assigned autoboosh to issue #1346: Alternative NEQ/GEQ/LEQ are missing in the grammar

[rubberduck-vba/Rubberduck] autoboosh added label bug to issue #1346: Alternative NEQ/GEQ/LEQ are missing in the grammar

[rubberduck-vba/Rubberduck] autoboosh opened issue #1346: Alternative NEQ/GEQ/LEQ are missing in the grammar

> Lexer rules:

```
GEQ : '>=';
ALTGEQ : '=>';
LEQ : '<=';
ALTLEQ : '=<';
NEQ : '<>';
ALTNEQ : '><';
```

The ALT-rules are currently missing. The VBE actually transforms the alternative into the "main" one but I don't know if every VBA editor out there does that.

[rubberduck-vba/Rubberduck] autoboosh assigned autoboosh to issue #1347: Resolver: Introduce Declaration Hierarchy

[rubberduck-vba/Rubberduck] autoboosh added label code-parsing to issue #1347: Resolver: Introduce Declaration Hierarchy

[rubberduck-vba/Rubberduck] autoboosh opened issue #1347: Resolver: Introduce Declaration Hierarchy

> ProjectDeclarations have a new priority id (?) which should probably be saved in the Declaration class. Since it's long overdue to introduce a class hierarchy we might as well use this opportunity to refactor the code.

One class per DeclarationType-enum member?

[rubberduck-vba/Rubberduck] autoboosh commented on issue #1247: Parser Error: Open "x" for Input As #iFNotAnIfStatement

> Found a solution to this, already tested locally but haven't had time to make a PR yet.

[rubberduck-vba/Rubberduck] autoboosh assigned autoboosh to issue #1348: Implements should allow expressions instead of mere identifiers

[rubberduck-vba/Rubberduck] autoboosh added label bug to issue #1348: Implements should allow expressions instead of mere identifiers

[rubberduck-vba/Rubberduck] autoboosh opened issue #1348: Implements should allow expressions instead of mere identifiers

> This is allowed:

`Implements VBAProject.ITest`

But destroys the parser.

"Correct" parser rule:

`implementsStmt : IMPLEMENTS whiteSpace valueStmt;`

I'd like to take this as an opportunity to sneak in expressions per VBA spec by creating an ExpressionParser. The expression parser could parse the valueStmt and build an expression tree. We could then hand off this tree to the binder. I'd start with simple-name-expressions in this case because that's what we currently support. This w

[rubberduck-vba/Rubberduck] autoboosh added label vba-grammar to issue #1348: Implements should allow expressions instead of mere identifiers

[rubberduck-vba/Rubberduck] autoboosh commented on issue #1247: Parser Error: Open "x" for Input As #iFNotAnIfStatement

> Found a solution to this, already tested locally but haven't had time to make a PR yet.

It's just like @retailcoder and @dimitrimitropulos said, by considering the fact that these conditional compilation keywords may only appear on a separate line we can catch the cases described in the first comment.

Duga

10:48

[rubberduck-vba/Rubberduck] Vogel612 commented on issue #1348: Implements should allow expressions instead of mere identifiers

> I'm not sure that this is a valueStmt. Isn't it more correct to see this as a FQN? or are these in fact valueStmts and only the resolver gives them the notion of being a Type?

[rubberduck-vba/Rubberduck] Vogel612 commented on issue #1346: Alternative NEQ/GEQ/LEQ are missing in the grammar

> Would it not be easier for the resolver if these were not separate Lexer rules, but combined ones?

GEQ : '>=' | '=>';
[...]

[rubberduck-vba/Rubberduck] Vogel612 commented on issue #1346: Alternative NEQ/GEQ/LEQ are missing in the grammar

> Would it not be easier for the resolver if these were not separate Lexer rules, but combined ones?

GEQ : '>=' | '=>';
[...]

[rubberduck-vba/Rubberduck] autoboosh commented on issue #1348: Implements should allow expressions instead of mere identifiers

> There's a bit of a chaos in the grammar right now because valueStmts should actually be expressions.

The VBA spec says that the implements-directive takes a class-type-name: https://msdn.microsoft.com/en-us/library/ee177434.aspx

A class-type-name in turn is defined as a defined-type-expression: https://msdn.microsoft.com/en-us/library/ee200228.aspx.

A defined-type-expression is then defined as simple-name-expression (basically an "ambiguousIdentifier") and a member-access-expression (

[rubberduck-vba/Rubberduck] autoboosh commented on issue #1348: Implements should allow expressions instead of mere identifiers

> There's a bit of a chaos in the grammar right now because valueStmts should actually be expressions.

The VBA spec says that the implements-directive takes a class-type-name: https://msdn.microsoft.com/en-us/library/ee177434.aspx

A class-type-name in turn is defined as a defined-type-expression: https://msdn.microsoft.com/en-us/library/ee200228.aspx.

A defined-type-expression is then defined as simple-name-expression (basically an "ambiguousIdentifier") or a member-access-expression (t

[rubberduck-vba/Rubberduck] autoboosh commented on issue #1348: Implements should allow expressions instead of mere identifiers

> There's a bit of a chaos in the grammar right now because valueStmts should actually be expressions.

The VBA spec says that the implements-directive takes a class-type-name: https://msdn.microsoft.com/en-us/library/ee177434.aspx

A class-type-name in turn is defined as a defined-type-expression: https://msdn.microsoft.com/en-us/library/ee200228.aspx.

A defined-type-expression is then defined as simple-name-expression (basically an "ambiguousIdentifier") or a member-access-expression (t

[rubberduck-vba/Rubberduck] autoboosh commented on issue #1346: Alternative NEQ/GEQ/LEQ are missing in the grammar

> You're right, that makes a lot more sense, thanks :)

Duga

11:14

[rubberduck-vba/Rubberduck] autoboosh created pull request #1349: shortliteral + neq/geq/leq cleanup (#1345 and #1346) to merge autoboosh/Rubberduck/shortliteralfix into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] build for commit 77f79942 on unknown branch: AppVeyor build succeeded

Duga

12:24

[rubberduck-vba/Rubberduck] dimitrimitropulos opened issue #1350: ByVal/ByRef with ParamArray not possible

> With the following code:
```
Public Sub test(ParamArray someArray() As Variant)

End Sub
```

the Code Inspections throws an "implicit ByRef parameter". The "Fix" then changes the code to this:
```
Public Sub test(ByRef ParamArray someArray() As Variant)

End Sub
```
which is a compile error.

Duga

12:44

[rubberduck-vba/Rubberduck] autoboosh added label bug to issue #1350: ByVal/ByRef with ParamArray not possible

[rubberduck-vba/Rubberduck] autoboosh added label vba-grammar to issue #1350: ByVal/ByRef with ParamArray not possible

[rubberduck-vba/Rubberduck] autoboosh assigned autoboosh to issue #1350: ByVal/ByRef with ParamArray not possible

[rubberduck-vba/Rubberduck] autoboosh commented on issue #1350: ByVal/ByRef with ParamArray not possible

> Nice find. Paramarrays are apparently always ByRef. I'll take a look :)

Duga

13:33

[rubberduck-vba/Rubberduck] retailcoder removed label vba-grammar from issue #1350: ByVal/ByRef with ParamArray not possible

[rubberduck-vba/Rubberduck] retailcoder added label code-inspections to issue #1350: ByVal/ByRef with ParamArray not possible

Duga

14:26

[rubberduck-vba/Rubberduck] autoboosh created pull request #1351: associate project references correctly and in the correct order (#1340, #1347) to merge autoboosh/Rubberduck/declarationhierarchy into...

rubberduck-vba/Rubberduck/next

> There's a new class ProjectDeclaration that contains the references to other projects with the priorities. Because this happens in the declaration gathering phase we can't add ProjectDeclarations as other references. I changed the Reference.ReferenceId to ProjectId and also switched from using the ProjectName as ProjectId for references to using a combination of the name + path. Otherwise could not differentiate between two libraries with the same name. Not sure if this will always work...

though, there might be some cases I don't know about. Not so sure whether it's possible to reference two libraries with the same name in the same DLL either.

[rubberduck-vba/Rubberduck] build for commit 729d09ca on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] autoboosh created pull request #1352: fix paramarray incorrectly flagged as missing ByRef param mechanism (… to merge autoboosh/Rubberduck/paramarrayfix into rubberduck-vba/Rubberduck/next

> …fixes #1350)

[rubberduck-vba/Rubberduck] build for commit 06b930c5 on unknown branch: AppVeyor build failed

BUILD FAILURE!

[rubberduck-vba/Rubberduck] autoboosh synchronized pull request #1352: fix paramarray incorrectly flagged as missing ByRef param mechanism (…

[rubberduck-vba/Rubberduck] build for commit debc72e9 on unknown branch: AppVeyor build succeeded

Duga

17:40

[rubberduck-vba/Rubberduck] comintern opened issue #1353: Auto-save should run any time the compiler is invoked

> I lost track of how many times I've done something bone-headed and either crashed the host application or put it into an irrecoverable not responding state and had to end-task it. One of the more overlooked features of Visual Studio is that it saves all of your open documents when you build a project. It would be nice to be able to recover from one's own stupid decision to run something before the host document is saved.

Duga

18:19

[rubberduck-vba/Rubberduck] retailcoder commented on issue #1353: Auto-save should run any time the compiler is invoked

> Any ideas on how to capture that "build" event?

puzzlepiece87

18:40

Good luck this weekend y'all! Hope the MVP presentation goes well.

Mat's Mug

aye, I need to package that pre-release build tonight - no ifs, no buts!

@puzzlepiece87 have you read the 2.0a release notes yet?

12 hours ago, by Mat's Mug

2.0a-pre release notes secret gist

puzzlepiece87

19:11

@Mat'sMug I have now, thanks for the link!

For EmptyStringLiteralInspection, I believe there may be a few edge cases that require it.

*that require "" instead of vbNullString

If I remember any of them I will pass them on

Oh oh

Mat's Mug

Interesting, got an example?

puzzlepiece87

It has to do with detecting empty cells I think

Let me try to find something I've written

Oh, I think I am confusing the case

Yes I am

Mat's Mug

lol

puzzlepiece87

I am confusing the case where there can be a Cell1.Value = "" but Cell1 Is Empty = False.

That's definitely a thing that happens, it's frustrating and requires clearing.

Mat's Mug

@INOPIAE @Vogel612 FYI I merged a new last-minute inspection yesterday, AFAIK it's the only missing German translation (see ObjectVariableNotSetInspection)

@puzzlepiece87 ah, yes I see.. perhaps that could be the basis of another inspection?

puzzlepiece87

19:17

I'm not sure how useful that would be to others, but I have had a few cases where I've needed empty cells and been frustrated by that.

I would not have mentioned it, for lack of value reasons, if I hadn't momentarily confused it for a ""/vbnullstring conflict

Mat's Mug

Keep in mind that inspection severity can be configured, from DoNotShow to Error

Duga

[rubberduck-vba/Rubberduck] comintern commented on issue #1353: Auto-save should run any time the compiler is invoked

> This would definitely go in the "experimental" category at this point - I need to poke around to see if I can find a suitable COM event. That would go far beyond lucky though.

If not, it would probably be substantially more work. My current thought from that angle would be to put a trampoline function on the entry points of the compiler that calls back to RD and delays compilation until RD says it's OK.

This is tangentially related to my pondering about replacing the code pane - it wou

Mat's Mug

IMO it's a very good inspection, could be worded as "Cells containing an empty string are not considered 'empty'. If this is looking for empty cells, the code will not work as intended." -- even an @Ignore EmptyCellCheckInspection annotation on top of that line would be informative about the code's intent I think.

puzzlepiece87

20:11

Oh cool, then I am glad it's potentially helpful :)

I'm not a real developer, but I figured that might add more overhead than it was worth. I'm glad it might be a useful feature :)

I'm going to try to reproduce that for you to make absolutely sure I have it straight

Mat's Mug

when you have is sorted out, make an issue for it on GitHub!

puzzlepiece87

okay!

Duga

20:43

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/Inspections/ImplicitByRefParameterInspection.cs of pull request #1352: fix paramarray incorrectly flagged as missing ByRef param mechanism (…

[rubberduck-vba/Rubberduck] retailcoder closed issue #1350: ByVal/ByRef with ParamArray not possible

[rubberduck-vba/Rubberduck] retailcoder merged pull request #1352: fix paramarray incorrectly flagged as missing ByRef param mechanism (… from autoboosh/Rubberduck/paramarrayfix into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] autoboosh pushed commit 06b930c5 to next: fix paramarray incorrectly flagged as missing ByRef param mechanism (fixes #1350)

[rubberduck-vba/Rubberduck] autoboosh pushed commit debc72e9 to next: fix failing test

[rubberduck-vba/Rubberduck] retailcoder pushed commit d5efed8d to next

Merge pull request #1352 from autoboosh/paramarrayfix

fix paramarray incorrectly flagged as missing ByRef param mechanism (…

[rubberduck-vba/Rubberduck] retailcoder merged pull request #1349: shortliteral + neq/geq/leq cleanup (#1345 and #1346) from autoboosh/Rubberduck/shortliteralfix into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] autoboosh pushed commit 77f79942 to next: shortliteral + neq/geq/leq cleanup (#1345 and #1346)

[rubberduck-vba/Rubberduck] retailcoder pushed commit ee0fdab5 to next

Merge pull request #1349 from autoboosh/shortliteralfix

shortliteral + neq/geq/leq cleanup (#1345 and #1346)

puzzlepiece87

Not able to reproduce at this time unfortunately

Not sure if they changed something from Excel 2010 to 2013, which we just upgraded to.

Duga

[rubberduck-vba/Rubberduck] retailcoder commented on RetailCoder.VBE/UI/Command/Refactorings/RefactorExtractInterfaceCommand.cs of pull request #1351: associate project references correctly and in the correct order (#1340, #1347)

puzzlepiece87

If I am able to reproduce in the future I'll report it on github.

Mat's Mug

cool!

Hosch250

@Mat'sMug The parser can be MIT now!

The only thing that needs to be GPL is the Smart Indenter module.

Mat's Mug

20:48

@Hosch250 meh

awgaya

when can we remove the gpl notice at the top? we changed it so much lol

Hosch250

@awgaya I thought you completely rewrote it?

Duga

[rubberduck-vba/Rubberduck] build for commit d5efed8d on next: AppVeyor build failed

BUILD FAILURE!

Mat's Mug

I think the best thing to do would be to ping the original author and ask permission

Duga

[rubberduck-vba/Rubberduck] autoboosh commented on RetailCoder.VBE/UI/Command/Refactorings/RefactorExtractInterfaceCommand.cs of pull request #1351: associate project references correctly and in the correct order (#1340, #1347)

Hosch250

20:49

Oh, I thought @awgaya rewrote it from scratch.

awgaya

yes but it broke everything so now I'm trying to sneak it into next without mat's mug noticing

2

because the entire codebase uses the parser contexts

Hosch250

Oh, I thought it was merged already for some reason.

awgaya

I did send a PR yes but we decided to close it again

Hosch250

@awgaya Need help?

I have merge powers, although I suspect I wouldn't have them long if I merge it :P

2

Mat's Mug

@Hosch250 IMO it's still derivative work

Hosch250

20:50

(No, I won't merge it.)

awgaya

it's ok but thank you, this approach is probably better. It's probably too risky to do such a major change right before the 2.0 release anyway.

Hosch250

@Mat'sMug BTW, why don't you just release the code explorer as-is anyway?

It is just incomplete like the rest of it.

Mat's Mug

because as-is, it's commented-out lol

Hosch250

Well, you could un-comment it.

Mat's Mug

also, ...it just doesn't work

and it's a source of errors

and frustrations

Hosch250

20:51

Oh.

Mat's Mug

basically it's not even ready for an "alpha" release :)

Duga

[rubberduck-vba/Rubberduck] build for commit ee0fdab5 on next: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.Parsing/Symbols/ReferencedDeclarationsCollector.cs of pull request #1351: associate project references correctly and in the correct order (#1340, #1347)

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.Parsing/VBA/ReferencePriorityMap.cs of pull request #1351: associate project references correctly and in the correct order (#1340, #1347)

[rubberduck-vba/Rubberduck] autoboosh commented on Rubberduck.Parsing/VBA/ReferencePriorityMap.cs of pull request #1351: associate project references correctly and in the correct order (#1340, #1347)

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.VBEEditor/QualifiedModuleName.cs of pull request #1351: associate project references correctly and in the correct order (#1340, #1347)

[rubberduck-vba/Rubberduck] autoboosh commented on Rubberduck.VBEEditor/QualifiedModuleName.cs of pull request #1351: associate project references correctly and in the correct order (#1340, #1347)

[rubberduck-vba/Rubberduck] retailcoder commented on pull request #1351: associate project references correctly and in the correct order (#1340, #1347)

> This only prepares the ground for the resolver to be able to correctly disambiguate colliding names, correct?

[rubberduck-vba/Rubberduck] autoboosh commented on pull request #1351: associate project references correctly and in the correct order (#1340, #1347)

> Yes, exactly. The next PR will include the first expression binding process using the new project declaration.

[rubberduck-vba/Rubberduck] retailcoder merged pull request #1351: associate project references correctly and in the correct order (#1340, #1347) from autoboosh/Rubberduck/declarationhierarchy into...

rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] autoboosh pushed commit 729d09ca to next: associate project references correctly and in the correct order (#1340)

[rubberduck-vba/Rubberduck] retailcoder pushed commit e2f75e82 to next

Merge pull request #1351 from autoboosh/declarationhierarchy

associate project references correctly and in the correct order (#1340, #1347)

[rubberduck-vba/Rubberduck] build for commit e2f75e82 on next: AppVeyor build succeeded

Mat's Mug

21:16

aaand that's it, I've committed:

The countdown has started - I'll post on Rubberduck News as soon as it's up! #AlmostThere #NextReleaseRocks #VBA https://t.co/3zjc83fdAy

Tweeted by rubberduckvba on April 22, 2016 at 9:15 PM

awgaya

awesome :)

that background picture lol

Mat's Mug

21:36

@awgaya aye I need to update that!

Mitropoulos Mitropoulos

21:52

I'm getting an ever flowing stream of RuntimeBinderExceptions in Rubberduck.dll and System.Core.dll that is bringing the VBE to it's knees. Any idea why?

(trying to find and file more grammar issues ATM)

awgaya

can you post some detailed exception messages?

Mitropoulos Mitropoulos

sure! from the output window?

awgaya

yeah

some example code would be nice too if possible

Mitropoulos Mitropoulos

well this is strange, now I'm getting a parser error (nothing has changed)

gimme one second and I'll get you what you asked for asap

here's the code that caused it (new project, new class, nothing else but this code in the class)

'no logging
Option Explicit

Private vaOutputHeaders() As String
Private vaOutputColumns() As Variant
Private vaOutputFormats() As Variant
Private vaInputHeaders() As Variant
Private vaInputColumns() As Variant
Private vaInstances() As Integer

Private dOutputHeaders As New Scripting.Dictionary
Private vaOriginalInputHeadersRow As Variant
Private vaOutput As Variant

Private i As Long
Private j As Long
Private headersMapped As Integer
Private haveMapped As Boolean

here's the Output

KeyboardHook: Attached
MouseHook: Attached
Hotkey '^`' hooked successfully to command 'Rubberduck.UI.Command.ReparseCommand'
Hotkey '^P' hooked successfully to command 'Rubberduck.UI.Command.IndentCurrentProcedureCommand'
Hotkey '^M' hooked successfully to command 'Rubberduck.UI.Command.IndentCurrentModuleCommand'
Hotkey '^T' hooked successfully to command 'Rubberduck.UI.Command.FindSymbolCommand'
Hotkey '^+I' hooked successfully to command 'Rubberduck.UI.Command.InspectionResultsCommand'
Hotkey '^+T' hooked successfully to command 'Rubberduck.UI.Command.TestExplorerCommand'

Mat's Mug

22:08

Dang, when did that break?

Mitropoulos Mitropoulos

it has been doing that for the last two days, but I figured it was me (didn't want to bother you guys until I was out of ideas)

I'm trying to isolate what causes it now

awgaya

looks like it can be parsed without errors

Mitropoulos Mitropoulos

yeah sorry, the parser error was different thing

no parser error, false alarm

(here, anyway)

awgaya

np

can you break on that runtimebinderexception find out where in the code the exception is being thrown?

Mitropoulos Mitropoulos

sure thing

Mat's Mug

22:14

@awgaya has to be in Declaration.IsArray or something.. that was fixed at one point

awgaya

we might be able to remove that dynamic code if we create subclasses for the declaration class, right? Just like with project declaration

Mitropoulos Mitropoulos

paramList = (VBAParser.ArgsCallContext)proc.argsCall();

Exception thrown: 'Microsoft.CSharp.RuntimeBinder.RuntimeBinderException' in Rubberduck.dll

Additional information: 'Rubberduck.Parsing.Grammar.VBAParser.ICS_S_VariableOrProcedureCallContext' does not contain a definition for 'argsCall'

Mat's Mug

Perhaps. I just got home, will be on shortly after dinner; if you can fix it quick and PR right away, I'll merge it right away and then focus on the release build

Mitropoulos Mitropoulos

paramList is Null

Duga

[rubberduck-vba/Rubberduck] ThunderFrame opened issue #1354: Copying text from the immediate window should remove line breaks in long text

> The immediate window limits line lengths to 1023 visible characters, so any value (that has lines with more than 1023 characters) is rendered in the immediate window with line-breaks inserted by the VBE.

If you want to subsequently use that text in another context (eg SQL in SSMS, XML in Visual Studio, or Lorem Ipsum in Notepad++), then the value that is copied from the immediate window needs to be pasted into the target context and then be manually corrected. That's tedious and error prone

Mat's Mug

22:24

@Duga sounds like an X-Y problem...

Mitropoulos Mitropoulos

So I have it reduced:

Option Explicit

Private x() As String

Public Sub Add(ByVal y As String)
ReDim Preserve x(0): x(0) = y
End Sub

This causes the RuntimeBinderException for me

(although before, since there were more lines doing the same thing, it was many more of them, or sometimes just an endless stream)?

I hate to go, but I have to jet. I will look deeper into this later if you guys don't figure it out

Mat's Mug

@MitropoulosMitropoulos oh I know exactly what's causing this

No worries

ThunderFrame

@Mat'sMug given there are three examples, maybe it's an X-YYY problem...

Mat's Mug

22:40

lol

Duga

22:54

[rubberduck-vba/Rubberduck] retailcoder commented on issue #1354: Copying text from the immediate window should remove line breaks in long text

> There's no [easy] programmatic access to the immediate pane; if debug output needs to be used, then it should be output to a file, no?

awgaya

23:10

how does that work with misusing the helpfile as project id exactly?

there's kind of a mismatch between project ids added as a reference and project ids of projects we actually parse. References use the project.name and parsed projects use the project helpfile?

Duga

[rubberduck-vba/Rubberduck] ThunderFrame commented on issue #1354: Copying text from the immediate window should remove line breaks in long text

> Copying the selection to clipboard,and then modifying the clipboard contents would do it.

Writing to a file would be better, but you don't always want to do that in a debugging situation.

If there was a COM exposed object, then you could do something like:

```vb
Public Debug2 as New RubberDuck.Debug

Sub dummy
s = String(10000,"*")
Debug2.Print s
End Sub
```

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

[retailcoder/Rubberduck] retailcoder pushed commit 61e0fb60 to next

Merge pull request #1342 from retailcoder/next

Disabled Code Explorer

[retailcoder/Rubberduck] retailcoder pushed commit b62455a1 to next

Merge pull request #1344 from retailcoder/next

ObjectVariableNotSetInspection

[retailcoder/Rubberduck] autoboosh pushed commit 77f79942 to next: shortliteral + neq/geq/leq cleanup (#1345 and #1346)

[retailcoder/Rubberduck] autoboosh pushed commit 729d09ca to next: associate project references correctly and in the correct order (#1340)

[retailcoder/Rubberduck] autoboosh pushed commit 06b930c5 to next: fix paramarray incorrectly flagged as missing ByRef param mechanism (fixes #1350)

[retailcoder/Rubberduck] autoboosh pushed commit debc72e9 to next: fix failing test

[retailcoder/Rubberduck] retailcoder pushed commit d5efed8d to next

Merge pull request #1352 from autoboosh/paramarrayfix

fix paramarray incorrectly flagged as missing ByRef param mechanism (…

[retailcoder/Rubberduck] retailcoder pushed commit ee0fdab5 to next

Merge pull request #1349 from autoboosh/shortliteralfix

shortliteral + neq/geq/leq cleanup (#1345 and #1346)

[retailcoder/Rubberduck] retailcoder pushed commit e2f75e82 to next

Merge pull request #1351 from autoboosh/declarationhierarchy

associate project references correctly and in the correct order (#1340, #1347)

[retailcoder/Rubberduck] retailcoder pushed commit af22ddf5 to next

Merge pull request #71 from rubberduck-vba/next

sync with main repo

Mat's Mug

@awgaya good point. if the reference is a VBProject in the IDE, its ID should match that of the project in question, yeah. be it just for semantics. also, we're ignoring any GUID-less reference for COM "reflection" (if there's no typelib then the COM "reflection" code bombs anyway)

awgaya

is there a way to associate a vb reference with a vb project? are they connected somehow?

Mat's Mug

a referenced VBProject will have a filename (you can't reference an unsaved project), so the paths/filenames will match

and, well, the Name too

awgaya

yeah but the name doesn't have to be unique right

and currently we use the hashcode of the vbproject instead of the filename

Mat's Mug

23:24

@awgaya the name does have to be unique - VBAProject can't reference VBAProject

> Name conflicts with existing module, project, or object library

awgaya

oh I see

Mat's Mug

..and we can't use filename until project is saved

awgaya

I meant to ask you something else as well, why did you use the referenceid in the rubberduckparser's synccomreferences method?

instead of building a project id

Mat's Mug

...it didn't occur to me that the two would actually be ProjectId's!

I only realized while reviewing your PR

yeah, dat "oh" moment

so.

awgaya

ok, so that reference id doesn't happen to be the vbprojects hashcode by chance then

Mat's Mug

23:29

Declaration._isTypeSpecified was never assigned. One down.

@awgaya nope. it's ...made-up :(

awgaya

would taking the filename work in case there is one? I have no idea what the consequences would be

e.g. take the filename if there's one, otherwise take the hashcode of vbproject or something like that

well the full path rather

Mat's Mug

the reference's filename?

there's always going to be one

awgaya

yes, but the projects we parse don't have to be saved. So it would be OK to use the project's full path if there's one, otherwise fall back to the hashcode?

Mat's Mug

for ProjectId?

awgaya

yeah

Mat's Mug

23:35

File > SaveAs would blow it up, no?

awgaya

of course we could just store it separately but I'm not sure what the point of the id would be then lol

Mat's Mug

the role of ProjectId is to be able to match a VBProject instance (from the VBE) to a VBProject instance (from the parser state) - basically it's fundamental that once assigned, it never changes.

otherwise we'll have a memory leak in the parser state

awgaya

ok

we could just store the filepath separately then, but it would have to somehow trigger a refresh of the state?

a saveas operation I mean

Mat's Mug

wouldn't that be wasteful?

awgaya

no idea lol

is there a better way?

thunderframe would probably start hooking the saveas dialog by now lol

Mat's Mug

23:40

lol

awgaya

why wouldn't there be an easy way of connecting a reference and a vbproject? what about that GUID on the reference class?

Mat's Mug

ok, how about this: when we establish that a reference is pointing to a project in the VBE, we fetch that project from parser state and use its ProjectId as the ReferenceId - whatever that ProjectId is

awgaya

ah well I guess a reference doesn't have to be a vbproject -_-

Mat's Mug

when the reference is a COM reference, the ReferenceID can be the filename, or the GUID

awgaya

yeah

OK I'll give that a try, thanks

Duga

23:46

[retailcoder/Rubberduck] retailcoder pushed commit 193c66a2 to next: reinstated cached results for IsArray, HasTypeHint and IsTypeSpecified.

Mat's Mug

BindingExceptions should be under control now

somehow the fields were still there, but they were never assigned

awgaya

weird

Mat's Mug

very

setting all module states to LoadingReferences just to get the parser to report "Loading References" state was a huge mistake

awgaya

huh?

ah btw, wasn't there a trick involved in finding out whether a project is saved? it throws an IO exception

Mat's Mug

catch(IOException) I guess ;-)

awgaya

23:58

kk

Transcript for