VBA Rubberducking

12:00 AM

RELOAD!

[rubberduck-vba/Rubberduck] 17 commits. 6 opened issues. 3 closed issues. 9 issue comments. 860 additions. 432 deletions.

Mathieu Guindon

1:15 AM

GitHub Extension for VBIDE 6.0+ ..."Fork me on GitHub", it says.

Duga

[rubberduck-vba/Rubberduck] retailcoder synchronized pull request #4082: Unify splash & about box version numbers

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4082: Unify splash & about box version numbers

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=h1) Report
> Merging [#4082](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/3489486aee1e5454ab71c938f7fb54fc29bba956?src=pr&el=desc) will **decrease** coverage by `0.01%`.
> The diff coverage is `0%`.

```diff
@@ Coverage Diff @@
## next #4082 +/- ##
===========================

Duga

1:33 AM

[rubberduck-vba/Rubberduck] build for commit 0a7e1017 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4082: Unify splash & about box version numbers

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=h1) Report
> Merging [#4082](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/3489486aee1e5454ab71c938f7fb54fc29bba956?src=pr&el=desc) will **decrease** coverage by `0.01%`.
> The diff coverage is `0%`.

```diff
@@ Coverage Diff @@
## next #4082 +/- ##
===========================

[rubberduck-vba/Rubberduck] build for commit 0a7e1017 on unknown branch: 52.62% (target 0%)

mansellan

@MathieuGuindon ~~"Requires Visual Studio 2015 or newer", it says. Not VB6 :-)~~

Mathieu Guindon

I guess we could ask the owners if a fork could be made to run in VB6/VBA

GitHub belongs to Microsoft now, so..

mansellan

Just re-read the link title. VBIDE6+, that's uh, a different thing

Mathieu Guindon

oh

I meant 6/7

mansellan

Office VBE?

Mathieu Guindon

1:40 AM

although the IDE itself is probably just 6.something

mansellan

yeah

Duga

[rubberduck-vba/Rubberduck] mansellan synchronized pull request #4111: VB6 - Add missing VBForm grammar cases

Mathieu Guindon

VBIDE 6.x

mansellan

@Duga Still gotta do the tests, but it seems to parse most of the cases when I kicked the tyres on it. Adding BARE_HEX_LITERAL to the lexer didn't work, as it would fail when the hex started A-F (picked it up as an IDENTIFIER).

Mathieu Guindon

the order matters, where was IDENTIFIER relative to your token?

in the grammar I mean :)

Duga

1:44 AM

[rubberduck-vba/Rubberduck] mansellan created comment on pull request #4111: VB6 - Add missing VBForm grammar cases

> Still gotta do the tests, but it seems to parse most of the cases when I kicked the tyres on it. Adding BARE_HEX_LITERAL to the lexer didn't work, as it would fail when the hex started A-F (picked it up as an IDENTIFIER). So I just allowed (numberLiteral | IDENTIFIER). A little too permissive, but that shouldn't matter.

The key was excluding braces from IDENTIFIER. I checked in the VBE and they are not legal, so I think this is a safe (and indeed, desirable) change.

Mathieu Guindon

thing is, it has to come after, otherwise goodbye Dim a

mansellan

@MathieuGuindon In the lexer, I had IDENTIFIER then BARE_HEX_LITERAL. It had to be that way around, otherwise identifiers that were also hex would break.

^^ yep

Sub DeadBeef() ' no es beuno

Mathieu Guindon

HEXLITERAL : '&H' [0-9A-F]+ INTEGERTYPESUFFIX?;

^ no bueno either?

mansellan

that would be awesome, except MS in their wisdom didn't put &H in front in this case...

go figure

Mathieu Guindon

oh

well then it's impossible to tell from an identifier if it starts with a letter

mansellan

1:48 AM

yep

Mathieu Guindon

without context, I mean

mansellan

hence "meh, we'll allow numbers or identifiers after the colon"

Mathieu Guindon

as a parser rule

mansellan

yeah

for this bit of the form header

Mathieu Guindon

yup

so.. why bother with BARE_HEX_LITERAL then?

mansellan

1:50 AM

I didn't in the end, was suggested on the PR but didn't work

aaaand, I forgot to remove it

gah

sync incoming

Mathieu Guindon

:+1:

Duga

[rubberduck-vba/Rubberduck] mansellan synchronized pull request #4111: VB6 - Add missing VBForm grammar cases

[rubberduck-vba/Rubberduck] build for commit bd472678 on unknown branch: AppVeyor build failed

BUILD FAILURE!

mansellan

probably a symptom of it being 02:53.... rest required.

Duga

[rubberduck-vba/Rubberduck] IvenBach opened issue #4122: Test Explorer>Group by Location duration never updates

> Duration always shows 0ms even when group by outcome shows a non-zero time. Originally a comment in #3590 I "rediscovered" this because I never reported this individually.

[rubberduck-vba/Rubberduck] IvenBach added label bug to issue #4122: Test Explorer>Group by Location duration never updates

Duga

2:04 AM

[rubberduck-vba/Rubberduck] build for commit afa628c5 on unknown branch: AppVeyor build failed

BUILD FAILURE!

Mathieu Guindon

@mansellan yikes! good night! ..whatever's left of it ;-)

Duga

[rubberduck-vba/Rubberduck] retailcoder synchronized pull request #4082: Unify splash & about box version numbers

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4082: Unify splash & about box version numbers

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=h1) Report
> Merging [#4082](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/3489486aee1e5454ab71c938f7fb54fc29bba956?src=pr&el=desc) will **decrease** coverage by `0.01%`.
> The diff coverage is `0%`.

```diff
@@ Coverage Diff @@
## next #4082 +/- ##
===========================

[rubberduck-vba/Rubberduck] retailcoder synchronized pull request #4082: Unify splash & about box version numbers

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4082: Unify splash & about box version numbers

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=h1) Report
> Merging [#4082](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/3489486aee1e5454ab71c938f7fb54fc29bba956?src=pr&el=desc) will **decrease** coverage by `0.01%`.
> The diff coverage is `0%`.

```diff
@@ Coverage Diff @@
## next #4082 +/- ##
===========================

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4082: Unify splash & about box version numbers

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=h1) Report
> Merging [#4082](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/3489486aee1e5454ab71c938f7fb54fc29bba956?src=pr&el=desc) will **decrease** coverage by `0.01%`.
> The diff coverage is `0%`.

```diff
@@ Coverage Diff @@
## next #4082 +/- ##
===========================

[rubberduck-vba/Rubberduck] build for commit a4159371 on unknown branch: AppVeyor build cancelled

BUILD FAILURE!

[rubberduck-vba/Rubberduck] retailcoder edited pull request #4082: Unify splash & about box version numbers

Mathieu Guindon

2:25 AM

v2.2.3422

Duga

[rubberduck-vba/Rubberduck] build for commit 1c46c5fe on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4082: Unify splash & about box version numbers

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=h1) Report
> Merging [#4082](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4082?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/3489486aee1e5454ab71c938f7fb54fc29bba956?src=pr&el=desc) will **increase** coverage by `<.01%`.
> The diff coverage is `11.54%`.

```diff
@@ Coverage Diff @@
## next #4082 +/- ##
=======================

[rubberduck-vba/Rubberduck] build for commit 1c46c5fe on unknown branch: 52.63% (target 0%)

[rubberduck-vba/Rubberduck] retailcoder merged pull request #4082: Unify splash & about box version numbers from retailcoder/Rubberduck/next into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] retailcoder closed issue #4052: Version in 'About' box doesn't match downloaded version #

[rubberduck-vba/Rubberduck] retailcoder pushed commit 30326e81 to next: fixes #4052

[rubberduck-vba/Rubberduck] retailcoder pushed commit 0a7e1017 to next: Merge branch 'next' of git://github.com/rubberduck-vba/Rubberduck into next

[rubberduck-vba/Rubberduck] retailcoder pushed commit a4159371 to next: inject Version into IVersionCheck service

[rubberduck-vba/Rubberduck] retailcoder pushed commit 1c46c5fe to next: Moved VersionCheck and dependencies from Main to Core.

[rubberduck-vba/Rubberduck] web-flow pushed commit e2468806 to next

Merge pull request #4082 from retailcoder/next

Unify splash & about box version numbers

Duga

2:44 AM

[rubberduck-vba/Rubberduck] retailcoder assigned retailcoder to issue #4052: Version in 'About' box doesn't match downloaded version #

[rubberduck-vba/Rubberduck] retailcoder published prerelease Rubberduck-v2.2.0.3423

[rubberduck-vba/Rubberduck] retailcoder created tag Rubberduck-v2.2.0.3423

[rubberduck-vba/Rubberduck] build for commit e2468806 on next: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit e2468806 on next: 52.63% (target 0%)

Mathieu Guindon

@this re-disposing shouldn't happen, but nothing in the IDisposable contract says that.. I'm sure I read somewhere they recommend an accidental 2nd invocation to be no-op rather than throw an exception (or worse, corrupt unmanaged state)

Mathieu Guindon

3:12 AM

well I'll be damned

windowType: Indeterminate

eventType: WinEvent.SystemMenuPopupStart, WinEvent.SystemMenuPopupEnd

idObject: 3

...which isn't a member of ObjId....yet

Mathieu Guindon

3:28 AM

hmm no, not it

idObject: 1 does it

Duga

3:44 AM

[rubberduck-vba/Rubberduck] retailcoder assigned retailcoder to issue #4087: AutoComplete interference with IntelliSense auto complete

Mathieu Guindon

whatever. I don't even need to isolate intellisense

hey @ThunderFrame

ThunderFrame

@MathieuGuindon hola

Mathieu Guindon

intellisense doesn't interfere with AC anymore :)

2

ThunderFrame

Nice. I haven't read the full history, but I gather you were trying to find the popup? Spy++ won't help, but WinDowse works.

Mathieu Guindon

Didn't need Spy++:-)

When you uncomment the firehose you see a handful of msgs per keypress, always the same. Then you hit space and suddenly a new one shows up as intellisense dropdown appears

then another when it's gone

ThunderFrame

3:54 AM

Cool. WinDowse has global keyboard shortcuts, so you don't need to change window focus, or move the mouse

And it's free for non-commercial use

Mathieu Guindon

the best part is that this was one of the most dreaded fixes to AC. rest should be smooth sailing :)

ThunderFrame

@MathieuGuindon what did you think of @IvenBach's idea for AC with foo++?

Mathieu Guindon

more autocorrect than autocomplete, but it could work

ThunderFrame

It becomes foo = foo + 1... Ditto for other operators

Mathieu Guindon

yeah

ThunderFrame

4:00 AM

I saw a tweet earlier about multicursors in Viscose snippets. Could be useful approach for the backing UDT and property thing we discussed last week.

twitter.com/johannesrieken/status/1009473137434681344?s=19

Mathieu Guindon

it's VSCode though

but wow

ThunderFrame

Oops, Viscose = VSCode

Mathieu Guindon

LOL

ThunderFrame

The syntax for defining the snippet looks useful and extensible

Mathieu Guindon

4:25 AM

the NameListWndClass aka intellisense dropdown is only created once. makes sense I guess. so ObjectCreate can't be used for fine-tuning. dang. I want the IntelliSense hwnd.

except now I know the classname is NameListWndClass, ...it's done.. building

bingo

IvenBach

hWnd is handleWindow? The pointer of the window?

Mathieu Guindon

yeah

except everything is a window in ...Windows

IvenBach

Which is why Excel has a hWnd property?

Mathieu Guindon

yup

IvenBach

msdn.microsoft.com/en-us/vba/excel-vba/articles/…

Is hwnd the parent object of other things it contains?

Mathieu Guindon

4:34 AM

it's just a pointer

IvenBach

If it's just a pointer why do you need it?

Mathieu Guindon

because it points to the window :)

IvenBach

I'm missing the obvious again.

Mathieu Guindon

13 mins ago, by IvenBach

Is hwnd the parent object of other things it contains?

hwnd isn't the parent object

it's a pointer to that parent object

IvenBach

#WordFail

Once you have the pointer to the parent object you can then dig around for what you want.

Mathieu Guindon

4:49 AM

plenty of User32 API functions take a hWnd pointer

IvenBach

I worry if I'll ever be able to adequately state what I think with programming concepts.

Mathieu Guindon

meh

ThunderFrame

6:08 AM

@IvenBach just think of hWnd as a unique ID. Want to get/set something about a window? Just pass its ID to one of a bunch of Win32 functions.

IvenBach

That's pretty much the determination I'd come to with hWnd.

Duga

[rubberduck-vba/Rubberduck] IvenBach created comment on issue #4050: PermissiveAssertClass not permissive enough for tests

> With github.com/IvenBach/Rubberduck/tree/… I've been able to get the results I've seen from VBA. I not trying to duplicate any work, but this is how I know to solve the restrictive permissiveness.

[rubberduck-vba/Rubberduck] retailcoder created pull request #4123: AutoComplete IV to merge retailcoder/Rubberduck/next into rubberduck-vba/Rubberduck/next

> Fixes most AC issues

[rubberduck-vba/Rubberduck] retailcoder added label feature-autocomplete to pull request #4123: AutoComplete IV

Mathieu Guindon

@IvenBach any more AC issues? ;-)

IvenBach

@Duga If anyone knows how to better replicate the VBA checks of ?"10" = 10 'True I'm open to suggestions.

@MathieuGuindon The comment about backspace vs del was the only one I had thought of with VS.

Mathieu Guindon

short of fancypants tooltips (#4110) and the initial wiring-up of the codepane hook (#4080), everything is working exactly as it should now AFAIK

IvenBach

6:16 AM

If I don't understand any suggestions with the PermissiveAssertClass just keep talking at me. I'll eventually connect the loose wire.

Mathieu Guindon

sure :)

but for now, sleep

<~ off to bed

IvenBach

I already mentioned foo++ becoming foo = foo + 1. Everything else will take time to code with and iron out any undiscovered issues.

Night mug.

I'm heading to bed myself.

Mathieu Guindon

yeah I think foo++ is a great idea

just doesn't fit the AC "input -> output" structure ATM

but not impossible

IvenBach

I'm incrementing counters a lot less nowadays. But when I find myself doing destinationRow = destinationRow + 1 I'd really like it.

Mathieu Guindon

might as well implement += and -= too

and /= and *=

and \=

IvenBach

6:23 AM

Don’t forget —

My brains toast. Night.

Duga

[rubberduck-vba/Rubberduck] build for commit 5d9bff53 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) created comment on pull request #4123: AutoComplete IV

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4123?src=pr&el=h1) Report
> Merging [#4123](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4123?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/e2468806b7da88f20ca28a4ea0f08edb92cacd78?src=pr&el=desc) will **decrease** coverage by `0.05%`.
> The diff coverage is `0%`.

```diff
@@ Coverage Diff @@
## next #4123 +/- ##
===========================

[rubberduck-vba/Rubberduck] build for commit 5d9bff53 on unknown branch: 52.58% (target 0%)

ThunderFrame

@MathieuGuindon it might be worth adopting a rule-based syntax like that VSCode snippet syntax

Mathieu Guindon

Yeah.. I'll sleep on it =)

Son Goku ssj4

6:52 AM

Hi folks :-)

is THIS the reason for the name? :)

imgur.com/gallery/NAulxvn

ThunderFrame

@SonGokussj4 exactly

and my avatar

M.Doerner

7:18 AM

@mansellan In my suggestion with BARE_HEX_LITERAL it was fully intentional that the rule did not get used in case the literal starts with a letter. That would be a legal identifier. That is why I suggested to use the alternative including BARE_HEX_LITERAL, numberLiteral and unrestrictedIdentifier to match the hex value.

For context, whatever the lexer matches cannot be changed anymore in the parser. So matching hex values starting with a letter as BARE_HEX_LITERAL would have caused parser errors in case there are used as identifiers.

Duga

7:37 AM

[rubberduck-vba/Rubberduck] MDoerner commented on Rubberduck.Parsing/Grammar/VBAParser.g4 of pull request #4111: VB6 - Add missing VBForm grammar cases

Duga

8:28 AM

[rubberduck-vba/Rubberduck] Vogel612 commented on Rubberduck.Core/AutoComplete/AutoCompleteBlockBase.cs of pull request #4123: AutoComplete IV

[rubberduck-vba/Rubberduck] Vogel612 commented on Rubberduck.VBEEditor/Events/VBENativeServices.cs of pull request #4123: AutoComplete IV

[rubberduck-vba/Rubberduck] Vogel612 commented on Rubberduck.Core/AutoComplete/AutoCompleteBlockBase.cs of pull request #4123: AutoComplete IV

[rubberduck-vba/Rubberduck] Vogel612 added label feature-unit-testing to issue #4122: Test Explorer>Group by Location duration never updates

[rubberduck-vba/Rubberduck] Vogel612 added label difficulty-02-ducky to issue #4122: Test Explorer>Group by Location duration never updates

Duga

9:25 AM

[rubberduck-vba/Rubberduck] SystemsModelling opened issue #4124: Documentation / examples for refactoring?

> I'm right at the start here. Take refactor/rename.
I have a control on a userform I want to rename. Say it's an optionbutton optComments and I want to call it optComment.

With the userform editor, Rubberduck > Refactor all items are greyed out.
Double click the control, now in Private Sub optComments_Click()
All refactorings still greyed out.
Double click the variable optComments to select it.
All refactorings still greyed out.

How do I rename it?

Quack Exchange

9:49 AM

0

Q: Find and replace placeholders in Word with values from Excel

I'm new at this and need assistance with shortening/streamlining the 2 pieces of code that is repeated three times each (this is a shortened version as these pieces of code is repeated many more times). Sub CreateWordDocTest() Dim wApp As Word.Application Dim wDoc As Word.Document Dim myStoryRa...

Duga

10:01 AM

[rubberduck-vba/Rubberduck] SystemsModelling edited issue #4124: Documentation / examples for refactoring?

Duga

10:12 AM

[rubberduck-vba/Rubberduck] SystemsModelling created comment on issue #4124: Documentation / examples for refactoring?

> OK, found it
Click the Ready button to reparse. In the sub now right-click and choose Rename from Rubberduck -> Refactor

BTW the option button is in a frame fraOptions so if I select the optionbutton in the form designer menu the only thing it offers to rename is the frame.

Duga

10:40 AM

[rubberduck-vba/Rubberduck] SystemsModelling opened issue #4125: Refactor: ^#M Extract nearly works

> I know this has been hidden but I wanted to try it anyway:

Sub testExtract()
Dim rng As Range, rowIndex As Long, colour As Long, colIndex As Long
Set rng = Worksheets.Add().Range("A1:Z1000")
'imagine some long procedure with 100+lines
For rowIndex = 1 To rng.Rows.Count

' this chunk should be elsewhere for reuse
' or to save duplication
colour = (rowIndex * 256 + rowIndex * 65536) Mod 16777216
colIndex = rowIndex / 2 + 1
rng.Cells(rowIndex, colIn

[rubberduck-vba/Rubberduck] SystemsModelling edited issue #4125: Refactor: ^#M Extract nearly works

Duga

10:57 AM

[rubberduck-vba/Rubberduck] SystemsModelling opened issue #4126: Refactor: Introduce Parameter might need TODO

> Start here:

Sub testIntroduceParameter()
Dim rng As Range, rowIndex As Long, colour As Long, colIndex As Long
Set rng = Worksheets.Add().Range("A1:Z1000")
For rowIndex = 1 To rng.Rows.Count
IntroduceParameter rowIndex, rng
Next rowIndex
End Sub
Sub IntroduceParameter(ByVal rowIndex As Long, ByVal rng As Range)
Dim colour As Long, colIndex As Long
colour = (rowIndex * 256 + rowIndex * 65536) Mod 16777216
colIndex = rowIndex / 2 + 1
rng.Cells

[rubberduck-vba/Rubberduck] mansellan created comment on pull request #4111: VB6 - Add missing VBForm grammar cases

> @MDoerner Excellent points on the unit tests, it's something I need to get into the habit of doing. Being reflective for a second, I think that my negative past experience with TDD has taught me some very bad habits, which I should correct. Specifically, I race to the code out of sheer impatience. I need to remember the old maxim that sometimes: "Slower is quicker."

Thanks for reminding me of this. Next sync will have proper test coverage.

Son Goku ssj4

11:21 AM

Just for info: Rubberduck Czech Translation status
545/1016 lines translated

Duga

[rubberduck-vba/Rubberduck] Vogel612 edited issue #4126: Refactor: Introduce Parameter might need TODO

[rubberduck-vba/Rubberduck] Vogel612 added label bug to issue #4126: Refactor: Introduce Parameter might need TODO

[rubberduck-vba/Rubberduck] Vogel612 added label feature-refactorings to issue #4126: Refactor: Introduce Parameter might need TODO

[rubberduck-vba/Rubberduck] Vogel612 created comment on issue #4124: Documentation / examples for refactoring?

> See #2058

[rubberduck-vba/Rubberduck] Vogel612 closed issue #4124: Documentation / examples for refactoring?

[rubberduck-vba/Rubberduck] Vogel612 added label duplicate to issue #4124: Documentation / examples for refactoring?

M.Doerner

11:44 AM

@mansellan I agree that TDD can be very tedious when still designing a feature. I have to admit that I do not really follow that approach in that situation. However, in fixing bugs it is much easier to apply: the structure is already there.

Duga

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.VBEEditor/Events/VBENativeServices.cs of pull request #4123: AutoComplete IV

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.Core/AutoComplete/AutoCompleteBlockBase.cs of pull request #4123: AutoComplete IV

Son Goku ssj4

12:07 PM

Edit: 693/1016
I'm on fire :-D
But that fire is gone. Have to go now :-)

Have a nice weekend!

Quack Exchange

12:57 PM

0

Q: VBA Latency Issues

I'm working on a large VBA project for work (around 1000+ lines of vba code) This module is called within another module where I disable everything possible like screen updating, events and setting calculation to manual to improve speed, yet this still seems to run slowly when I'm working with sh...

vba

Duga

1:49 PM

[rubberduck-vba/Rubberduck] bclothier commented on Rubberduck.Core/AutoComplete/AutoCompleteBase.cs of pull request #4123: AutoComplete IV

[rubberduck-vba/Rubberduck] bclothier commented on Rubberduck.Core/AutoComplete/AutoCompleteFunctionBlock.cs of pull request #4123: AutoComplete IV

[rubberduck-vba/Rubberduck] bclothier commented on Rubberduck.VBEEditor/Events/VBENativeServices.cs of pull request #4123: AutoComplete IV

[rubberduck-vba/Rubberduck] bclothier commented on Rubberduck.Core/AutoComplete/AutoCompleteFunctionBlock.cs of pull request #4123: AutoComplete IV

Vogel612

@QuackExchange "large (1000+ lines of code)"... boy are you in for a ride

Mathieu Guindon

lol

Duga

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.Core/AutoComplete/AutoCompleteBase.cs of pull request #4123: AutoComplete IV

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.Core/AutoComplete/AutoCompleteFunctionBlock.cs of pull request #4123: AutoComplete IV

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.VBEEditor/Events/VBENativeServices.cs of pull request #4123: AutoComplete IV

this

@Duga ah, didn't consider that we already have a ) typed, so it's implicitly "there". But I still don't get it -- don't we normally type Function foo As, not Function foo As)?

Duga

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.Core/AutoComplete/AutoCompleteFunctionBlock.cs of pull request #4123: AutoComplete IV

M.Doerner

2:04 PM

I have to think very hard to come up with any VBA program I have written that is below 1000 lines of code.

Mathieu Guindon

@this function foo() as

this

sorry, i'm confused. I understood that we are checking the expression when one has typed a )? If that's already typed, then what I'm typing right after the as?

Mathieu Guindon

if you delete the type name up to the space that follows As and hit tab, you don't want AC to go Function Foo() As As Variant

this

ah!

IvenBach

@Vogel612 For non-daily coders 1000 lines feels like a mountain on top of you.

Vogel612

2:10 PM

I know.

IvenBach

I felt that way with my own VBA code.

How blind I was then...

Duga

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.Core/AutoComplete/AutoCompleteFunctionBlock.cs of pull request #4123: AutoComplete IV

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.VBEEditor/Events/VBENativeServices.cs of pull request #4123: AutoComplete IV

Inarion

2:31 PM

Uh, I just realized that I broke my code library while refactoring. (Using RD's autofix on class modules with attributes set certainly was not my smartest idea.) Luckily, 20 min later, thanks to code versioning and diffing, I'm back to working status. Could have been worse.

Crisis averted

Mathieu Guindon

2:41 PM

@Inarion that's gotta be the single most annoying side-effect of being required to work off the code pane contents: rewriters lose member attributes

this

would it be possible to do a short term fix of providing a warning whenever non-default attributes are detected before executing rewriter?

Mathieu Guindon

yes

this

#TODO: Find a sucker to write the fix

Mathieu Guindon

lol

not quite a fix though

a fix would be figuring out a way to translate a codepanepass Interval into an attributespass Interval

Inarion

I know. And I also knew about this exact limitation. (Or I had previously known...) - It was just not present in my mind when I pressed the autofix-button. ;)

this

2:50 PM

#TODO: Find a sucker to write the fix

Mathieu Guindon

lol

@Inarion FTR I broke my own code myself multiple times because of this

Inarion

@MathieuGuindon Ha, I can easily imagine that. (How would one remember those attributes when the darn VBE doesn't show them?)

this

^

I don't know what they were thinking when they implemented that "feature"

Inarion

I really wonder what the decision process was when they added support for showing a selected few of them, like the Name attribute, but not for others...

this

"oooh, let's make it easy on the noobs and hide those super complicated attributes. it's easier that way."

Mathieu Guindon

2:55 PM

^ also "oooh, and let's make members invocable implicitly, so everyone and their dog can write VBA code, but nobody can read a line out of context and know what it's doing"

Inarion

Aside from the 5 minutes of confusion that my brain needed to connect the dots (and draw the correct conclusion), though, no damage done on my side. :)

On a completely unrelated note: It's Friday. It's 5pm. That is (in your words, @MathieuGuindon) almost beer o'clock! (Cya around!) :)

Mathieu Guindon

@Inarion I wish :) 11:02AM here

Hosch250

@SonGokussj4 You too!

IvenBach

3:55 PM

@Inarion I asked myself "What are these attributes these people are speaking of... I don't see anything." When I actually payed attention and read about needing to export the modules to see them. "Hrm... even more stuff I had no clue about and don't understand".

I now understand why programming stuff should require effort to do correctly. It's too easy to hide the layers of the lasagna. You end up with majority of users having NFI what they're doing.

commitstrip.com/en/2018/06/22/are-you-any-good-at-it

^ Is this how #ImposterSyndrome is borne?

Duck check: With moving parts of the solution into their own project do namespaces get auto updated or does that need to be manually done?

Mathieu Guindon

4:14 PM

@IvenBach kinda, yeah

"oh, you know VBA! we'd need XYZ, bet you can do that eyes closed!"

IvenBach

^ Me muttering under my breath: "Only if you want it to contain unqualified ranges, dot select with selection dot, and a smattering of other macro recorder jewels".

Hosch250

@Inarion Look up Ten Little Bottles, by Johnny Bond.

That is, only if Country doesn't make you want to commit suicide, like some people I know.

IvenBach

That checkbox was marked when I was built.

Duga

[rubberduck-vba/Rubberduck] autoboosh forked us into autoboosh/Rubberduck

3

Mathieu Guindon

@Duga was wondering when that was going to happen :)

awgaya

4:23 PM

I'll try to create a PR for the redim variable bug, OK?

Mathieu Guindon

awesome!

IvenBach

@awgaya chat.stackexchange.com/transcript/message/45289482#45289482 ran through my mind when I read this.

awgaya

hah :P I was the one who failed to implement it back then so it's only fair

Mathieu Guindon

@IvenBach lol

IvenBach

I think your hiatus started right before I started coming here.

For any issues I raise, I feel a sense of duty to fix them.

awgaya

4:28 PM

:)

Mathieu Guindon

@IvenBach speaking of issues you raised, mind testing this build for all the AC fixes? :)

IvenBach

Will I get those fixes if I pull from next?

Mathieu Guindon

the PR isn't merged, so, not yet

IvenBach

There's issues with doing an install with a dev build already though?

I need to get comfortable with testing with an installer and getting the dev build working again.

Mathieu Guindon

hmm @this could confirm, but I think RD.Deploy handles it smoothly now

damn, IsBlockCompleted is failing

Function test() As Variant
    msgsbox "While" 'all good

    While True

    Wend

    Wend

End Function

...but only when an InputToken is in a string literal. I must have forgotten to StripStringLiterals somewhere

otherwise the inline completion feels very very very much exactly where it needs to be now

except msgbox " isn't right, you get MsgBox ""| ...and that's @this comment about assuming whitespaces

Hosch250

4:36 PM

@MathieuGuindon LOL.

this

5:11 PM

@MathieuGuindon not that'd have mattered since VBA'd have pretty printed it the same way.

@MathieuGuindon @IvenBach the key is that you just need to build the deployment project before you debug. Actually, the deployment ought to be the startup project.

Mathieu Guindon

which solves itself if the startup proj... ^ that

this

Because it'll update the registry for each build, that can stomp any installed version. It's generally best to uninstall the installed version if you intend to use the installed version. Even better, keep it separate in another VM.

Mathieu Guindon

@this yeah.. but I'd want it to be MsgBox "|" regardless... but that means no longer assuming that not-same code means VBE pretty-printing added just one space, i.e. accounting for weird edge cases where the user is probably trying to deliberately break the AC

i.e. wasting critical cycles for little gain

(not that the current code is super-efficient already)

this

you said VBE will add space somewhere. Will it really add a space in the middle? I'd thunk it'd be always at the end?

Mathieu Guindon

when you type msgbox" and leave the line, you get MsgBox ""

this

5:16 PM

oooh

oomf

that sucks

Mathieu Guindon

ditto for msgbox("foo")

^ actually that was the scenario I was correcting for - extra parens getting shifted

this

actually it can get worse.... foo a,b,c => foo a, b, c

that's 3 new spaces added. :\

Mathieu Guindon

yeah ~~but that doesn't affect AC~~, crap

Duga

[rubberduck-vba/Rubberduck] BZngr created pull request #4127: Unreachable Case Inspection: follow-up to merge BZngr/Rubberduck/UnreachableCaseInsp_Comments2 into rubberduck-vba/Rubberduck/next

> This is a refactored version of the UnreachableCaseInspection. This version addresses many of the unresolved comments from the initial PR as well as some improvements/corrections: 1. Refactored tests into multiple files - IMO, the single file had become unwieldy. Also allows better identification of unit tests versus functional tests. 2. Simplified the inspections processing to a single pass - prior design essentially evaluated each RangeClause context a couple times. 3. Adds Like...

expression support. 4. Adds & operator support 5. Enhanced/Corrected Eqv and Imp operator support 6. Flags value range clauses like 'x To y' where x > y. These RangeClauses are unreachable due to invalid order...but VBE still compiles and executes without errors. 7. Strengthened the handling of RangeClauses containing variables 8. Respects Option Compare Binary/Text setting.

Mathieu Guindon

okay so I need to figure out where the caret needs to shift

what would be a good algorithm for that?

I know I've inserted "" and where that is in the "before" code

I need to find that insertion point in the "after" code

I can determine the count of InputToken in the "before" code, then get the IndexOf(InputToken) - and that would be the caret position in the "before" code

Hosch250

5:23 PM

@MathieuGuindon I think you need a diff tool.

Mathieu Guindon

no, I need to write one

Hosch250

Which, anecdotally, is overdue for inspections.

@MathieuGuindon Same thing.

Mathieu Guindon

What's the use case of that for inspections?

Hosch250

Showing the diff(s).

Preview

Mathieu Guindon

I see

thing is, a preview box that takes 200ms to render doesn't matter

Hosch250

5:25 PM

@MathieuGuindon To what?

You mean it's too slow?

Mathieu Guindon

with AC, 200ms can make the difference between "awesome" and "typing is somewhat sluggish"

Hosch250

For inspections?

No, no preview for AC.

Just for inspections. You can reuse the diff tool.

Mathieu Guindon

not if diffing allocates a ton of objects. No I'd rather have AC work out the positions

Hosch250

Is it just whitespace?

Mathieu Guindon

I know what characters I'm adding and where; all I'm missing is the instance of it in the string

Hosch250

5:28 PM

Or do you need to account for auto-complete changes too?

Compare the two strings and skip whitespace in both to map the old position to the new position?

Mathieu Guindon

@Hosch250 the problem to solve is that the code I insert is not the code that gets inserted, because VBE pretty-prints

Hosch250

@MathieuGuindon If it's just whitespace, that's easy to solve...

Mathieu Guindon

 Dosomething(42,(|))

Will be inserted as

Dosomething (42, ())|

I need to work out the instance of the character I'm inserting in the "expected string", and find the index of the same instance# in the "actual string"

this

just to check --- why can't you take the current position, then search forward for the new occurence of the typed character?

Hosch250

@MathieuGuindon Yep.

this

5:32 PM

so if the typed char is ", look forward for the " then put the cursor there?

Mathieu Guindon

@this because that wouldn't account for edge cases like msgbox (

Duga

[rubberduck-vba/Rubberduck] build for commit c24b2f4f on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) created comment on pull request #4127: Unreachable Case Inspection: follow-up

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4127?src=pr&el=h1) Report
> Merging [#4127](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4127?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/e2468806b7da88f20ca28a4ea0f08edb92cacd78?src=pr&el=desc) will **increase** coverage by `0.08%`.
> The diff coverage is `92.25%`.

```diff
@@ Coverage Diff @@
## next #4127 +/- ##
=======================

Mathieu Guindon

@this I can't assume it's going to be the next one

this

hmm.

Duga

[rubberduck-vba/Rubberduck] build for commit c24b2f4f on unknown branch: 52.71% (target 0%)

this

5:33 PM

ok, what about this....

Mathieu Guindon

I think I've figured it out

this

strip all the whitespace from both original & new code, then compare the position to put the caret?

or maybe what you have in your head is better. :)

Hosch250

@this That's exactly what I suggested.

Mathieu Guindon

3 mins ago, by Mathieu Guindon

I need to work out the instance of the character I'm inserting in the "expected string", and find the index of the same instance# in the "actual string"

Hosch250

Exactly...

this

5:35 PM

@Hosch250 #SkimFail

Mathieu Guindon

First get the n, then get the nth index of

Stripping ws is just extra work

Hosch250

Yeah.

this

I wonder if VS actually parse for AC....

Mathieu Guindon

Also the selection isn't correct when completing the astypeclause of functions/propget

it's still assuming the width of 'variant', but now the type might not always be that

that's pretty easy to work out though

IIRC that leaves 2 AC issues... that IDK how to fix

this

which are?

Mathieu Guindon

5:42 PM

injecting the quickinfo message in the msgloop

and hooking on existing code panes

this

oh, just do User32.SendMessage

why can't you loop the open Windows on the startup?

Hosch250

17

Q: What are the consequences of wishing for 'all' the gold?

A lot of fantasy stories involve someone using magic or technology to solve a problem, only to find out the spell or thing they used to solve their problem was a little too potent. Let's say a wizard invents an alchemical like spell that creates gold. It's not created from nothing though, it's ...