VBA Rubberducking

Duga

00:00

RELOAD!

[Hosch250/CheckersWebsite] 3 commits. 340 additions. 78 deletions.

[rubberduck-vba/Rubberduck] 2 opened issues. 1 closed issue. 24 issue comments.

Tweeting Duck

02:12

"I've always wanted to create something new." "Do it." "What if I screw up?" "You'd make a new, screwed up thing no one ever has before!"

Tweeted by ASmallFiction on June 16, 2018 at 1:06 AM

2

ThunderFrame

05:08

@Duga @this IIUC, the members exposed by Global are not implicit instances, but rather the instances will be created automatically if you use one of the members. i.e. Having a Word project reference Excel and Access, will not create instances of either Excel or Access, unless you use one of their respective global members.

mansellan

11:58

so i was thinking, now that SC has been removed from RD (which I think was absolutely the right call), once I've finished up the to-do's in VB6 go-live I'd like to start my own GH project to implement VBE source control as a separate com add-in.

SC was always pretty orthogonal to RD's core purpose, a separate add-in makes more sense I think

I'd like to use RD code as a base though, with full accreditation

awgaya

sounds cool :) could make a different one for git, svn etc. and might even integrate it back into rubberduck as a source control provider kind of like other IDEs do it?

mansellan

yeah I have a bunch of ideas for making it pluggable for different repos. An RD plugin was my initial thought, but realistically that'll be a while off I think. However, I'll stick as closely as I can to RD's arch (for example, using Castle Windsor as IoC) to make it easy to reintegrate down the line if that's ever wanted.

mansellan

12:31

So I set up my GH org for it... github.com/orangesource-vba/orangesource

(thought the name was perfect - hope the duck doesn't get scared!)

bbl

this

13:23

@ThunderFrame That's correct. I was referring more to correctly resolving stuff like Selection.XXX -- this exists both on Word and Excel OM which means that for a Word project that references Excel, it will be never possible to use say, Selection.Copy and have it copy from Excel rather than Word. But the same code would work fine in an Access project that references Excel then Word. Those implicit references was what I was worried for the resolver.

Duga

14:08

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

2

Duga

15:18

[rubberduck-vba/Rubberduck] bclothier synchronized pull request #3956: Introducing Busy state and SuspendParser action

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #3956: Introducing Busy state and SuspendParser action

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3956?src=pr&el=h1) Report
> Merging [#3956](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3956?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/c5bb40f9981a7e759da7009d9a6d11a73aa9099b?src=pr&el=desc) will **increase** coverage by `0.04%`.
> The diff coverage is `48.84%`.

```diff
@@ Coverage Diff @@
## next #3956 +/- ##
=======================

[rubberduck-vba/Rubberduck] bclothier created comment on pull request #3956: Introducing Busy state and SuspendParser action

> A thought occurred to me --- should inspections be considered a "busy" action? Any reasons for it or against it?

Duga

15:30

[rubberduck-vba/Rubberduck] build for commit 10f40449 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #3956: Introducing Busy state and SuspendParser action

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3956?src=pr&el=h1) Report
> Merging [#3956](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3956?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/c5bb40f9981a7e759da7009d9a6d11a73aa9099b?src=pr&el=desc) will **increase** coverage by `0.02%`.
> The diff coverage is `45.24%`.

```diff
@@ Coverage Diff @@
## next #3956 +/- ##
=======================

[rubberduck-vba/Rubberduck] build for commit 10f40449 on unknown branch: 52.65% (target 0%)

Duga

15:55

[rubberduck-vba/Rubberduck] MDoerner created comment on pull request #3956: Introducing Busy state and SuspendParser action

> No, running inspection should not be considered as a busy state. They should get canceled along with the parse. There is no point to run them to completion on outdated data.

Mathieu Guindon

17:10

@mansellan I'd suggest you hook up @Duga to it =)

Duga

[rubberduck-vba/Rubberduck] retailcoder closed issue #4056: Refactor/rename of type within a class doesn't change the identifier of the type definition itself

[rubberduck-vba/Rubberduck] retailcoder merged pull request #4092: Fix rename of UDTs from MDoerner/Rubberduck/FixRenameOfUdt into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] MDoerner pushed commit beff9f00 to next: Added failing tests covering issue #4056.

[rubberduck-vba/Rubberduck] MDoerner pushed commit 585a324c to next: Made handling of UDT visibility in th grammar with all other visibility handling.

[rubberduck-vba/Rubberduck] MDoerner pushed commit 04880f53 to next: Enhanced com release in VBENativeServices.

[rubberduck-vba/Rubberduck] MDoerner pushed commit 9745875b to next: Added rename tests for UDTs with a reference in another module.

[rubberduck-vba/Rubberduck] MDoerner pushed commit 6ac12aa0 to next: Improved com wrapper release in VBENativeServices in case of exceptions.

[rubberduck-vba/Rubberduck] web-flow pushed commit 27f33e32 to next

Merge pull request #4092 from MDoerner/FixRenameOfUdt

Fix rename of UDTs

[rubberduck-vba/Rubberduck] retailcoder closed issue #4046: VB6 project events not working

[rubberduck-vba/Rubberduck] retailcoder merged pull request #4093: VB6 - Fix project and component events from mansellan/Rubberduck/4046 into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] mansellan pushed commit 877c1ee7 to next: Closes #4046

[rubberduck-vba/Rubberduck] mansellan pushed commit 4b6147c1 to next: Nits

[rubberduck-vba/Rubberduck] web-flow pushed commit 38be7373 to next

Merge pull request #4093 from mansellan/4046

VB6 - Fix project and component events

[rubberduck-vba/Rubberduck] retailcoder commented on Rubberduck.VBEditor.VBA/Rubberduck.VBEditor.VBA.csproj of pull request #4098: VB6 - Do not delete source files during attributes parse

[rubberduck-vba/Rubberduck] retailcoder published prerelease Rubberduck-v2.2.0.3397

[rubberduck-vba/Rubberduck] retailcoder created tag Rubberduck-v2.2.0.3397

[rubberduck-vba/Rubberduck] build for commit 27f33e32 on next: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit 27f33e32 on next: 52.66% (target 0%)

Duga

17:42

[rubberduck-vba/Rubberduck] retailcoder published prerelease Rubberduck-v2.2.0.3398

[rubberduck-vba/Rubberduck] retailcoder created tag Rubberduck-v2.2.0.3398

[rubberduck-vba/Rubberduck] build for commit 38be7373 on next: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit 38be7373 on next: 52.65% (target 0%)

awgaya

18:08

anyone know where to watch world cup online for free?

awgaya

18:21

nvm found a stream :P

mansellan

18:41

@MathieuGuindon Duga, AV, a readme.md... lots to do. parking it for now until vb6 is ship-shape. But I wanted to grab the org and domain names in advance.

You like the name btw? Duck a L'Orange :-)

awgaya

lol

so thats where the name is from I had no idea

mansellan

Duck with Orange ~~sauce~~ source!

Mathieu Guindon

19:02

@mansellan :+1: =)

Duga

20:32

[rubberduck-vba/Rubberduck] bclothier synchronized pull request #3956: Introducing Busy state and SuspendParser action

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #3956: Introducing Busy state and SuspendParser action

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3956?src=pr&el=h1) Report
> Merging [#3956](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3956?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/38be73730ae7a05dcca540a83c1c029fc6505fb0?src=pr&el=desc) will **increase** coverage by `0.01%`.
> The diff coverage is `45.24%`.

```diff
@@ Coverage Diff @@
## next #3956 +/- ##
=======================

[rubberduck-vba/Rubberduck] bclothier created comment on pull request #3956: Introducing Busy state and SuspendParser action

> I ended up adding a `Cancel()` so that we can ensure that inspections are cancelled. However, from what I observe, it might not have much of a effect because the throw for cancellation is performed after `GetInspectionResults`, not prior, which may be why. Additionally, the `Inspecting` is not a true state, so the appearance of busy state will not look right for that reason.

I also noticed some inspections got fired multiple times, which I thought strange as I would hope it would only fire

[rubberduck-vba/Rubberduck] build for commit 59d3a2fa on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #3956: Introducing Busy state and SuspendParser action

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3956?src=pr&el=h1) Report
> Merging [#3956](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3956?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/38be73730ae7a05dcca540a83c1c029fc6505fb0?src=pr&el=desc) will **increase** coverage by `0.03%`.
> The diff coverage is `45.67%`.

```diff
@@ Coverage Diff @@
## next #3956 +/- ##
=======================

[rubberduck-vba/Rubberduck] build for commit 59d3a2fa on unknown branch: 52.67% (target 0%)

Duga

21:15

[rubberduck-vba/Rubberduck] mansellan synchronized pull request #4098: VB6 - Do not delete source files during attributes parse

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4098: VB6 - Do not delete source files during attributes parse

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4098?src=pr&el=h1) Report
> Merging [#4098](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4098?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/c5bb40f9981a7e759da7009d9a6d11a73aa9099b?src=pr&el=desc) will **decrease** coverage by `0.01%`.
> The diff coverage is `16.33%`.

```diff
@@ Coverage Diff @@
## next #4098 +/- ##
=======================

[rubberduck-vba/Rubberduck] mansellan removed label PR-Status: Review Requested from pull request #4098: VB6 - Do not delete source files during attributes parse

[rubberduck-vba/Rubberduck] mansellan created comment on issue #3817: Rubberduck.Extension should add CommandLineSafe DWORD value

> Closing this for now, as the flag is set to 0 (not command-line-safe) in #3959.

Will open a new issue for actually implementing a command-line safe switch for VB6.

[rubberduck-vba/Rubberduck] mansellan closed issue #3817: Rubberduck.Extension should add CommandLineSafe DWORD value

[rubberduck-vba/Rubberduck] build for commit 4428f607 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4098: VB6 - Do not delete source files during attributes parse

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4098?src=pr&el=h1) Report
> Merging [#4098](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4098?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/c5bb40f9981a7e759da7009d9a6d11a73aa9099b?src=pr&el=desc) will **increase** coverage by `0.01%`.
> The diff coverage is `22.95%`.

```diff
@@ Coverage Diff @@
## next #4098 +/- ##
=======================

[rubberduck-vba/Rubberduck] build for commit 4428f607 on unknown branch: 52.63% (target 0%)

[rubberduck-vba/Rubberduck] mansellan added label enhancement to issue #4099: VB6 should offer a command line option for UI-less operation.

[rubberduck-vba/Rubberduck] mansellan added label installer to issue #4099: VB6 should offer a command line option for UI-less operation.

[rubberduck-vba/Rubberduck] mansellan opened issue #4099: VB6 should offer a command line option for UI-less operation.

> See also #3817.

Rubberduck for VB6 should offer an option to specify a command-line flag to operate in a UI-less mode, to allow RD to be used to run unit tests in automated build environments (CI). This will then allow us to set the command-line safe registry key to 1.

[rubberduck-vba/Rubberduck] mansellan added label vb6-specific to issue #4099: VB6 should offer a command line option for UI-less operation.

[rubberduck-vba/Rubberduck] mansellan added label feature-unit-testing to issue #4099: VB6 should offer a command line option for UI-less operation.

[rubberduck-vba/Rubberduck] MDoerner commented on Rubberduck.Parsing/VBA/ParseCoordinator.cs of pull request #3956: Introducing Busy state and SuspendParser action

[rubberduck-vba/Rubberduck] mansellan edited issue #4099: VB6 should offer a command line option for UI-less operation.

M.Doerner

21:41

@Duga Did we ever get to the point where we pass the cancellation token to the Inspector?

this

@M.Doerner AFAICT, it does check the cancellation token provided from the parse run. It does not have its own cancellation token.

you can see that we throw an exception but only after GetInspectionResults has been performed: github.com/rubberduck-vba/Rubberduck/blob/next/…

I am thinking there should be a token.ThrowIfCancellationRequest() prior to the GetInspectionResults, too.

M.Doerner

We also throw between all inspections: the cancellation token gets provided to the Parallel.ForEach.

this

Right. But since those will run in parallels, if they've already started but has not yet entered the GetInspectionResults, we can't cancel those?

M.Doerner

Whatever inspection reached that method cannot be canceled. The inspections have not been written with cancellation in mind.

this

Not even just before entering the GetInspectionResults?

M.Doerner

21:54

That method is the very first thing in the parallel execution.

this

Yeah.

M.Doerner

Feel free to add a throw before it.

It should not make much of a difference.

this

I just don't know if TPL will queue the tasks and that would make difference. That's why I wondered.

Alright.

@Duga I do not follow. A suspension requires a write lock to be acquired. A reparse request requires a read lock to be acquired. So how could a parse have passed the read lock check?

M.Doerner

The way I understand it is that Parallel.ForEach strives to only have one task running for each CPU.

No, the check I meant is not acquiring the read lock. It is the check for being put on the stack instead of running.

There is a slight window between this check and acquisition of the read lock.

this

but read lock must be taken prior to the check

github.com/rubberduck-vba/Rubberduck/pull/3956/…

the if statement only exists for the unlikely case of thread re-entrancy (which would violate the no recursion policy)

M.Doerner

22:00

But then you release it again before taking it another time at the place where it is relevant.

this

Are you referring to the one that used to be in the ParseInternal? I removed that.

The ParseAll never ever had a read lock taken.

M.Doerner

That place being the one right next to the Monitor.Enter.

this

Yeah that was removed.

M.Doerner

Then the suspension in not water tight again.

Exactly because of the situation I have been referring to.

this

Hmm. If I put it back, then I have to modify the ParseAll

because I wanted to keep the suspension lock close as possible to the same point where we do the Monitor.Enter/Monitor.Exit

and what I really do not like is that I can't test the ParseAll

I considered moving the timer and logger out into their own method and have both path call same ParseInternal with the run & suspension locks being taken.

Duga

22:15

[rubberduck-vba/Rubberduck] MDoerner commented on Rubberduck.Parsing/VBA/ParseCoordinator.cs of pull request #3956: Introducing Busy state and SuspendParser action

Duga

22:25

[rubberduck-vba/Rubberduck] bclothier commented on Rubberduck.Parsing/VBA/ParseCoordinator.cs of pull request #3956: Introducing Busy state and SuspendParser action

Duga

22:44

[rubberduck-vba/Rubberduck] mansellan commented on Rubberduck.VBEditor.VBA/SourceCodeHandler.cs of pull request #4098: VB6 - Do not delete source files during attributes parse

[rubberduck-vba/Rubberduck] bclothier commented on Rubberduck.VBEditor.VBA/SourceCodeHandler.cs of pull request #4098: VB6 - Do not delete source files during attributes parse

Duga

23:07

[rubberduck-vba/Rubberduck] mansellan synchronized pull request #4098: VB6 - Do not delete source files during attributes parse

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4098: VB6 - Do not delete source files during attributes parse

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4098?src=pr&el=h1) Report
> Merging [#4098](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4098?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/c5bb40f9981a7e759da7009d9a6d11a73aa9099b?src=pr&el=desc) will **increase** coverage by `0.01%`.
> The diff coverage is `22.95%`.

```diff
@@ Coverage Diff @@
## next #4098 +/- ##
=======================

Duga

23:19

[rubberduck-vba/Rubberduck] build for commit b4fee199 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4098: VB6 - Do not delete source files during attributes parse

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4098?src=pr&el=h1) Report
> Merging [#4098](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4098?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/c5bb40f9981a7e759da7009d9a6d11a73aa9099b?src=pr&el=desc) will **increase** coverage by `<.01%`.
> The diff coverage is `22.58%`.

```diff
@@ Coverage Diff @@
## next #4098 +/- ##
=======================

[rubberduck-vba/Rubberduck] build for commit b4fee199 on unknown branch: 52.63% (target 0%)

Transcript for