VBA Rubberducking

this

7:01 PM

hmm. IDK if that's really feasible.

Mathieu Guindon

debugger is off-limits AFAIK

this

^

Mathieu Guindon

until technical advancements enable otherwise

this

I could see doing keyboard shortcuts to convert a f8 into f8+shift but.... IDK about performance.

mansellan

so - I added a catch/eat (not much info available to log due to call depth), it's now collecting meta for the errant libraries, but the parser issue I was seeing before has resurfaced (expected, as the parser expects everything to be named)

fwiw, the IDL for (one of) the types is:

[
  uuid(a32b7ac3-f454-4f8b-bc2e-a73b4b40bed9),
  helpcontext(0x00030d40)
]
coclass CommonDialog {
	[default] interface ICommonDialog#i;
	[default, source] interface ICommonDialogEvents#i;
};

can't see much wrong with it, except maybe that it has a helpcontext but no helpstring - not sure if that's legal (I know little about COM...)

Son Goku ssj4

7:08 PM

@MathieuGuindon Okay, Build succeeded :-)

this

@mansellan that was from olewoo, right?

IvenBach

^ :+1:

this

cos the #i don't belong there.

mansellan

@this yes

this

ok good then that's a copy artifact

mansellan

7:09 PM

yes

Mathieu Guindon

@SonGokussj4 congratulations, you no longer need to install Rubberduck ever again - just pull the latest next and build :)

mansellan

tried googling, but either not on teh webs or my googlefu is weaksauce

Son Goku ssj4

@MathieuGuindon: So now
1) I uninstall current version
2) how do I register it so that Excel knows about it?
3) Will there be automagically my Czech translations or I have to wiggle with something

this

@mansellan and how did you conclude it's the index 4?

mansellan

from the call stack. other indexes in the same lib work. its certain indexes in certain libs.

this

7:13 PM

rephrase - you verified from the callstack that the index 4 for the typelib is the CommonDialog coclass?

mansellan

its the only coclass in that lib

this

well yes, but if you did ITypeLib::GetTypeInfo(4...) it'll return that?

mansellan

can test, but expect so, all of the other meta ties up (implemented interfaces etc)

this

my concern is whether it's really the index 4

mansellan

ok, let me test

this

7:15 PM

I ran into a similar situation w/ the funcdesc

thinking that I could use MEMBER_ID -- turn out that I had to call a different function to get the index which isn't even MEMBER_ID

just to clarify did you try stepping via the olewoo's source code to see what it is doing to get the documentation?

for the same thing?

mansellan

@this ooh no... it's c++ right?

I have 0 c++...

this

both c++ and c#

mansellan

oh cool ok

that should be enough for me to diag then :-)

IvenBach

@SonGokussj4 Any time you pull from next or a repo for RD it should take care of registration for you, at the time you build it.

Son Goku ssj4

@IvenBach So if I built with Local install installed, that was I mistake I take it. So I have to uninstall, then build again = it should be there?

this

7:23 PM

it's not that bad - it's easier to read than to write anyway

and you're basically dealing with the same API.

you do need ot have C++ enabled on your VS and use native debugger so you can enter the C++ section

mansellan

@this yeah was thinking I could probably grok if I squint a bit :-)

this

well, HHICB?

IvenBach

@SonGokussj4 You shouldn't need to worry, ever, about uninstalling and re-installing.

mansellan

ikr!

this

I mean, there's a PR on olewoo by someone who basically fumbled around in C++....

IvenBach

7:24 PM

That's all thanks to the work of @this.

mansellan

'bout time I used C++ anyway, the c# footgun isn't deadly enough :-)

this

C# only gives you a gun. C++ gives you a gun, a noose, and poison.

mansellan

lol

ok cool, will have a peek & see what I find

this

@IvenBach complicated the build process in order to simplify contribution setup....

mansellan

we only just shipped 2.2 so there's no huge rush for 2.3 right?

Son Goku ssj4

7:26 PM

Okay. So without uninstalling my Local copy and builded once, I need now to:
- Git fetch rubberduck-vba
- Git pull ?

And build again to have the newest thingie.

this

meh, we don't ship on a fixed schedule. At least not without @rubberduck around. :p (he's one who keeps Mat to it)

mansellan

lol

IvenBach

git pull upstream next is what I do to ensure I'm synchronized with the next repo.

this

hmm maybe we need to pull him back so he can nag Mat. LOL

IvenBach

I think @Vogel612 is the local git guru around here.

mansellan

7:27 PM

tbh VB6 is pretty functional now. there's bugs and annoyances, but it mostly works.

it's funny to see how many inspections pop up on the MS-provided templates :-)

Son Goku ssj4

@IvenBach Oki, little problem, I don't have upstream repo. But i think it's the rubberduck-vba one?

It did pull quite ok.

IvenBach

git-scm.com/book/en/v2 is your git bible.

Son Goku ssj4

I know git. Partly. I didn't know what to pull from where. :)

Because I!ve got 2 repos. Retailcoder and Rubberduck-vba

IvenBach

That's ought to be tidied up.

Son Goku ssj4

which should be, by custom, named upstream. That I get. Will do next time.

IvenBach

7:30 PM

IIRC any lingering Retailcoder stuff should be able to be deleted without any issues.

Son Goku ssj4

there are none.

IvenBach

Deja vu... When did I become helperful?

2

Son Goku ssj4

:D

Okay. Nearly there. Last thing. How do I enable Czech language on which I work on?

Not in UI Settings - Display language

Mathieu Guindon

retailcoder/Rubberduck is my fork...

Son Goku ssj4

I know. I had this for the translation change.

this

7:38 PM

@SonGokussj4 going forward, rubberduck-vba/rubberduck should be your primary source

Son Goku ssj4

so I suppose I can delete this remote now :-)
So how do I enable my language? Just add some line in some config file?

this

should be it, yeah.

Son Goku ssj4

Any tip where and what? (will be back in an hour. Going from work)

Duga

[rubberduck-vba/Rubberduck] IvenBach created pull request #4029: Add missing resource string to merge IvenBach/Rubberduck/FixUnitTestLabels into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] IvenBach edited pull request #4029: Add missing resource string

IvenBach

It's nice knowing how/where to fix something.

Mathieu Guindon

7:44 PM

@IvenBach github.com/IvenBach/Rubberduck/blob/…

the key was there already, with the value in all available translations; it just didn't have the correct key. the correct fix would have been to rename that key in all TestExplorer.*.resx files

github.com/IvenBach/Rubberduck/blob/…

IvenBach

I feel smrt...

Mathieu Guindon

actually the correct fix would have been to change the xaml to refer to the correct, existing key

Duga

[rubberduck-vba/Rubberduck] IvenBach synchronized pull request #4029: Add missing resource string

this

i wonder if it's possible to convert all missing keys into a compiler error...

IMO, it should be. Why ship something without a key?

Duga

[rubberduck-vba/Rubberduck] build for commit e81474ba on unknown branch: AppVeyor build succeeded

IvenBach

7:54 PM

@MathieuGuindon You are getting what you paid for. Still got a long ways to go...

Duga

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) created comment on pull request #4029: Add missing resource string

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4029?src=pr&el=h1) Report
> Merging [#4029](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4029?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/e9418e26d0b6518da56d33cc9bcb86d061d4025e?src=pr&el=desc) will **increase** coverage by `<.01%`.
> The diff coverage is `0%`.

```diff
@@ Coverage Diff @@
## next #4029 +/- ##
===========================

[rubberduck-vba/Rubberduck] build for commit e81474ba on unknown branch: 53.73% (target 0%)

Duga

8:07 PM

[rubberduck-vba/Rubberduck] build for commit aa1ad098 on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] [codecov[bot]](github.com/apps/codecov) edited comment on pull request #4029: Add missing resource string

> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4029?src=pr&el=h1) Report
> Merging [#4029](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/4029?src=pr&el=desc) into [next](https://codecov.io/gh/rubberduck-vba/Rubberduck/commit/e9418e26d0b6518da56d33cc9bcb86d061d4025e?src=pr&el=desc) will **increase** coverage by `<.01%`.
> The diff coverage is `n/a`.

```diff
@@ Coverage Diff @@
## next #4029 +/- ##
==========================

[rubberduck-vba/Rubberduck] build for commit aa1ad098 on unknown branch: 53.73% (target 0%)

[rubberduck-vba/Rubberduck] retailcoder merged pull request #4029: Add missing resource string from IvenBach/Rubberduck/FixUnitTestLabels into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] IvenBach pushed commit e81474ba to next: Add missing resource string

[rubberduck-vba/Rubberduck] IvenBach pushed commit aa1ad098 to next: Correctly preface resource string.

[rubberduck-vba/Rubberduck] web-flow pushed commit d165aa9e to next

Merge pull request #4029 from IvenBach/FixUnitTestLabels

Fixes broken resource ref in Test Explorer UI.

[rubberduck-vba/Rubberduck] retailcoder added label PR-Status: Conflicting to pull request #4021: VB6 - Exclude ResFile and RelatedDocument from parse

Mathieu Guindon

@M.Doerner did you want to review #4020? IMO it's good to go

Duga

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #4002: OnErrorResumeNext Inspection can fail while walking ParseTrees

> Alternatively, the quickfix could be made unavailable at project and module levels, "forcing" the user to review each "issue" individually at a procedure scope. If making the quickfix operate at procedure level makes our lives easier and the quickfix more reliable, I'm all for it.

[rubberduck-vba/Rubberduck] retailcoder published prerelease Rubberduck-v2.2.0.3283

[rubberduck-vba/Rubberduck] retailcoder created tag Rubberduck-v2.2.0.3283

[rubberduck-vba/Rubberduck] build for commit d165aa9e on next: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] build for commit d165aa9e on next: 53.73% (target 0%)

Son Goku ssj4

8:24 PM

@MathieuGuindon: I'm home. So will I enable Czech if I add here: github.com/rubberduck-vba/Rubberduck/blob/… lines:

<data name="Language_CZ" xml:space="preserve">
    <value>Czech</value>
</data>

Mathieu Guindon

you'll definitely need a Language_CZ key; that is what's used for populating the language dropdown in the general settings tab. IIRC the key only needs to exist in the .resx file of that language though.. right? we're fetching these strings from the specific resource file they're for are we not?

Son Goku ssj4

On the other thought... I believe it won't be inside any resx file. That was stupid.

Any hint where do I add it then? I tried to find "German" or "French" text inside repo but didn't found the right place.

Hmm.

Mathieu Guindon

there's definitely Language_EN, Language_FR, and Language_DE keys somewhere

under RubberduckUI.*.resx

this

does .cs resource files need to be added back to the project?

Mathieu Guindon

just for consistency's sake, have the key in each one

@this huh?

M.Doerner

8:30 PM

@MathieuGuindon I am rather busy atm. I skimmed through it and it looks good to me now.

Mathieu Guindon

@this I've made sure to include a .cs.resx file for every .resx I've pulled out of RubberduckUI

this

oh, nm then

Mathieu Guindon

@M.Doerner thanks!

Duga

[rubberduck-vba/Rubberduck] retailcoder closed issue #4002: OnErrorResumeNext Inspection can fail while walking ParseTrees

[rubberduck-vba/Rubberduck] retailcoder merged pull request #4020: Redesign of UnhandledOnErrorResumeNext inspection from rkapka/Rubberduck/rkapka-master into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] rkapka pushed commit d14613ee to next: Clear all collections in OnErrorResumeNext inspection's context clearing

[rubberduck-vba/Rubberduck] rkapka pushed commit eaa5205b to next: Merge remote-tracking branch 'upstream/next' into rkapka-master

[rubberduck-vba/Rubberduck] rkapka pushed commit a6f11961 to next: Fixed label number for OnErrorResumeNext inspection

[rubberduck-vba/Rubberduck] rkapka pushed commit 626552cb to next: Simplified finding body element for issue in OnErrorResumeNext inspection

[rubberduck-vba/Rubberduck] rkapka pushed commit 873ac173 to next: Redesign of UnhandledOnErrorResumeNext inspection

[rubberduck-vba/Rubberduck] rkapka pushed commit af26c307 to next: Merge remote-tracking branch 'upstream/next' into rkapka-master

[rubberduck-vba/Rubberduck] web-flow pushed commit c5c11ebe to next

Merge pull request #4020 from rkapka/rkapka-master

Redesign of UnhandledOnErrorResumeNext inspection

Son Goku ssj4

So I just need to add key to ResX manager?

Mathieu Guindon

8:34 PM

yup

also, I confirm: the key is only really needed in the language of the .resx it's for:

the idea being that if you picked "Japanese" then you could still be able to revert back to English without taking a crash course in Japanese

but creating the language key in every .resx file makes more consistent resource files :)

@SonGokussj4 I'm worried about the contents of the left pane there

at this point every resource string should be under Rubberduck.Resources.*

Dain Ironfoot III

@MathieuGuindon Isn't there an enum somewhere he needs to add a value to?

Mathieu Guindon

@DainIronfootIII digging

this

@DainIronfootIII wouldn't ti be a smell to use an enum? Shouldn't it be able to deduct what to show bsaed on resx files?

Dain Ironfoot III

Umm, no.

Mathieu Guindon

@SonGokussj4 you'll want to add the language here

Dain Ironfoot III

8:40 PM

That would be a smell because it's bloody inobvious.

Son Goku ssj4

Oh there it is. Thank you.

Mathieu Guindon

        Languages = new ObservableCollection<DisplayLanguageSetting>(
            new[]
        {
            new DisplayLanguageSetting("en-US"),
            new DisplayLanguageSetting("fr-CA"),
            new DisplayLanguageSetting("de-DE")
        });

it's hard-coded

Son Goku ssj4

I just added this: i.imgur.com/GVLw30a.png

but it didn't appear. :)

That's what I was looking for.

Dain Ironfoot III

Also, we need to store it in the settings somehow....

Mathieu Guindon

@SonGokussj4 looks right

@DainIronfootIII that's literally what adding it to the Languages collection does

Dain Ironfoot III

8:42 PM

@MathieuGuindon Right.

FreeMan

Something's not quite right in Denmark...

  Dim NPSValues As NPSSettings
  Set NPSValues = New NPSSettings

is generating a valid `Variable 'NPSValuse' is not used.

Mathieu Guindon

so far so good...

@FreeMan yes?

FreeMan

However, the fix of Remove unused declaration leaves uncompilable code.

because

Mathieu Guindon

oh

FreeMan

Set NPSValues = New NPSSettings is left behind

Duga

8:43 PM

[rubberduck-vba/Rubberduck] retailcoder published prerelease Rubberduck-v2.2.0.3284

FreeMan

how best to write that up?

Duga

[rubberduck-vba/Rubberduck] retailcoder created tag Rubberduck-v2.2.0.3284

[rubberduck-vba/Rubberduck] build for commit c5c11ebe on next: AppVeyor build succeeded

Mathieu Guindon

I'm 80% sure it's already reported

FreeMan

it's not a false positive, but it is an incomplete fix.

Son Goku ssj4

WoHooooo, Finally!!!! :D

3

FreeMan

8:44 PM

OK, I'll be sure to 80% ignore it for now. ;)

Mathieu Guindon

TBH the wording and general way unused/unassigned variable inspections work has been annoying me for a few years already

Son Goku ssj4

I can finally translate with context :-D That would be totally different and more pleasure :)

THANKS guys for the time. :)

Dain Ironfoot III

NP at all.

this

so it's Céstina, eh? (don't know how to get the accented C here). Funny how we come up with different spelling for any foreign locations.

Mathieu Guindon

@SonGokussj4 that's awesome :)

Duga

8:45 PM

[rubberduck-vba/Rubberduck] build for commit c5c11ebe on next: 53.73% (target 0%)

FreeMan

Yeah, getting 2 warnings - one for not assigned and 1 for not used is a bit of a bother & somewhat confusing until you get used to it

IvenBach

@SonGokussj4 Time to convert all the rest of the resource strings.

Mathieu Guindon

@SonGokussj4 now do note, that as new inspections and features are added, new resource strings will show up regularly - you don't want all that work to do the way of the v1.2 Japanese translations =)

what I'm saying is, please try to keep them as up-to-date as possible

Son Goku ssj4

@this "Che" from "check" -> če "sh" from "shell" -> š "ti" is like "ti" from "time" but not hard "TY" but really sweet "ti" and "na" is "na" :)

@IvenBach: Yeah. It will be much faster now. :)

Mathieu Guindon

Japanese ended up dropped for lack of maintenance; by the time it was gone, a mere fraction of the UI had a Japanese translation

(which is sad, really)

Dain Ironfoot III

8:49 PM

We had someone else offer to pick it up, but Mat never pinged them when we were "done" with 2.0.

Son Goku ssj4

@MathieuGuindon You did warn me that if I don't want to maintain it, I shouldn't even start with it. I remember. I won't be telling you "don't worry, that won't happen", but I really don't plan on that.

Mathieu Guindon

+1

Dain Ironfoot III

chat.stackexchange.com/users/191800/user97218

Mathieu Guindon

whenever a "green release" is scheduled, I update the sidebar countdown on the WordPress blog

Dain Ironfoot III

Feb 10 '16 at 6:40, by Duga

[rubberduck-vba/Rubberduck] kotaro4 forked us into kotaro4/Rubberduck

That guy.

Mathieu Guindon

8:51 PM

I don't mind missing keys for pre-release builds (I don't keep FR translations up-to-date either), but for a "green release" I strive to have a FR translation for all resource keys

> This branch is 6706 commits behind rubberduck-vba:next.

this

oh, just slightly out of date...

Mathieu Guindon

I wonder how many forks are actually active, beyond ours

Son Goku ssj4

Got that. So I should watch the blog closely. :)
Or you just send a warning message to me a week before.

this

I think there's strong correlation between actively contributing and active fork. If they aren't contributing, there's not much point to using a fork is there?

Mathieu Guindon

right

well, short of "lemme build it instead of downloading an installer"

this

8:54 PM

the only reason to pay attention to any forks would be due to slackness on the primary fork's part.

Son Goku ssj4

The last thing I need to know:
Until I've translated everything, I don't do any PR. Just often do Pulls. Right?
Or maybe Pushing to my fork.

From time to time.

this

@MathieuGuindon OBTW, I use sandpaper for toilet paper!

2

Mathieu Guindon

uh

@SonGokussj4 do push to your fork often - every time you've translated a "batch" of keys

once you're ready to PR, we'll bring it in

this

I'd say to pull from rubberduck-vba/rubberduck now and then so you can keep getting the new keys and also push to your fork.

Mathieu Guindon

^

Son Goku ssj4

8:57 PM

Nice. Finally getting the vibe of "jelly cooperation" :D

Mathieu Guindon

that said, there shouldn't be any new major architectural changes affecting resources for a while now :)

IvenBach

That was a pretty big change that was made.

Mathieu Guindon

yeah and it's not 100% completed, but already having the .resx files under Rubberduck.Resources limits the damage

FreeMan

Public Sub foo(ByVal myRange As Range)
  myRange.Value2 = "Bar"
End Sub

Technically, that's ^^^ legit.

this

@MathieuGuindon at least until Mat is so guilt-ridden over the mistake he've just made. :p

FreeMan

9:00 PM

but it generates a passed ByVal and assigned a value inspection.

I get that it's not intuitive...

Mathieu Guindon

no, it's wrong

we seem to have a problem reliably identifying the target of a member assignment

FreeMan

ok, what's wrong? the code or the inspection?

Mathieu Guindon

the inspection

FreeMan

whew!

:)

Mathieu Guindon

Set myRange = ActiveCell would have been a ByVal assignment in this case

one problem is that every inspection that needs it is implementing its own version of "what's being assigned in this expression" ...that can only end up badly

FreeMan

9:03 PM

looks like a new issue? a quick search for "passed byval" doesn't seem to turn up any fruit

maybe because a turnip is a vegetable

IvenBach

Reference to Howl's Moving Castle?

this

fwiw, we definitely shouldn't release 2.3 until we fix the crash on exit thingee

FreeMan

me? no...

this

it's really annoying...

FreeMan

OK, lovely inconsistency...

Mathieu Guindon

9:05 PM

@this haven't crashed on exit for well over a year here

FreeMan

my MCVE does not generate the error, however, my real code does...

@MathieuGuindon no crash on exit for me, either, just a ghost ACCESS.EXE if I'm using that...

Mathieu Guindon

not even that

this

@FreeMan sorry I should be more precise - that's what I meant.

2.1 didn't have that; 2.2 does.

Mathieu Guindon

2.2 unloads and reloads from add-ins manager just fine (albeit with an obvious, massive, managed memory leak)

this

yep

I need to confirm but I think 2.2 is causing more system resources exceeded errors, too

Mathieu Guindon

9:08 PM

that could conceivably be behind the lingering host process in Access

Dain Ironfoot III

@this I only wish I could see the mortician's face.

this

@DainIronfootIII That'd be only possible if it was literal. (God forbid that anyone literally do it.....)

Duga

[rubberduck-vba/Rubberduck] daFreeMan opened issue #4030: Inconsistent false positive on 'passed ByVal and assigned a value' inspection

> Win10, Excel2016 (desktop), RD .3280

```vba
Public Sub foo(ByVal myRange As Range)
myRange.Value = "Bar"
End Sub
```

Generates no inspection. However, my real code (snippet):

```VBA
Public Sub SetSummaryCell(ByVal SummaryRange As Range, ByVal value As Variant, ByVal FormatFlag As String)

SummaryRange.value = CStr(value)
'other stuff
End Sub
```

generates the inspection.

Dain Ironfoot III

@this Is that a dare?

FreeMan

that's enough for one day... Do a bit of clean up at home, get my fork up to date and time to learn me some C#

TTFN

IvenBach

9:10 PM

^ Attaboy.

Dain Ironfoot III

Me, I'm going to go home and dive into a SignalR PR.

Then play a game or two of Heroes of the Storm.

IvenBach

I'll be back to the same once gma is 6ft under, and life gets back to normalcy.

this

@DainIronfootIII Uhhhh.... That's why one shouldn't do dares....

Dain Ironfoot III

@this :)

IvenBach

@DainIronfootIII I can be your medic.

Dain Ironfoot III

9:13 PM

@IvenBach You play too?

YEAH!

Join us here to talk about it and home-made games and other stuff, if you like: chat.stackexchange.com/rooms/16134/…

Mathieu Guindon

TTGH

Dain Ironfoot III

9:42 PM

TTQW.

Duga

10:27 PM

[rubberduck-vba/Rubberduck] MDoerner created comment on issue #4030: Inconsistent false positive on 'passed ByVal and assigned a value' inspection

> This is not inconsistent, but still wrong. In the code snippet, for some reason the inspection thinks that the member value refers to the parameter value, which gets passed by value. Since the potential MCVE does not have the parameter value and it does not reproduce the issue.

[rubberduck-vba/Rubberduck] MDoerner edited comment on issue #4030: Inconsistent false positive on 'passed ByVal and assigned a value' inspection

> This is not inconsistent, but still wrong. In the code snippet, for some reason the inspection thinks that the member value refers to the parameter value, which gets passed by value. Since the potential MCVE does not have the parameter value it does not reproduce the issue.

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #4030: Inconsistent false positive on 'passed ByVal and assigned a value' inspection

> Hmm, so a resolver bug more than an inspection issue then?

IvenBach

@Duga That's actually a good catch.

Duga

11:42 PM

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #4030: Inconsistent false positive on 'passed ByVal and assigned a value' inspection

> Could also be an artifact of how the older inspections kind of tried to do the resolver's job - this looks like it could be a case of the inspection not using the resolved member call; investigation is needed.

IvenBach

Home Time.

Mathieu Guindon

11:55 PM

Damn, I'll be landing in Orlando in a thunderstorm

Transcript for