VBA Rubberducking

Duga

12:01 AM

RELOAD!

[banane-io/PDB] 5 commits. 2 closed issues. 2 issue comments. 99 additions. 73 deletions.

[banane-io/pdb-frontend] 5 commits. 218 additions. 844 deletions.

[rubberduck-vba/Rubberduck] 1 opened issue. 6 issue comments.

[SimonForsberg/minesweeper-server] 1 commit.

[Zomis/minesweeper-flags-client] 7 commits. 1 closed issue. 1 issue comment. 297 additions. 103 deletions.

M.Doerner

12:17 AM

@this I do not see the big problem with the lists of context menu items in the IoC installer. At some point we have to state which items exactly should be injected into the parent menu.

mansellan

1:02 AM

? Application.VBE.VBProjects(2).Type
 101

:-)

Comintern

StandAlone?

this

1:27 AM

@mansellan next step: decompile that bad boy

@M.Doerner I presumed that this would incur an additional maintenance overhead in requiring to remember to update the list, as opposed to merely implementing an interface. The matter of specifying where it should be shown seems to me should be possible to do in a more declarative manner.

Comintern

@this Whatever that refactoring looks like, it should make ordering the commands (and adding the separators) easier.

this

Yeah

Comintern

A lot of the ordering is more juggling than design at this point.

this

Suddenly, the CommandBars doesn't seems that screwed up of a OM. :-D

Comintern

When in Rome...

this

1:37 AM

We could bring ribbon to VBE.

Comintern

Isn't that patented or something?

this

For the record, I was deadly unserious with the suggestion.

That said, we're using a Microsoft technology to bring a Microsoft idea to to a Microsoft tool that they don't maintain. What's the problem?

They should be seeing it as one huge favor

Comintern

^

this

(while millions of VBE developers cry bloody murder and lynch the whole RD team for it)

Comintern

TBH I don't even know if there are ribbon APIs or not these days.

this

1:40 AM

there has to be

Comintern

One would think so.

this

you can write C# addins for Office that customizes the ribbon

(doesn't have to be C#, even. I'm sure I seen that done in Delphi)

and there are standalone applications that helps with creating ribbon UIs.

What I definitely don't like about ribbon from programmability POV is that it's basically late-bound; you won't know you're doing it wrong until you run it.

Comintern

docs.microsoft.com/en-us/dotnet/api/…

@this That sounds very WPF-like.

this

wouldn't surprise me.

after all, we declare it using XML, no?

Comintern

Pretty much. I've never messed with it.

I find mucking around with the ribbon obnoxious.

this

1:44 AM

Because Access is its own thing, we can write XML in a table and boom, custom ribbon for a specific application

Comintern

Well, it stands to reason that since Excel is a database too...

this

But like you, writing XML by hand is decidedly unfun.

Comintern

It beats the hell out of most designers though.

this

That is probably why there's ribbon creator tools.

Comintern

Should I WIP the PR for #4724 before I start adding Document specific functionality to the DocumentModuleDeclaration subclass?

The diff is fairly extensive given the amount of places where checks needed to be adjusted.

this

1:55 AM

if you can break it up into smaller bite chunks, I'd do that way.

Comintern

Can't really do that given that it has resolver changes.

The bit about "the fix is trivial" was a wee bit of an understatement.

15 more failing tests first though. I apparently I jacked something up in the ShadowedDeclarationInspection...

this

don't you love rabbit holes?

Comintern

That inspection strikes me as a bit too complex in general.

this

there are few open issues around it, too

Comintern

We should already know every single shadowing instance from the name binding.

this

2:00 AM

don't we also have to consider different context?

e.g. a name could be resolved at different levels

Comintern

Right. The resolver stops when it gets an unambiguous match.

I wonder if you could simply "walk" up the scope tree.

this

we should be ... but IINM, that's not always the case?

esp. w/ UDTs or Enums

Comintern

Test for global collisions, then walk up to the project.

UDT's and Enums are always going to be special cases, no?

this

why should it be? that's just one more level for resolver to transverse.

Comintern

They're kind of pseudo-global.

this

2:06 AM

right - i was thinking that the name comparisons should be always fully qualified

Comintern

#4714 rears its head again.

this

yeah

if everyone had their own fully qualified name, the name comparison becomes much simpler

Comintern

^

this

even if the actual code doesn't use the fully qualified name form

Comintern

I'm kind of curious what else is broken by #4714 that we don't know about.

More-so about what will get broken fixing it...

this

2:10 AM

lemme see - we had a discussion about parentscope, didn't we?

Comintern

Yep. Couple days ago IIR.

this

hmm.... chat.stackexchange.com/transcript/14929?m=46213636#46213636

Comintern

FML, the ShadowedDeclarationInspectionTests are a monumental PITA to debug.

this

i originally thought there was an GH issue too but not finding it, so probably just here in chat.

Comintern

@this Interesting.

That still leads to code like this though:

    var matching = updated.FirstOrDefault(decl =>
        Declaration.DeclarationType == decl?.DeclarationType &&
        Declaration.QualifiedName.Equals(decl.QualifiedName) &&
        (Declaration.ParentDeclaration is null || Declaration.ParentDeclaration.QualifiedName.Equals(decl.ParentDeclaration?.QualifiedName)));

That should simply be Declaration.QualifiedName.Equals(decl.QualifiedName)

this

2:16 AM

^

we shouldn't be special casing name matching like that

Comintern

(it actually was before 2 bug fixes...)

What sucks is that if you build another segment into the QMN, 99% of them are going to carry around an "empty".

this

i guess sparsely populated struct isn't a thing.

OTOH, they all are strings. So that's only 4/8 bytes more.

Comintern

On the other hand, if you get rid of the "duplicate" from ParentScope...

this

that would be an improvement in the usability, I thik

declarations needs to be trimmed

Comintern

^

That's one of the reasons I want to split the classes up. The base Declaration carries some baggage the isn't always needed.

this

2:21 AM

^

and that would make the segmenting much easier

need a 5th segment? No problemo

Comintern

Yep. Gets rid of crap like this too:

this

just update the QualifiedName and the equality/hash

Comintern

public string Scope
{
    get
    {
        switch (DeclarationType)
        {
            case DeclarationType.Project:
                return "VBE";
            case DeclarationType.ClassModule:
            case DeclarationType.Document:
            case DeclarationType.ProceduralModule:
                return QualifiedModuleName.ToString();
            case DeclarationType.Procedure:
            case DeclarationType.Function:
                return $"{QualifiedModuleName}.{IdentifierName}";
            case DeclarationType.PropertyGet:

this

whee!

Comintern

You could potentially do away with DeclarationType entirely, and just use the typeof...

this

2:23 AM

and rebirth the ducky in the process, too.

Comintern

Oh wait... that's on a public COM interface.

:interfacepalm:

this

for API?

meh. break it.

we never advertised it as 100% functional anyway. I think there are other pain points around the API that I have not gotten around to addressing.

Comintern

I don't care so much about breaking it as about breaking it multiple times over a series of refactorings though.

this

^ that is a thing to worry about, esp that Max is working on them

so if you do want to break the API COM interface, you have my go for it.

it'll be a good change in the end

Comintern

It's been a long time since I looked at it.

this

2:27 AM

Another thought - it occurs to me that we could apply similar thing to the IdentifierReferences

haven't looked at code around but if we're bending over backward to find all the implemented interface/event members compared to finding all the usages....

Comintern

Not entirely sure I follow.

The TypeHierarchyPass is fairly clean IIR.

Duga

[rubberduck-vba/Rubberduck] comintern assigned comintern to issue #4724: DeclarationType is not correct for documents

this

the half-baked thought was something like EventHandlerReferences, InterfaceImplementationReferences, InvocationReferences, which would then simplify the finding of different types of references

Comintern

That wouldn't be horribly hard to implement.

It would certainly improve inspection performance.

Likely significantly.

this

because I think we do this in too many places, manually filtering the references and the QMNs for one we ant

OTOH, I really feel like we need a proper querying service

... like a .... database

Comintern

2:42 AM

Nah, that feels kind of heavy.

this

the issue is that if we have to manually tweak each inspection and the declaration finder to be performant, which could then break when we update something, that becomes burdensome.

Comintern

The other thing is that you'd be at the mercy of its selection caching and a bunch of other performance things that are easier to approach in code.

this

The real challenge, though is finding a good in-memory in-process database engine with a good optimizer.

Comintern

You mean like DeclarationFinder?

:ducks:

this

For general purpose querying, I'd rather trust a well-tuned optimizer over myself.

Heh. It is practically a database.

It's only missing an optimizer, though.

(indexing, too, I guess)

Comintern

2:44 AM

The problem though, is that tweaking performance in the inspections would mean a crapload of indexes, no?

this

right now, DF has to invalidate a lot of caches

Comintern

That's a function of the resolver though. I'm not sure how changing the "backend" solves that.

this

and rebuild them all from scratch. With indexing support, we might be able to incrementally refresh and just not worry about the queries returning stale data

No, the idea is that you invalidate only declarations that should be invalidated, and add new declarations from the resolver. The indexing service takes care of keeping the data current. No need for us to invalidate any cache.

Comintern

Wouldn't that also make the memory profile worse though?

this

Again, one'd have to find a good in-memory, in-process database with the optimizer to achieve that.

maybe, maybe not. If there's a candidate, it'll have to be benchmarked

but for now, I'll just put down the pipe.

Comintern

2:47 AM

Agreed.

Probably the best way to do that would be to wrap it in the declaration finder's interface.

this

yes that does make sense.

we do want to reduce the number of places where we write queries, too.

e.g. out from the inspections

Comintern

Hell, you could make it user settable.

Ugh. The ShadowedDeclarationInspection is borderline unmaintainable. I'm tempted to do a bottom-up re-write.

It strikes me that it's approaching the problem domain backward.

I might be able to make some slight changes to SimpleNameDefaultBinding and just handle it exactly like the resolver does.

Probably not the best idea from the SoC standpoint, but definitely tempting from a DRY standpoint...

this

again, I do think we wnat to minimize the code duplication esp in the area of querying.

the less "querying" code we have, the better & easier it will be to refactor for performance

after all, resolver + declaration finder is practically database, albeit without an indexing service or an optimizer

Comintern

So what the inspection really wants to do is call SimpleNameDefaultBinding.Resolve() while ignoring itself.

If it binds, it's a result.

this

hmm. does it have to re-resolve?

why can't we cache the information that we couldn't bind?

Comintern

3:00 AM

It should be like 10 lines of code instead of 394, and it would simply "work".

this

I think the same can be said for a numbers of inspection, actually.

Comintern

@this The problem is that it does bind. Shadowing implies that it would bind to something else if the declaration didn't exist.

this

right let me rephrase - record the number of binding attempted

Comintern

So the inspection should be, if my declaration didn't exist, would my name bind to anything.

We don't want the binder to stop short circuiting though.

The inspection is on a very limited sub-set of declarations.

Well, actually it isn't.

And you'd want the order to be different anyway.

this

considering that hte resolver is the slowest part of the parsing process (barring inspections which technically isn't a part of the parsing process), it seems preferable to cache the information for the inspection.

Comintern

3:08 AM

Except it stops when it binds.

It would basically need to find the first 2 possible bindings, then return the first one.

this

hmm

Comintern

If the second binding is not null, then it's an inspection result.

The main issue with that is that if the inspection is disabled, you'd still take the performance hit in the resolver.

this

yeah

Mathieu Guindon

4:39 AM

> Rubberduck.Setup.2.4.0.4488.exe (4.03 MiB) - downloaded 1,076 times. Last updated on 2019-01-28

M.Doerner

6:59 AM

@Comintern Two comments regarding the ShadowedDeclarationInspection: first, AFAIU, it does not try to find declarations that shadow other declarations, but all pairs of declarations where one sdhadows the other. Second, shadowing is context sensitive, so using the resolver, you would have to resolve the identifiers in all contexts in which the declarations are accessible. Using accessibility rule logic without really resolving is probably faster for this.

mansellan

11:03 AM

@Comintern ugh, that was literally what I was working on a few weeks ago, before I got bored and wandered off...

Hmm, most illuminating...

File menu gets New Project/Open Project/Save As.../Make. Some of the other menus get additions too

Interestingly, It doesn't use a .vbp project file... The Standalone projects save to a file with a .vba extension, which is binary but has bits of ini-file-style cleartext too.

Plus source code control, which doesn't load yet because I haven't installed VSS yet

Quack Exchange

1:09 PM

0

Q: Excel VBA define a range based on passed arguments

TL;DR I'm self teaching VBA, learning about functions and passing arguments from a subroutine into the function and want to know if my very basic function to define a range is optimal and follows correct referencing/standards/practices. I'm teaching myself VBA (mostly in MS Excel) as a starting ...

Hosch250

1:20 PM

@Comintern Sure can. VS has a remove unneeded fix :)

@IvenBach That your interview?

Learn about AutoFac.

And look up basic stuff around ASP.NET MVC.

FreeMan

food for @puzzlepiece87 to chew on!

Comintern

@M.Doerner Agree on the performance aspect. I had a rough idea of how to re-implement this late last night.

@mansellan Don't let me stop you - we were just discussing it. I'm down a different rabbit hole.

@Hosch250 Typing fail - that should have been "without using VS".

M.Doerner

1:36 PM

I am also down another rabbit hole with the refactorings.

Hosch250

Sorry for all my rabbit holes.

FreeMan

1:49 PM

you gotta watch out for rabbit holes. They breed like rabbits.

this

RabbitDuck?

seems better than a CatDog anyway

@Hosch250 S'ok - you had some help from me in digging some rabbit holes. Probably others, too.

Comintern

@FreeMan My favorite bit there was: "Consider this real world example a_crszkvc30LastNameCol. It took a team of maintenance engineers nearly 3 days to figure out that this whopper variable name described a const, reference, function argument that was holding information from a database column of type Varchar[30] named "LastName" which was part of the table's primary key."

Hosch250

What does the sz stand for?

Size?

Comintern

Null terminated string. Isn't HN fun?

String ... Zero

Hosch250

LOL.

this

2:01 PM

it's decidedly not standard HN, though

FreeMan

@this which simply adds to the obsfucatority of the whole thing!

Comintern

It's all over the Win32 APIs. Usually as lpsz.

FreeMan

a_ -> argument?

Comintern

> Encoding the type of a function into the name (so-called Hungarian notation) is brain damaged. - Linus Torvalds

Quack Exchange

0

Q: Need to leave out 2 parameters out of loop excel calculation

Thanks for taking the time to look into this. My code takes the contract length into account and calculates depreciation divided by contract length and also training costs divided by contract length. What I need to do is to have the calculator calculate depreciation and training costs ONLY for y...

vba excel

Comintern

2:07 PM

@FreeMan I think a_ stand for "arbitrary".

FreeMan

ah! that's where I got lost...

mansellan

@Comintern OK cool

Hosch250

2:24 PM

> For example insist on the PowerBuilder l_ and a_ {local and argument} scoping prefixes and always use the VB-esque style of having a Hungarian wart for every control type when coding to C++.

KySoto

3:14 PM

you know, its kinda funny

if you have never see anything to do with code review before, youd think that would be where you would troubleshoot your code based on the name

Mathieu Guindon

that's why you take the time to familiarize with the site, browse around, take the tour, read how-to-ask, before asking :)

of course nobody does that

this

Need a lurking engine instead of a search engine.

Duga

[rubberduck-vba/Rubberduck] fernandreu starred us

3

Mathieu Guindon

@Duga coincides with an upvote on my answer to "how to unit test vba code"

KySoto

3:31 PM

@MathieuGuindon to be fair, they dont know what they dont know

and they dont know that they dont know it

FreeMan

^ very valid points

Mathieu Guindon

fair enough - and that's why we should avoid snarks when commenting on a question we vote to close

FreeMan

unfortunately, though, even taking the minute or less to read the help ("Code Review is a question and answer site for seeking peer review of your code") doesn't make it abundantly obvious that this is for working code only. You have to delve into the Help and the "What topics can I ask about" question to find "If you have a working piece of code from your project and are looking for open-ended feedback in the following areas:"

Now, it is right there on the "Ask a question" page, in the box to the right, but that still ass-u-mes that people will take the 30 seconds to read and comprehend. It also does not direct them to SO for broken code.

Mathieu Guindon

that's /help/on-topic

FreeMan

my point. You have to read through several sections of the Help before you find that. That's more reading than most people are willing to do. Sadly.

Mathieu Guindon

3:40 PM

@FreeMan we could have "WORKING CODE ONLY" in bold flashing marquee all over the site, and we'd still get "please help me fix this" questions though

Comintern

^

FreeMan

sadly...

just gotta be nice when it comes to redirecting them to SO.

Mathieu Guindon

hey @FernAndr! welcome to our little corner of the Internet!

FreeMan

Now, my code isn't working. Anyone have any ideas why?

Mathieu Guindon

and thanks for the GH star!

FreeMan

3:45 PM

^^ Improper code reuse. :(

this

@FreeMan sorry, my crystal ball looks very... much like a crystal ball.

Mathieu Guindon

@QuackExchange answered

FreeMan

@this I was getting myself prepared to ask a VBA question on SO. I though I provided the correct amount of detail...

Comintern

4:02 PM

@QuackExchange +1 for "I'm teaching myself VBA (mostly in MS Excel)" instead of "teaching myself Excel VBA".

Mathieu Guindon

Is Rubberduck loaded? It's a useful add-in, but takes its toll in terms of memory available to the host process - is it 32-bit or 64-bit Access? — Mathieu Guindon 22 secs ago

puzzlepiece87

@FreeMan Is it a pull request or an issue, and what #?

Following the link doesn't take me to anything in particular, sorry.

FreeMan

@puzzlepiece87 you don't see this?

You were pinged with love... :)

You'll want to scroll to the top for context, though.

puzzlepiece87

4:24 PM

@FreeMan laughs and starts typing that url into his phone too imgur also blocked at work

"you don't see this?" "image not found" xD

FreeMan

sometimes I get imgur at work, sometimes not.

IvenBach

4:43 PM

@Hosch250 Yes.

KySoto

5:12 PM

-_- looks like its time to SO my problem.

the detail_format event doesnt seem to wanna let me change the visibility of a control

depending on a field value

all the stuff ive seen on hte net says in hte format event its supposed to let me do what im trying to do.

FernAndr

Thanks @MathieuGuindon! I was mostly having a quick look at the GitHub repo because I am thinking in doing some small contributions (i.e. implementing easy issues), so hopefully you will hear from me again in the next weeks!

Feeds

CommitStrip: Estimates

posted on March 06, 2019 by CommitStrip

FernAndr

Also, have you guys considered adding Rubberduck in here? github.com/MunGell/awesome-for-beginners I am surprised there is no C# projects there at all. I first started having a look at vscode but it is quite a large project to start with.

Mathieu Guindon

@FernAndr awesome! don't hesitate to come here and ask about literally anything! you'll find the initial setup instructions in the wiki

Rubberduck isn't exactly small either, but that's a very nice find!

...and wow, no C#!

hmm, many long-standing pull requests there

Hosch250

TBH, the only thing that makes RD a good first-contribution project is the current team is so helpful and it was architected pretty well.

Otherwise, you'd need a good bit of in-project and VBA/VB6 knowledge.

Mathieu Guindon

5:25 PM

depends what for

Hosch250

TBH, can't think of anything you wouldn't.

Refactorings and inspections? Need in-depth VB knowledge.

Mathieu Guindon

making a UI for regex search & replace?

Hosch250

FWIW, that should be removed. Pretty sure it would need to be redone from scratch anyway.

And the replace part would need in-project knowledge of how to use the SCP's to replace the code.

And the search the same, for reading the code.

Mathieu Guindon

meh

Hosch250

TBH, those SCPs would be fairly surprising to most programmers not experienced with COM.

Mathieu Guindon

5:29 PM

it's meta-programming, so, of course there's a bit of domain knowledge that's needed -- but at the end of the day all you need to know is "how to I get ahold of the code contents"

IvenBach

@FernAndr Welcome to the pond.

KySoto

welp, i made my question...

0

Q: How do i hide a control in a report based on a field value

I have a checkbox control that i need to hide if a Boolean value is false. I have tried using the Detail_Format event to no avail. i put the field value into another checkbox, then tried setting the visible property from the value in the checkbox, no dice. In the image below, the rightmost checkb...

vba ms-access

Hosch250

@MathieuGuindon And that's still in-project knowledge, since we don't use the VBE API directly.

Mathieu Guindon

@Hosch250 pretty sure SCP doesn't stand for "Self-Closing Pairs" here... what SCP are you talking about?

still, if you want to implement an easy inspection (there are a bunch of 'em still), you start by looking at how other inspections are made, no?

Hosch250

Safe code panes.

Mathieu Guindon

5:33 PM

oooh

SCW then - "Safe COM Wrappers" ;-)

#TooManyAcronyms

Hosch250

Oh, I knew I had the wrong name for it.

@MathieuGuindon Yeah, it's pretty easy, but you either need knowledge of how to get the declarations/references (and not all of them do this correctly ;), or knowledge of which parse tree node is what, which sends you to the grammar.

Mathieu Guindon

yeah, well, that's why we're here :)

and why pull requests are reviewed

Comintern

There's a ton of UI stuff that doesn't require any domain knowledge at all (or very minimal).

Mathieu Guindon

^

hi @MarkBalhoff!

Comintern

I.e., it would be awesome if somebody with a solid set of WPF chops would go through and consolidate the styling, etc.

Mathieu Guindon

5:39 PM

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Hosch250

Yeah, the UI is pretty much straight WPF. It's integrating the UI to the backend, though, that starts to require the internal knowledge.

Mathieu Guindon

I get the feeling someone with solid WPF chops would look at our XAML and go "ew"

*mine, at least :)

Comintern

Only if they were trying to be polite.

Hosch250

Shadow of Silicon figured out how to create resources.

Mark Balhoff

hey @MathieuGuindon been a busy couple weeks so I haven't looked in here much

Hosch250

5:41 PM

Someone could probably check his work and move a bunch of standard styling to there.

Mathieu Guindon

I know the feeling... swamped with work-work here, haven't pushed a commit in ages

Comintern

@MathieuGuindon Mine too. I shudder at the thought that I'm starting to be considered as somebody to ask about WPF things.

Mathieu Guindon

lol

Hosch250

TBH, to this day, I don't see why in the world he thought it was a good idea to use two resx strings to force one sentence of text to display on two lines :/

Mathieu Guindon

lol IKR

Hosch250

5:43 PM

Or why he rage-quit over me saying that it was unacceptable to the project standards, and would get redone, after he wouldn't change it.

Mark Balhoff

At least it's been a ecstatic frantic couple weeks as I turned down one job and accepted another offer for full time work for the first time in 3 years

4

FernAndr

I guess I am a bit biased, as I write VBA code regularly and I used rubberduck beforehand (although not very intensively). In fact, I was looking at how to write unit tests when I realised this project is on GitHub

Mathieu Guindon

curious, how did you learn about Rubberduck's existence?

@MarkBalhoff congrats!

FernAndr

I think at some point years ago I googled something like "VBA IDEs"

Mark Balhoff

I've dabbled with WPF on personal projects to learn it since I'm used to WinForms but I'm not great at it, especially dynamic generation and custom controls

Hosch250

5:45 PM

Congrats, @MarkBalhoff.

Now you won't have any free time anymore for a while :)

Mathieu Guindon

food o'clock here, bbiab

Mark Balhoff

@MathieuGuindon @Hosch250 Thanks! Yep next few months will be narrow focus. If you're talking to me, I think I just heard originally from monitoring the SO excel/access/vba tags back when I answered questions as a mental break from more monotonous work.

KySoto

6:12 PM

sometimes i hate access -_-

well... more than normal.

Mark Balhoff

lol @KySoto I get it. Access is nice for the quick and dirty front end it is with a few nice features but yes it can get annoying and limit your flexibility

KySoto

7:03 PM

@MarkBalhoff Yeah, in this particular project, ive already put in a LOT of time and effort building it, i doubt my boss would let me rebuild it in .net

plus its fairly complex just due to what they wanted it to be able to do.

not including me adding things that made sense to me to put in

KySoto

7:32 PM

Co-Worker sent me this,

apparently if you do it, it randomly deletes half of your project

Mathieu Guindon

LOL

Comintern

@KySoto When I try to run it I just keep getting an error "Thanos could not locate the JavaScript mind stone".

KySoto

lol

IvenBach

8:29 PM

Is the only difference between .NET Framework and Core that Core can run on Linux and MacOS?

Hosch250

No.

IvenBach

Is there an easy spoon-feedable summary between them?

Hosch250

That's one of the biggest changes.

ASP.NET Core has a built-in dependency injection.

Comintern

@IvenBach This isn't strictly accurate - other versions of .NET could be written in a way that was cross-compilable with Mono. AIUI what Core does is unify the target framework so that you don't have to deal with the portability concerns yourself.

mansellan

@IvenBach .Net Framework is dying, .Net Core is not.

(only partly kidding)

FreeMan

8:35 PM

@mansellan seems a relatively straightforward differentiation.

Comintern

I've heard rumors that WinForms is dying too. Been hearing those for 20 years.

Hosch250

@Comintern Part of what it does.

Mathieu Guindon

lol, VBA has been dying for 20 years too

Hosch250

And no, .NET Framework is not "dying", but they are doing their best to get people off it.

IvenBach

^^ Still has 40 more years before it's on its deathbed...

Hosch250

8:36 PM

They have explicitly said they are not killing it, but if you can get off it, to do so.

mansellan

@Hosch250 yep, was being hyperbolic. But it's not getting any more shiny toys any time soon.

Comintern

I wonder if that was because the slaying of VB6 went so well for them.

this

The new term is "mature technology "

mansellan

@Hosch250 Which would be considerably easier for us at work if they had bothered to port WCF server...

Hosch250

@mansellan Might .NET Core 3 help with that?

They are providing shims for a ton of Windows-only components.

It'll still only actually run on Windows, but you'll be able to target Core.

mansellan

8:38 PM

@Hosch250 not that I've heard, no. IIUC the sticking point is MSMQ, which few people care about anyway...

TBH we're migrating all our WCF SOAP to REST anyway, so its better in the long term. Just lots of work.

this

The mistake was that they used WCF....

Comintern

You can handle MSMQ in non-Windows environments.

mansellan

@this hey, it was once the new shiny too :-)

this

been regretting it since that one project we stupidly used

i know. Doesn't make it any less of a mistake

mansellan

^

Comintern

8:43 PM

With MS it often helps to not drink the Kool-Aid until you make sure it doesn't kill all the cultists.

2

this

23 hours ago, by this

That's "Widespread & Crushing Frustration"

mansellan

I've come to the conclusion that the best free Windows VM for me is VirtualBox. VMWare Player doesn't do snapshots, which was a dealbreaker.

this

@Comintern The hardest one is the one that kills slowly....

Comintern

ASP Classic?

this

@mansellan for vm on same machines?

mansellan

8:46 PM

@this yeah

Comintern

@mansellan That's what I run.

this

FWIW if you have a spare box, i would go for VMWare ESXi

Comintern

Make sure you install the guest extensions though.

this

its free.

mansellan

Yeah but bare metal right? Not is a positition to do that atm.

Comintern

8:48 PM

@this You mean gratis?

this

yes

Both

Been running one for more than a year much better proposition than sharing with host

FernAndr

Alright, I have finally been able to build Rubberduck. The contributing page explained it well, but I still faced an error running heat.exe (the parameter 'exePath' is invalid). I tracked this down to the fact that my solution files were inside a folder named 'C#'. Namely, the '#' was causing the issue, because it also fails with other folder names having that '#'. Is that something you were aware of? Should I create an issue?

Mathieu Guindon

@FernAndr that's a new one :)

M.Doerner

Hm, we might want to add that to the contributing page.

@this Do you know of any further restrictions heat.exe has on file names?

this

No i was not aware of that

I thought quoting would have taken care of that

@FernAndr @IvenBach. Applies to you, too?

I would create an issue. Remind me to research and see if i can make it a bit more robust

M.Doerner

8:57 PM