VBA Rubberducking

IvenBach

18:01

There's enough people that know about RD at MS. At the very least they're smiling at the work we're doing. Right?

this

At least two.

M.Doerner

I'd be even more interested in them open sourcing the VBE.

Mathieu Guindon

^

this

isn't that implied with OSing the VB6?

IvenBach

That'd be the black box we'd need?

M.Doerner

18:02

However, they will certainly never do that.

Comintern

Woot! Found the bug I've been trying to find for the last 2 weeks.

2

Mathieu Guindon

@IvenBach that would probably make Wayne go frantic

IvenBach

Because his lunch'd be eaten?

M.Doerner

I would be interested in the hack they pull regarding host interaction stuff, like document modules.

Comintern

^ I'm guessing it's more than one hack.

Mathieu Guindon

18:04

lol no, because then he would get to see all the hidden doors just waiting to be slammed open

Comintern

There might be one thing that isn't a hack.

this

@IvenBach more like everyone's lunch getting eaten

IvenBach

:click: Ah....

this

@Comintern nice. get a plate and hang it on the wall!

Mathieu Guindon

step 1: fork VBE. step 2: fork VSCode. step 3: ...

Hosch250

18:06

@MathieuGuindon Remind me again why we can't just transpile VBA to C# and interact with the proper COM libraries as needed?

IvenBach

There'd no longer be hidden knowledge of "This is how you do Foo on the Bar"

Hosch250

With ThunderFrame's ways to get the data from a closed document, we could write our own editor.

Comintern

@Hosch250 I'm not entirely convinced that you couldn't do that.

Mathieu Guindon

@Hosch250 we probably could... assuming a magical clean way for .NET to play nice with COM libraries

Hosch250

Doesn't .NET play nice with COM mostly?

Vogel612

18:08

if a .NET program can't play nice with COM, then the VBE probably doesn't either

this

@Hosch250 mostly because of GC

@Vogel612 VBE is COM

Mathieu Guindon

@this VBE isn't playing nice either ;-)

MdiChildHack

this

Yes, that's on VBE, though.

and most likely doesn't have anything to do with COM.

Vogel612

hmm ... can't we use an unsafe code block in C# to sidestep the GC issue?

Mathieu Guindon

hmm

this

18:10

(I'd be surprised if COM was involved at that level. Normally you only do so if you intend it to be a component that's used by others (e.g. activex controls)

Vogel612

I have no idea how GC and unsafe work, but it feels like unsafe should modify the GC semantics

Comintern

MdiChildHack is most likely due to laziness in implementing the docking system.

this

@Vogel612 one has nothing to do with other.

Hosch250

@Vogel612 If that's the case, could we maybe obsolete VBA?

this

In fact, I'd argue that we are already writing unsafe C#, without the unsafe

6

Comintern

18:11

:-D

this

because .NET GC has absolutely no idea about the COM objects, and its RCW/CCW abstractions works okish.... until it doesn't and then it leaks something bad.

Comintern

TBH, transpiling VB6 would be much, much simpler than VBA.

this

^

VBA is bascially code + host

Mathieu Guindon

how so?

Comintern

No host interaction required.

this

18:13

how do you transpile ThisWork.WorkSheets.Add()?

Mathieu Guindon

I see "host interaction" as nothing more than "just another COM type library reference"

this

Ah, but to transpile, the target language must support COM

remember, COM is ABI.

Comintern

True, but there's some other Voodoo going on there.

Mathieu Guindon

the persistence is irrelevant to.. oh wait

yeah

Comintern

The compiled binary than got shoved back into the document would need to have all the right interfaces available in a way the host could call them

this

18:14

IDK if you even can implement the necessary ABI using javascript....

Mathieu Guindon

@Comintern not impossible, but probably requires significant input/cooperation from MS

this

certainly not using javascript itself. You'd have to write the javascript compiler to handle that and that'd be no longer "just another javascript"

FreeMan

wow... I had no idea that those VB6 fans were so... fanatical!

Comintern

Hmmm... Although, you could just drop a shim into the document... VBAInterop...

Mathieu Guindon

that would make someone happy. #shim

FreeMan

18:17

Sure there are 1000s of COBOL programmers out there, but I can't imagine that any of them are clamoring to write more COBOL. I'd guess that most of 'em are doing it because A) that's all they've ever done and/or B) it pays well

Comintern

LOL - I was envisioning the VBA project bin in, say, an .xlsm file would be a shim.

IvenBach

> Doing a thing well is often a waste of time.

That's what I envision all project managers saying to themselves.

Comintern

@FreeMan I'd bet for VB6 programmers it's heavily in favor of "A".

@IvenBach I had that argument this morning.

FreeMan

@Comintern probably, though I keep considering COBOL for B)

then I bang my head on the desk to remind myself how bad it hurts.

IvenBach

@Comintern You'll have to fight Vogel for my last condolence.

@this ABI???

this

18:21

you know what API is, right?

IvenBach

A Basic Instruction?

Vogel612

~hands Iven's last condolence to Comintern

IvenBach

API = Application Programming Interface.

Vogel612

ABI = Application Bullsh*t Interface?

this

ABI = Application Binary Interface

Vogel612

18:22

I'll show myself the door ...

KySoto

i think @Vogel612 's definition is best

this

Consider how you have to write the Declare statement in VBA.

IvenBach

So instead of Worksheets.Add you'd need 000011101011100.0111000001110?

this

To do that, you need to know a bit of C++

to set up the data types correctly, etc., etc.

e.g. map VBA's Long to C++'s int

Vogel612

actually to std::int32_t

this

18:23

BUT --- you are still doing it the C++'s way

Vogel612

because for reasons int does not have a defined size

this

so if you can't invoke the method using C++'s way (e.g. using stdcall calling convention, etc.), then you're screwed.

#TIL, Vogel (goes to show how often I code in C++ )

ABI avoids that by giving you the specification of the layout of everything in binary format.

IvenBach

So instead of just simple Dim foo as Long it'd be along the lines of Public Declare PtrSafe Foo system32.dll ... as shown in docs.microsoft.com/en-us/office/vba/language/reference/…?

this

which is why COM can be language agnostic - all it matters is that you adheres to the same binary format for calling a method, defining parameter, etc. etc.

Vogel612

> If no length modifiers are present, it's guaranteed to have a width of at least 16 bits. However, on 32/64 bit systems it is almost exclusively guaranteed to have width of at least 32 bits (see below).

IvenBach

18:26

Since I know nothing of C++ I'm still struggling to use your breadcrumbs.

this

hmm. not sure how I can explain - API and ABI are similar but differ primarily in how they make themselves available to consumers.

FreeMan

IIUC - an API specifies a Long while an ABI specifies 32-bit integer. With the API, if the calling language's definition of Long doesn't match the internal definition of Long Bad Things Happen™. With an ABI, the caller sends a 32-bit integer, no matter what it wants to call it (Long, LongLong, Fred, Whatever) and the internal code gets its 32-bits and Good Things Happen™

@this - @IvenBach ^ ??

this

yeah

FreeMan

#winning!!

Mathieu Guindon

in Discussion between miroxlav and Comintern on Stack Overflow Chat, 9 mins ago, by Mathieu Guindon

hmm, you wouldn't happen to be looking for an OSS project to contribute to? Rubberduck can use devs that know VBA and C# (it's written in C# 7.x/latest)

FreeMan

18:35

weirdly, I'd never heard of an ABI more than 10 minutes ago, and just assumed it was a typo

in Discussion between miroxlav and Comintern on Stack Overflow Chat, 6 mins ago, by Mathieu Guindon

and no Slovak translations so far ;-)

Comintern

@FreeMan I'm thinking if you pass a LongLong or a Fred and it gets truncated on the copy to the stack, Bad Things Still Happen™

IvenBach

The idea of a 32-bit integer, regardless of language, is unambiguous. Hence language agnostic.

FreeMan

But if Fred is a 32-bit integer and the ABI is expecting a Long 32-bit integer, then it won't (shouldn't) get truncated. right?

Comintern

No, but if Fred is a 64 bit double and the ABI is expecting a 32 bit long, then it would.

FreeMan

You know, like we call him Mug, but really he's Mathieu

IvenBach

18:47

@FreeMan I think that was the breadcrumb I needed.

Comintern

If Fred is a double and the ABI expects a pointer...

IvenBach

@Comintern But it it's a 16-bit would it be expanded? IE #BadSTuff = overwriting stuff it shouldn't.

FreeMan

@Comintern right. The calling language passes in a Fred because that's what a 32-bit integer is called in "Callese" while the ABI, written in ABIlene calls a 32-bit integer a George. The key is they're both talking about a 32-integer.

If I pass a 64-bit double to something that's expecting a 16-bit string, well, then that's on me for being an id10t.

Comintern

In practice, yes. It would have 32 bits allocated on the stack (and assuming that the byte order was the same and the bit layout was compatible) it would get expanded.

IvenBach

Abstractions.

Comintern

18:51

What happens if it's wider is that the stack just grabs the low (or high depending on the byte order) 32 bits and pushes that to the stack.

FreeMan

@IvenBach I think you and I are understanding on a high level, while @Comintern is delving much deeper than either of us are thinking about.

Comintern

That protects the stack from being smashed by an overflow.

@FreeMan Oh, yeah - I'm off in assembly land.

FreeMan

I've written some of that. For a VAX. Back in '89ish...

@Comintern that's the thing about the ABI - you shouldn't be worrying about the stack, right? you just match up the bit-size in your language to what bit size the ABI specifies and everything is good...

IvenBach

@FreeMan We're off in the clouds while the big ducks are on the ground. Occasionally they go into the trenches like Wayne.

Comintern

@FreeMan Correct. The ABI manages the stack interaction.

Vogel612

18:56

~heavy sigh

IvenBach

@FreeMan An accurate understanding of computers would make this a lot easier for me.

I just kind of fake it, learning along the way.

FreeMan

@IvenBach join the club

IvenBach

Feels like just us are in it. Everyone else already understands it, or they fake it better than we do.

Comintern

Instruction cycle

The instruction cycle (also known as the fetch–decode–execute cycle or the fetch-execute cycle) is the basic operational process of a computer system. It is the process by which a computer retrieves a program instruction from its memory, determines what actions the instruction describes, and then carries out those actions. This cycle is repeated continuously by a computer's central processing unit (CPU), from boot-up until the computer has shut down. In simpler CPUs the instruction cycle is executed sequentially, each instruction being processed before the next one is started. In most modern...

^ That's surprisingly good.

Vogel612

@IvenBach mostly faking it better :)

and having become pretty good at googling

IvenBach

19:04

I've gotten better at that too. #UseTheSearchNoob

FreeMan

@IvenBach lol!

IvenBach

May 11 at 15:40, by this

@IvenBach use the search, Luke.

It was said a bit nicer though. I've learned to #ReedBetweenTheLines

Duga

[rubberduck-vba/Rubberduck] daFreeMan opened issue #4587: Object variable {x} assigned without Set - Quick fix is broken

> Just installed build .4244 on Win10 Excel 2016:

Received a false positive `Object Variable '{x}' is assigned without the 'Set' keyword` on the `Application.StatusBar = ...` line, so I used the `Ignore Once` quick fix to ignore it until the underlying false-positive is resolved.

Starting code:
```VBA
Private Sub PrintDataRows(ByRef Report As Worksheet, ByVal HeaderRow As Long, ByVal clinicList As ProcessClinic, ByVal MSS As MonthScoreSettings)

Dim MS As MonthScore
Dim CI As Cl

FreeMan

^sigh...

Comintern

@Duga Help me out there. What did the quick-fix break?

Duga

19:11

[rubberduck-vba/Rubberduck] retailcoder edited issue #4587: IgnoreOnce QuickFix puts the annotation under the target

Comintern

Ahhh...

Duga

[rubberduck-vba/Rubberduck] retailcoder added label bug to issue #4587: IgnoreOnce QuickFix puts the annotation under the target

[rubberduck-vba/Rubberduck] retailcoder added label feature-inspection-quickfixes to issue #4587: IgnoreOnce QuickFix puts the annotation under the target

Comintern

@FreeMan You didn't have a stale parse did you? That looks suspiciously like the code module was dirty when it ran.

Mathieu Guindon

Hmm especially with all the green tests for the quickfix

Quickfix CanExecute should be disabled if the target module is dirty... can we do that?

IvenBach

@Comintern Yes it is.

FreeMan

19:19

I was able to replicate it after a fresh parse.

Yeah, I didn't want to have to report something like that - I tried it several times.

Mathieu Guindon

@FreeMan ignore once isn't tied to any specific inspection.. they all do that?

FreeMan

Ignore once when applied to a Consider renaming parameter inspection:

'@Ignore HungarianNotation
Private Function SetSheetTitle(ByVal generatingQuarterly As Boolean, ByVal rptDate As Date) As String

i.e. it worked correctly

Mathieu Guindon

hmm that helps narrowing it down a bit

Comintern

I wonder if the code that finds the target context is borked.

FreeMan

it's 19:30 UTC on a Friday... Everything is borked...

Comintern

19:32

And if I worked in UTC I'd be off work now.

Duga

[rubberduck-vba/Rubberduck] daFreeMan created comment on issue #4587: IgnoreOnce QuickFix puts the annotation under the target

> `Ignore once` when applied to a `Consider renaming parameter` inspection:
```VBA
'@Ignore HungarianNotation
Private Function SetSheetTitle(ByVal generatingQuarterly As Boolean, ByVal rptDate As Date) As String
```
i.e. it worked correctly

FreeMan

^for the record

Comintern

Granted, I would have been really late this morning.

FreeMan

#MeToo

Yay! The wife just called with the message I've been waiting for. TTGH!

;)

Mathieu Guindon

later!

Comintern

19:34

^

Hmmm... probably not a context problem - it issues an IdentifierReferenceInspectionResult.

...and of course, I was the last one to touch it. sigh

Mathieu Guindon

@Comintern I know the feeling :)

this

20:01

yay! I'm a id10t!

everyone else has their cache invalidation and off-by-one errors; I have my double negations!

Mathieu Guindon

clearly we're a bunch of idiots

:)

this

Rubberduck: Written by a bunch rubber-necking idiots who need to rubberduck more

Comintern

LOL

I don't have no double negations though.

this

I recall reading somewhere that English is in minority in how it treats double negatives.

Comintern

Well it certainly isn't not in the minority

this

20:10

Should I go and get that dog and rabbit for you?

Comintern

:-)

Mathieu Guindon

FWIW, I'm sorry if my self-deprecation has been rubbing off on any of you guys - please don't do that; you're all anything but idiots, don't let anyone tell you otherwise.

Comintern

Not at all - If I do something stupid, I'll call myself on it.

this

^

I already had the self-deprecating sense of humor before I came here

Comintern

Same.

I think in general it's probably better than arrogant denialism.

this

20:22

^^

mansellan

20:42

TFW you spend ages trying to work out why the webcam isn't working on your eBay laptop, then realise that it doesn't have a webcam

facepalm

IvenBach

#TIL what a tetronimo is. Suddenly Tetris has so much more meaning.

What else have I missed out on by not studying algorithms!

Vogel612

21:01

four-block-shape?

IvenBach

Yes.

Vogel612

hmm .... I just now notice that Tetris uses all rotationally invariant cojoined four-block shapes

brainfilter for stupid graph theory jargon turned itself off...

:(

IvenBach

Aug 9 '17 at 21:03, by IvenBach

I'm not dumb... Just slow.

Duga

[rubberduck-vba/Rubberduck] MicroSur starred us

5

IvenBach

@Duga One more! Then His Holy Mugness can tweet our achievements to the interweebs.

Mathieu Guindon

21:04

lol

Vogel612

@IvenBach note that interweebs and interwebs are very different things

IvenBach

Wut...

Mathieu Guindon

that race has been going on since what, 300-350 stars?

this

oh that's not a new thing?

i thought oletools used to be way ahead of us until relatively recent

Comintern

I think it's interweeds, not interweebs. CompuServe, AOL, BBS's etc. - The Interweeds.

this

21:09

I knew there was a reason for 4:20. You don't make up an appointment randomly.

Mathieu Guindon

hmm

Aug 27 at 8:49, by ThunderFrame

wow, 646 stars... RD just passed decalage2/oletools to be the 2nd-most starred repo (that returns in a search for "vba") after VBA-tools/VBA-Web... roll on 650

interesting:

Nov 13 '16 at 22:09, by Duga

[rubberduck-vba/Rubberduck] decalage2 starred us

IvenBach

:whips RD: Giddy-up. They're gaining on us.

Vogel612

~grabs popcorn to enjoy watching Iven get stomped by RD

this

I think you'll find it just mite bit difficult to ride a rubberduck.

The best it could manage is a squeak, I'm afraid.

IvenBach

@Vogel612 Even without teeth RD is more than capable of trouncing me.

Vogel612

21:14

unrelated heavy sigh...

Comintern

The irony is that by openly talking about our race with "the other project", we're promoting them in our Google results for things like DISPID.

Mathieu Guindon

The ID is in the listbox and I have pulled it. But I'm just not sure how to designate to not use quotes around the IN criteria without determining the type of the bound column. Otherwise I get a query like SELECT foo FROM bar WHERE myID IN ('1','2','3') which results in a type error. I want ...myID IN (1,2,3) but if I have an actual string, I want the function to work there as well ...myStr IN ('a','b','c') — Elizabeth Ham 2 mins ago

@this ^ any ideas?

Comintern

By the stats page, to get stars we should add a Rubberduck.WebScraping API...

Mathieu Guindon

I'd say "don't use IN and parameterize the query, but this is likely Access/ADO

Vogel612

 * !!!!!!!!!!!!!!!!!!!!! This is a dirty copy of code of the …! Instructed by […].
 * !!!!!!!!!!!!!!!!!!!!!

Mathieu Guindon

21:15

@Comintern IKR!

Vogel612

lightly anonymized to give the semblance of protecting the innocent

Mathieu Guindon

@Comintern "the repo that shall not be named" then

Comintern

LOL

IvenBach

The more I play with XML documentation in throwaway stuff the more I want it in a real code base.

Vogel612

you're highly welcome to add some to the RD codebase

IvenBach

21:19

:thinks-about-it: I'll start with my own code I modify

Comintern

@MathieuGuindon If the ListBox is bound, shouldn't you be able to determine the data type of the column it's bound to? Pretty sure you can do that with UserForms.

I'm clueless about Access forms.

Mathieu Guindon

@Comintern x-y problem, I think. she's concatenating the IDs into a WHERE clause "IN" statement

this

I see that Mat already answered.

Mathieu Guindon

Not familiar with Access/DAO, but it looks like this might help. If possible, prefer a parameterized query so that the database engine can work out the correct parameter type, instead of string-concatenating the where clause. Benefits include correctness, performance, ...and security. — Mathieu Guindon 1 min ago

this?

this

@Comintern Yes you can infer the type but not from where she's looking - that would be from a different property altogether.

Comintern

21:21

Right. Access doesn't let you create table valued functions, does it.

this

Yes, though I'd say that security is overrated in the context of Access.

No, not really. We're basically talking SQL-89.

OOOOOOOLLLLLLLLDDDDDDDD SKKKKKKOOOOOOOOLLLLLLL

Mathieu Guindon

but that xkcd can't possibly be over-linked

this

for the context it's alluding, nope.

See, the critical difference is that you're hijacking a website remotely.

you can't really hack an Access database remotely.

mansellan

@Vogel612 what's the preference with XML docs? Obviously on the external API and inspections, but what else? public members in general?

this

By time you've reached it, you already can pwn everything else.

Mathieu Guindon

21:23

lol IKR

this

e.g. you're on a remote desktop

and then there's the fact that it doesn't do multiple statements

Mathieu Guindon

still not a reason to not learn to use parameterized queries. not learning to use them causes php

this

so the worst thing that will happen out of a SQLI could do is a buggy application

Unfortunately for a dynamic IN.... That's not easily solved with parameterized queries.

Comintern

@this ORLY? That sounds like a challenge if I ever heard one.

this

You can, don't misunderstand me.

Just call a VBA function in query, pwned.

Comintern

21:25

lol

this

My point is that the typical injection attacks won't work on Access.

Comintern

True, but I'm guessing you could still do some damage.

this

Anyway, a better solution would be probably to insert into a temporary table and join that. Pesto, dynamic filtering!

Comintern

^

Mathieu Guindon

Good suggestion but I'm afraid it won't work. For example, ID 1 would be in the string ("2,6,10") but that doesn't mean 1 is one of the IDs in the list. — Elizabeth Ham 1 min ago

this

21:26

Again, cost/ROI works against it, I think.

mansellan

@this unless they're just a UI to a SQL server... then they work just fine...

Mathieu Guindon

oi, wrong level assessment there

Comintern

I was thinking more along the lines of LEFT JOIN TableValuedFunction(ListBox)

this

That presumes you're using PTQ, though.

mansellan

yeah

this

21:27

LEFT JOIN TempTableInsertedFromListBox would owrk, though.

Comintern

I don't see why it wouldn't.

Mathieu Guindon

@this that

Comintern

I wonder if you could use Access SQLI to DoS a machine.

this

hmmm. convoluted but i can see a possible way

Comintern

Like some kind of SQL fork bomb.

this

21:30

something like SELECT * INTO [C:\Temp\fooN].Foo FROM MSysObjects o1, MSysobjects o2, MSysObjects o3;, with N being incremented

hmm. need a way to create a blank file first, I think.

Comintern

Can you put arbitrary VBA code in a query somehow?

this

6 mins ago, by this

Just call a VBA function in query, pwned.

Comintern

I think you used to be able to - that's a bus sized security hole.

this

Still can.

Mathieu Guindon

Access is a bus sized security hole

this

21:32

Note that there's the sandbox mode thing, though.

so you can't just plop down a random Access file and run VBA code.

Comintern

Double decker bus sized security hole. Sideways.

Mathieu Guindon

LOL

this

might be a wee bit challenging driving it sideway.

Anyway, I think you still have to do a bit of social engineering to get them to trust the file, then all bets' off.

Mathieu Guindon

doesnt the bus driver in Harry Potter manage that?

(nah)

this

hmm. i do remember him turning a bus like crazy but actually driving it sidway?

Comintern

21:34

Remember though, we're crashing the bus, not driving it.

this

lol

Maybe there's a way to run VBA without sandbox but TBH I wouldn't know. If it's possible, then it'd be true for any Office programs.

Comintern

I would kind of doubt it. What makes Access "special" in that regard is that you can open connections remotely.

I don't think there's a way to call an Excel UDF via ADODB.

If there was, I would have done something incredibly cool and probably somewhat stupid and dangerous with it by now.

this

For vBA to run, it has to be the current file that's opened; you can't run VBA that's housed in other file

Mathieu Guindon

hmm, "show designer" doesn't work, but doesn't log any exceptions either

this

and you cannot run VBA without first enabling the contents of the file.

Hence why I think you still have to first breach the trust center somehow.

Mathieu Guindon

21:41

the typelib API is going to make a few folks nervous I think

Apr 10 '17 at 3:36, by ThunderFrame

Ohai, Unviewable+ modules

this

Huh. I thought VBA handled forward declarations.

Comintern

Because we un-obscured their "security"?

this

~~Not sure if that's the same thing.~~ I'm an id10t, ignore me - I forgot the TAPI exposes the private members

Comintern

Because we can write invisible code?

I'd be more nervous about that, personally.

Mathieu Guindon

to be fair it's just the modules and members, not the actual code, right?

this

21:44

we still can't see the actual implementation, so there's that.

Mathieu Guindon

(or is it??)

this

right

we'll know that there's a AFilthyDirtySecret method housed in MyCoolAPI module

but we won't know what it actually does.

Comintern

Based on the "hidden" code I've seen, they hide it because it's trashy.

The Adobe VBA code is horrendous.

this

But remember - you can execute any code, private or not from Immediate.

didn't they deprecate it?

(in the literal meaning of the word)

Comintern

I think they defecated it.

Mathieu Guindon

21:46

@Comintern ~~The~~ Adobe ~~VBA code~~ is horrendous. #FTFY

this

^

Comintern

I think that's a fair statement.

Mathieu Guindon

@this but we can presumably run it and break and then... what's the VBE listing

Comintern

The thing is, knowing that a FilthyDirtySecret exists is 90% of the battle.

this

^

@MathieuGuindon how would you break if you can't put a breakpoint in the FilthyDirtySecret?

it'd have to be long running that you could ctrl + break but... ?

IDK.

Comintern

21:49

I'm not entirely sure I care though. If your business model is based on hiding code in Office documents, then that is your bug, not the code itself.

Mathieu Guindon

I'd guess same question applies to UV+ protected code that pops a msgbox and then you just ctrl+break onto it

just a blank page? junk?

 's?

this

Good question. I honestly don't know.

In ACCDE, though, the source code are really gone

you'd be left with only the p-code.

in theory you can "debug" that but there's no tool that works at that level, is there?

Mathieu Guindon

> I work for a non-profit and someone has kindly donated a Mac mini (late 2012) to us. However, it has Windows 7 installed on it (why, I don't know).

bwahahaha

Comintern

Doesn't that make the documents non-portable?

this

.... Ok.

@Comintern backward, no.

forward is fine.

Comintern

21:53

What about bitness?

this

locked to the Access you made ACCDE from

therefore you must distribute both 32-bit ACCDE and 64-bit ACCDE from one same ACCDB

(and you must have both 32-bit Access and 64-bit Access in order to compile the ACCDB; neither can cross-compile)

Comintern

Lovely. It kind of sounds like a PITA.

this

yes. that's why we don't recommend making shrinkwrapped commerical application with Access.

It's fine as a internal LOB application where you have more control over the environment and that's 95%+ of our clients' usage.

Comintern

Hopefully amongst other reasons.

this

I know there has been a number of software that are Access-based and continued to be sold even today

Mathieu Guindon

21:59

ttqw

this

but they are actually more like "consultingware". Not a true shrinkwrapped application.

Comintern

Viewable++ might have to be my next major project.

this

lol

and as an extra bonus -- a p-code debugger

HHICB?

Comintern

IKR?

The problem with unviewable code like that is that it makes bypassing malware checks easier.

As evidenced by "the project that shall not be named".

this

aren't they supposed to be looking at p-code?

if the AV software only reads source code then i have to say it's... dumb.

Comintern

22:02

In theory, but a lot of them take shortcuts.

this

probably explains a lot about Symantec.

Comintern

There was one I saw with a "bad functions" list.

this

lol wut?

Comintern

Like I'm a malware author because I use Kill or have reference to an FSO.

this

i guess they didn't thunk that I could just obfuscate that code or use some wicked wizardry to make it look innocent.

Comintern

22:04

"OMG! You could delete files with that!"

Vogel612

22:15

@mansellan everything you want, I guess?

Rubberduck is sorely underdocumented

add docs wherever you want, preferrably docs that are helpful :)

this

The art is in documenting what isn't obvious from the source code. That's usually the intention or some high level architectural design.

Vogel612

@this not always. For a lot of internal API even just documenting what's in the code can help, because putting it into prose already can help understanding the code

this

my main problem is that when code changes, the documentation can become obsolete

Vogel612

that's always a problem

It's something we're going to have to start looking for in PullRequests

this

from a POV of ROI, documentation that is at high level usually serves well in long run.

But yes, I do agree that we're awfully underdocumented.

Vogel612

22:23

that's true

Hosch250

Let's start with a contributing.md

That outlines the different paths to contributing.

Like, "here's what you'll touch if you just want to work on inspections".

"Here's the code explorer".

"Here's the parser. Buyer Beware"

this

we already have that.

Hosch250

Oh?

this

github.com/rubberduck-vba/Rubberduck/blob/next/CONTRIBUTING.md

github.com/rubberduck-vba/Rubberduck/wiki/Contributing

hmm this is different from this:

so we actually have two contributing pages.

M.Doerner

22:55

I think what could use some (xml) documentation are our internal API interfaces.

Duga

23:40

[rubberduck-vba/Rubberduck] comintern created comment on issue #4393: Missing Open Designer menu

> This appears to be a binding issue (as opposed to being related to setting the object to Nothing). When I put break-points at the top of OpenDesignerCommand.EvaluateCanExecute and OpenDesignerCommand.OnExecute I can't hit either one of them no matter what I do in the UI.

IvenBach

:click: another "egghead" light-bulb moment.

Transcript for