VBA Rubberducking

EBrown

21:00

Just as well Case "A" To "Z" is valid...

Stack Exchange

0

Q: Loop trough a column cells and compare value with a Collection in VBA

I have a Collection of 166 elements: Dim MyCollection As New Collection MyCollection.Add ("%") MyCollection.Add ("%O") MyCollection.Add ("D") MyCollection.Add ("CMS") ... I want to compare them to the value of the cells in a column. If the value of a cell is not in MyCollection, I change the b...

EBrown

@Mat'sMug Umm...so I put it in the wrong room, but INTEGERLITERAL will not suffice for the Select Case statement To blocks.

Duga

[rubberduck-vba/Rubberduck] EBrown8534 commented on issue #750: Parser error: Select Case, multi Case

> Additionally, upon further research, in VB.NET the `To` style of `Case` statements can have alphabetical values, and double values as well. I.e.:

Select Case Foo
Case "a"
Case "bolts" To "nuts"
Case Else
End Select

And:

Select Case Foo
Case Is < 15, 19.5
Case 15 To 22.5, 46
Case 22.5 To 40, 45, Is > 47
Case 40 To 47
Case Else
End Select

This means that `INTEGERLITERAL` in the `TO` syntax tree

Mat's Mug

@Hosch250 unit test the resolver

but even that wouldn't be enough

EBrown

@Mat'sMug Can you move all my messages in The 2nd Monitor starting at chat.stackexchange.com/transcript/message/23564347#23564347 here?

(Only four of them.)

Mat's Mug

21:09

← 5 messages moved from The 2nd Monitor

oops, moved 5

EBrown

THat's fine.

I'll throw some updated grammar to try to fix those problems in that issue tonight...

Provided you can do a quick VBA test of the two case statements I added in my latest comment, to verify it's allowed.

Mat's Mug

I need to document the "workflow" I do when I modify the grammar file..

it's a bit convoluted

EBrown

I believe it, especially with ANTLR.

INTEGERLITERAL : (PLUS|MINUS)? ('0'..'9')+ ( ('e' | 'E') INTEGERLITERAL)* ('#' | '&')?;
DOUBLELITERAL : (PLUS|MINUS)? ('0'..'9')* '.' ('0'..'9')+ ( ('e' | 'E') (PLUS|MINUS)? ('0'..'9')+)* ('#' | '&')?;

Those are bugged as well, it appears.

Mat's Mug

huh, how so?

EBrown

As far as INTEGERLITERAL stands, it could be 1e1e5e1.

That should validate true by that rule.

Mat's Mug

21:12

oh wow

EBrown

:)

Like I said, second pair of eyes.

Ismael Miguel

  <data name="DeclarationType_Control" xml:space="preserve">
    <value>control</value>
  </data>

Where's that one applied?

EBrown

I'll comment some updated grammar rules tonight.

Mat's Mug

@IsmaelMiguel same as the other DeclarationType_Xxxx ones - that one's a bit special though, "control" stands for a "form control" (a button, textbox, dropdown, whatever)

Ismael Miguel

That one is a kick in my nuts :/

Mat's Mug

21:14

sorry

Ismael Miguel

Lets see what I can make up here

It's not your fault

You did a nice work

Mat's Mug

Simon and Vogel had quite a bit of a hard time translating that one

thanks

Ismael Miguel

It really is hard to translate

But then again, it's different languages and this kind of difficulty will always happen when translating

Mat's Mug

@EBrown one assumption RD is making, is that the VBA code is compilable when it parses

we're making an add-in, not a compiler ;-)

@IsmaelMiguel I believe you

EBrown

@Mat'sMug Still doesn't mean the grammar cannot be thorough.

Ismael Miguel

21:16

@Mat'sMug I believe me too

Mat's Mug

@EBrown indeed

EBrown

INTEGERLITERAL : (PLUS|MINUS)? ('0'..'9')+ [ ('e' | 'E') ('0'..'9')+ ] ('#' | '&')?;
DOUBLELITERAL : (PLUS|MINUS)? ('0'..'9')* '.' ('0'..'9')+ [ ('e' | 'E') (PLUS|MINUS)? ('0'..'9')+ ] ('#' | '&')?;

Mat's Mug

still, it was better to start with a grammar that works for 80% of the code, than to write one from scratch

EBrown

Provided ANTLR can properly work with the optional ([]) blocks, that should fix those two.

Mat's Mug

optional would be (...)?

EBrown

21:17

No, that's just grouping.

Option is brackets ([]).

Ismael Miguel

@EBrown What about 1.1e2? It is an integer and that treats it as decimal

Mat's Mug

"?" was part of it...

Ismael Miguel

No

EBrown

INTEGERLITERAL : (PLUS|MINUS)? ('0'..'9')+ ( ('e' | 'E') INTEGERLITERAL)* ('#' | '&')?;
DOUBLELITERAL : (PLUS|MINUS)? ('0'..'9')* '.' ('0'..'9')+ ( ('e' | 'E') (PLUS|MINUS)? ('0'..'9')+)* ('#' | '&')?;

Ismael Miguel

It's a question, it ends with ?

EBrown

21:18

That's the original definition.

Ismael Miguel

I'm just saying an edge-case

EBrown

Notice, all I did was change ( ... INTEGERLITERAL)* to [ ... ('0'...'9')+ ] on the INTEGERLITERAL.

Ismael Miguel

I noticed

I'm just informing you

EBrown

And changed ( ... )* to [ ... ] on the DOUBLELITERAL.

@IsmaelMiguel I wasn't talking to you.

Mat's Mug

the operators work like in regex. ? is "0 or once", + is "at least once" and * is "0 or more"

Ismael Miguel

21:19

Oh

Mat's Mug

and | is "or"

anyway, TTQW

EBrown

@Mat'sMug Right, and what you had was ( ( 'e' | 'E') INTEGERLITERAL)*, which would read as "the group of 'e' or 'E' followed by another integer literal, repeated any number of times."

Mat's Mug

yup

EBrown

Which means, 1e1e5e1 would pass that test.

Mat's Mug

indeed

EBrown

21:21

Which is...not ideal...lol

So, by using [ ... ] in place of ( ... )*, it should treat that like an optional. Though, you could replace it with ( ... )? instead.

Mat's Mug

I'll be back later, will catch up on the transcript - feel free to "rubberduck" at will :)

EBrown

Alright, I won't be on much after now. I'm leaving work as well.

I don't usually get on "chat" at home.

I will try to update that issue with new grammar, though. (Hopefully valid.)

I'll setup a test-environment where I can test just the grammar updates.

Mat's Mug

LMK if you have any questions about how to build the updated grammar into Rubberduck.Parsing.dll, if you need to

Ismael Miguel

  <data name="RemoveParamsDialog_RemoveIllegalSetterLetterParameter" xml:space="preserve">
    <value>Cannot remove last parameter from setter or letter.</value>
  </data>

Was it really supposed to be "letter"?

Mat's Mug

21:38

@IsmaelMiguel there's no real term for it. It's a setter for a value-type property

Property Set Foo(value)

^^ only valid for reference types

Property Let Foo(value)

^^ value is a value type

"Let-mutator" would be appropriate too

Ismael Miguel

I'll leave it in English then

Mat's Mug

I wonder what Portuguese MSDN calls them

Ismael Miguel

Garbage, probably

Some Brazillian mashup

Or another hedious name

Mat's Mug

lol. Ok, gone/driving now

Ismael Miguel

Alright

EBrown

22:25

@Mat'sMug I'll leave that to the professionals. :)

EBrown

22:40

@Mat'sMug Should I submit an issue for the INTEGERLITERAL and DOUBLELITERAL grammar issues?

I have one prepared, just need to press "Submit new issue" if you would like.

Duga

22:57

[rubberduck-vba/Rubberduck] EBrown8534 commented on issue #750: Parser error: Select Case, multi Case

> Alright, so after reading through a lot of the grammar, I think the following solution should take care of this issue:

Replace the current `Select Case` grammar:

> selectCaseStmt :
SELECT WS CASE WS valueStmt NEWLINE+
sC_Case*
WS? END_SELECT
;

> sC_Case :
CASE WS sC_Cond WS? (':'? NEWLINE* | NEWLINE+)
(block NEWLINE+)?
;

> // ELSE first, so that it is not interpreted as a variable call
sC_Cond :

[antlr/antlr4] parrt created tag 4.5.1-1

Hosch250

New version of ANTLR?

EBrown

The link to that repo is bugged.

Mat's Mug

@EBrown sure!

EBrown

@Mat'sMug Already did before you responded...lol

@Duga hasn't picked it up yet.

I also added two other concerns in it, in case you wanted to implement them.

Duga

23:16

[rubberduck-vba/Rubberduck] EBrown8534 opened issue #753: Grammar of INTEGERLITERAL and DOUBLELITERAL is too permissive

> I noticed an issue with the following grammar when I was reading over it:

> INTEGERLITERAL : (PLUS|MINUS)? ('0'..'9')+ ( ('e' | 'E') INTEGERLITERAL)* ('#' | '&')?;
DOUBLELITERAL : (PLUS|MINUS)? ('0'..'9')* '.' ('0'..'9')+ ( ('e' | 'E') (PLUS|MINUS)? ('0'..'9')+)* ('#' | '&')?;

These allow for invalid values, such as `1e1e5e1`, and `1.5e4e3`.

Possible solution:

INTEGERLITERAL : (PLUS|MINUS)? ('0'..'9')+ ( E ('0'..'9')+ )? ('#' | '&')?;
DOUBLELITERAL : (PLUS|MINUS)?

EBrown

There's @Duga with the win.

Duga

23:31

[rubberduck-vba/Rubberduck] retailcoder commented on issue #753: Grammar of INTEGERLITERAL and DOUBLELITERAL is too permissive

> :+1:

Mat's Mug

Thanks @EBrown, that's awesome!

EBrown

@Mat'sMug I may as well help all you guys out with this stuff, after all, you guys have helped me here on CR plenty before. :)

So, if all I'm really able to do is go around and fix minor issues, then so be it. :)

So now rubberduck-vba/Rubberduck shows up under "Repositories you contribute to" on my Git homepage.

Mat's Mug

Eh, any grammar issue is a major issue that ripples across the whole app

EBrown

@Mat'sMug Well, then maybe I can help fix some "major" issues then. :)

RubberDuck

Uh oh. I see grammar snippets. =;)-

EBrown

23:44

@RubberDuck Heh, figured I'd chip in some grammar help a bit.

I'm going to run through the entire grammar file at some point (maybe this weekend) and take some notes and open some issues if I think any changes would be beneficial.

RubberDuck

Ohhhh. 87 stars!

Happy to have any help we can get @EBrown.

EBrown

@RubberDuck Glad I can offer some. :) I'm thinking I'm going to use this (and having one of my dozen+ projects on Git) to get myself back in the C# groove.

I've lost my mojo lately.

Trying to get it back.

Though, I'm probably not going to make any PR's for the grammar stuff, so you guys will need to do that. (I'll provide the fixes in issues so you can run-and-gun with them, though.)

RubberDuck

Good call on the calculation Chip. — RubberDuck 36 secs ago

It's cool @EBrown. However you want to pitch in. I find myself with a lot less time and motivation these days. Which sucks because there's a not quite done feature that needs to be finished.

Transcript for