VBA Rubberducking

Duga

00:02

RELOAD!

[Cardshifter/Cardshifter] 2 opened issues. 8 issue comments.

[Cardshifter/HTML-Client] 2 commits. 8 issue comments. 14 additions. 6 deletions.

[Hosch250/ResxEditor] 2 commits. 42 additions. 13 deletions.

[Hosch250/VSDiagnostics] 23 commits. 2110 additions. 902 deletions.

[retailcoder/Rubberduck] 1 commit. 820 additions. 472 deletions.

[rubberduck-vba/Rubberduck] 1 opened issue. 3 issue comments.

[skiwi2/Fragments] 2 commits. 11 additions. 3 deletions.

[Vannevelj/VSDiagnostics] 16 commits. 9 opened issues. 7 closed issues. 3 issue comments. 1167 additions. 860 deletions.

[Vannevelj/VSDiagnostics] 1 commit. 7 additions. 8 deletions.

Stack Exchange

00:28

Our pleasure to support OSS! Maintainers of active non-commercial OS projects, apply here: https://www.jetbrains.com/devnet/sponsorship/open-source/ https://twitter.com/chadmyers/status/634108324636524548

Tweeted by resharper on August 19, 2015 at 10:56 PM

Duga

[rubberduck-vba/Rubberduck] ckuhn203 commented on pull request #746: Logging support

> Yeah. Looks good to me, but I'd like @retailcoder to give it a once over. He's the one who installed NLog to begin with.

Duga

01:18

[rubberduck-vba/Rubberduck] retailcoder commented on pull request #746: Logging support

> @zippy1981 If you know how to add the event ID, by all means go ahead and do it - I'd have to do some research, as I never logged to the EventLog... but I love the idea, thanks for this PR!

Mat's Mug

05:36

coming along nicely:

Rossco

05:46

@Mat'sMug What on earth are you doing online at this time?

Mat's Mug

WPF

Rossco

isn't some crazy hour of the morning where you are?

mmmm WPF....

Mat's Mug

1:47AM

yeah, getting ~~late~~ early

Rossco

dedication or stupidity...???

Mat's Mug

both

Rossco

05:48

Well, your screenshot does look purrrtty.

Mat's Mug

better

Rossco

the menu is a little bit in the wrong place.

Mat's Mug

eh, it's a designer view

Rossco

or is that just the designer?

Mat's Mug

run-time will look much uglier than that!

Rossco

05:50

snap

@Mat'sMug excellent.

Mat's Mug

the bottom part is a tree view

I need to add buttons to control how tests are regrouped

and the captions are hard-coded into the XAML for now

Rossco

You'd just want to group them on pass/fail wouldn't you? What other logical grouping is there?

Mat's Mug

by project, module, test name, outcome

Rossco

Possibly...

Mat's Mug

actually you're right

that's just candy

I'll just group them by outcome

Rossco

05:53

Anyway, I better go home now. Catch you later, enjoy.

Mat's Mug

and 2.1 can have more stuff

eh, going to bed

later!

Duga

[retailcoder/Rubberduck] retailcoder pushed commit 33237b1f to ICommand_Shelved: implemented TestExplorerControl.xaml, added ExportResultsCommand and resources. todo asap: reference .resx file in the xaml instead of hard-coding the strings.

Duga

06:51

[rubberduck-vba/Rubberduck] kevstark starred us

Meehow

07:54

Morning lads

vba4all.com expires in less than a month. I will not remain the owner any longer so if you or any of your friends want to take over we can make some arrangements

Stack Exchange

08:37

0

Q: Optimising "Find position of String in Dimension of Array" Function

Purpose of the function: Given an array and a string and bounds within which to search, find the position of the string within those bounds. What I want to optimise: If possible, I want to re-design the function so that it doesn't have 10 optional arguments, some of which are not actually opti...

RubberDuck

10:37

Someone wrote us just to say "thanks".

> Subject
Great software project!
Message
Dear developers of Rubberduck, after months or even years I had to write an Excel macro again – and I stumbled across Rubberduck more or less by accident because I was searching for some refactoring options in the default editor on the web ... To cut a long story short: This is a wonderful idea and a nice piece of software which I already felt in love with! :) Thanks for your effort! Best, Stefan

@Mat'sMug @Hosch250 great work guys.

Hey @Meehow. I'm sorry to hear that, but happy you've moved on to bigger things. I might get in touch. You've a ton of great content there that shouldn't vaporize. You made the Internet a better place friend.

Duga

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

Mat's Mug

12:16

> Rubberduck.Setup.1.4.3.0.exe(4.32MB) - Downloaded 428 times.
Last updated on 2015-07-08

Stack Exchange

12:28

0

Q: Performance Optimisation: Matching between files and a VERY long list of filenames

Purpose of the Macro: Given a folder typically containing ~250,000 small (20-100kb) files in HTML format (they open as single-sheet workbooks in excel) and a list of ~ 1 Million Filenames, open all the files that match the list and Do Stuff. Optimisation Parameters: It works, but will typically ...

performance file-system vba excel

Duga

13:25

[rubberduck-vba/Rubberduck] yadimon opened issue #749: Parser error on Split(x, "x")(0) and Dir$

> Hi, I tried some RD function over some projects,
the parser shows error, but for different lines, somewhere near the code examples.
The following code examples generates errors for Parser:

`somevar = Split(anyvar, "anystring")(0)` (the last part is the problem)

and

`Dir$`

Duga

13:43

[rubberduck-vba/Rubberduck] ckuhn203 merged pull request #748: fix links in readme.md from yadimon/Rubberduck/patch-1 into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] yadimon pushed commit c50b0edf to next

fix links in readme.md

from "retailcoder" to "rubberduck-vba" repo.

[rubberduck-vba/Rubberduck] ckuhn203 pushed commit d09a2e6a to next

Merge pull request #748 from yadimon/patch-1

fix links in readme.md

[rubberduck-vba/Rubberduck] ckuhn203 commented on pull request #748: fix links in readme.md

> You're awesome! Thank you very much!

[rubberduck-vba/Rubberduck] ckuhn203 commented on issue #749: Parser error on Split(variable, "text")(0) and Dir$

> Any chance you have the error message from the parser handy?

[rubberduck-vba/Rubberduck] ckuhn203 added label bug to issue #749: Parser error on Split(variable, "text")(0) and Dir$

[rubberduck-vba/Rubberduck] ckuhn203 added label code-parsing to issue #749: Parser error on Split(variable, "text")(0) and Dir$

Duga

14:02

[rubberduck-vba/Rubberduck] mgrojo commented on issue #262: Function/Sub never called Inspection

> When the use of the subprogram is performed through the `Call` clause, the procedure/function is reported as not used. I have removed these `Call` clauses since it makes apparently no difference. Then should `Call` be marked as obsolete as `Rem` ist?

I'm using 1.4.3.2343 (MSIL). Thanks for your work.

Mat's Mug

@Duga isn't it? or is the "obsolete call syntax" inspection not working?

Duga

[rubberduck-vba/Rubberduck] retailcoder commented on issue #749: Parser error on Split(variable, "text")(0) and Dir$

> The grammar doesn't understand that a function can return an *array* that's accessed immediately; the grammar rule allows for chained *member calls*, but the *array or procedure call* rule is currently standing on its own, which means...

Repro:

Sub Foo()
bar = Split("abc,def", ",")(0)
End Sub

Work-around:

Sub Foo()
temp = Split("abc,def", ",")
bar = temp(0)
End Sub

By splitting the call on two instructions, the grammar rules are satis

[rubberduck-vba/Rubberduck] retailcoder added label vba-grammar to issue #749: Parser error on Split(variable, "text")(0) and Dir$

[rubberduck-vba/Rubberduck] retailcoder removed label code-parsing from issue #749: Parser error on Split(variable, "text")(0) and Dir$

Mat's Mug

@RubberDuck it's purely a grammar issue ;-)

Duga

14:34

[rubberduck-vba/Rubberduck] retailcoder assigned retailcoder to issue #749: Grammar doesn't understand function's array return value being accessed in the same instruction

RubberDuck

@Mat'sMug I figured as much.

@Mat'sMug good question.

Duga

[rubberduck-vba/Rubberduck] ckuhn203 commented on issue #262: Function/Sub never called Inspection

> @mgrojo that's interesting. Did you receive an "Obsolete Call Statement" warning?

Mat's Mug

nope. the inspection is borked. dammit

and that one is more complicated

Duga

[rubberduck-vba/Rubberduck] retailcoder commented on issue #262: Function/Sub never called Inspection

> @ckuhn203 @mgrojo congratulations, you've found a bug ...well, two actually! :+1:

- `Call Foobar` should trigger a code inspection result saying 'Use of obsolete Call statement'.
- `Call Foobar` should be caught as a usage of the `Foobar` identifier... and it's not. Which means if you rename `Foobar` to `Foobaz`, the refactoring will break the code if `Call` is used.

Mat's Mug

Public Sub Foo()
    Call Bar
End Sub

Sub Bar()

End Sub

^^ right-click Bar in Sub Bar(), and select "find all references"

then remove the Call; "find all references" works as expected

there's a grammar bug, a resolver bug, and an inspection bug

RubberDuck

14:45

Ouch.

Duga

[rubberduck-vba/Rubberduck] mgrojo commented on issue #262: Function/Sub never called Inspection

> @retailcoder that makes my day! Thanks.

Duga

15:08

[rubberduck-vba/Rubberduck] ckuhn203 commented on issue #262: Function/Sub never called Inspection

> @retailcoder do you want to log those separately, or do you consider them to be part of this issue?

Stack Exchange

15:20

2

Q: Long running time on VBA macro which filters and splits dataset into new workbooks

The VBA macro below uses a worksheet (source_data_worksheet) to filter and split out the records based on about a dozen unique column (D) values in a dataset of about to separate workbooks which are saved to a folder. There are over 10,000 records x 18 columns total with text and numbers (the o...

performance vba excel macros

Duga

[rubberduck-vba/Rubberduck] yadimon commented on issue #749: Grammar doesn't understand function's array return value being accessed in the same instruction

> and what about Dir$ error?

```
Sub `Foo()
Dir$ ("/")
End Sub
```

Duga

15:40

[rubberduck-vba/Rubberduck] retailcoder commented on issue #262: Function/Sub never called Inspection

> @ckuhn203 I'll have to investigate. ObsoleteCallInspection worked fine in 1.22, and the inspection itself wasn't modified AFAIK, but there's a [known flaw that should also be addressed](https://github.com/rubberduck-vba/Rubberduck/blob/next/RetailCoder.VBE/Inspections/ObsoleteCallStatementInspection.cs#L24), even at the cost of performance:

//note: this misses calls to procedures/functions without a Declaration object.
// alternative is to walk the tree and listen f

[rubberduck-vba/Rubberduck] yadimon opened issue #750: Parser error: Select Case, multi Case

> example:

```
Sub Foo()
Dim a As Variant

Select Case a
Case 1 To 3, 4 To 6

Case Else

End Select
End Sub
```

[rubberduck-vba/Rubberduck] retailcoder added label vba-grammar to issue #750: Parser error: Select Case, multi Case

[rubberduck-vba/Rubberduck] retailcoder added label bug to issue #750: Parser error: Select Case, multi Case

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

> Nice one, good catch!

RubberDuck

I will never understand why so many people won't RTFM. More often than not, TFM has your answer.

Hosch250

Most people don't like to read anything, much less something technical.

RubberDuck

I know. It just baffles me the lengths programmers will go through to do something simple.

Duga

15:59

[rubberduck-vba/Rubberduck] retailcoder commented on issue #749: Grammar doesn't understand function's array return value being accessed in the same instruction

> Interesting. The same code using `Dir` instead of `Dir$` doesn't blow up the parser.

My guess is that the `$` trips the parser because in this context `Dir` is interpreted as a procedure call, and it seems only a *function* call is legal with a type hint - this code doesn't blow up the parser:

Sub Foo()
bar = Dir$("/")
End Sub

We might need to special-case `Dir$`.

Mat's Mug

hi @ya_dimon!

ya_dimon

Hi, i wanted ask you if someone used AccUnit already?

http://accunit.access-codelib.net/

because the unit testing process there is already nice implemented, its c# addin too.
maybe its useful to merge some features to RubberDuck

the problems its documented mostly in german ...
but i think its not a such big problem.

Mat's Mug

never used it

looks like it requires a truckload of boilerplate comment syntax to work

RubberDuck

I've used it and that ^

Any feature in particular that you're interested in @ya_dimon?

Mat's Mug

^^

exporting test results is in the making

ya_dimon

16:08

the comments used to pass "many params at one test method" only.
but i only used this, did not programmed it. so i dont really know how it works internal.
what i know, the unit test features was good :)

e.g. generate test by selected method/class.
does RD have something like this already?

Mat's Mug

not currently, but thinking that any given method should have only 1 test is fundamentally wrong.

ya_dimon

nobody thinks so :)
i m not sure, but generation was everytime like "methodname_insertyourcase_insertwhatitshoudlbe"
every generation -> new test. even for the same method.

Mat's Mug

I'd rather generate code from the tests ;-)

but yeah

ya_dimon

would be nice feature. lol

Mat's Mug

it wouldn't be too hard to generate stub test methods named after the members of a selected module/class

we could do that

RubberDuck

16:13

Sounds good to me. Submit the feature request. =;)-

Mat's Mug

good timing, I'm rewriting the unit testing UI for RD 2.0

(and the actual engine)

does AccTest work in Outlook VBA?

ya_dimon

another feature request, maybe linkage and submenu/hotkey for "go to test"/"go to source"

the generation could be on rightclick on module/method.

should i write all the requests on github issues?

dont know about outlook. i tried it in Access and Excel.

Mat's Mug

make 'em separate issues, it's easier to track that way :)

reason I'm asking about Outlook is that running VBA code from a VBE add-in in Outlook is different than in Access/Excel, in a major way; looks like RD 2.0 is going to support it.

@ya_dimon wait, what do you mean exactly with "go to test/go to source"?

Hosch250

Like R#, you mean?

Show Covering Tests?

Mat's Mug

not R#, DotCover

test coverage tooling would be beyond awesome, but isn't something that's currently possible.

ya_dimon

16:23

e.g. you are in the class, by implementing some method. and you want go fast to the test implementation (or create it). And form other side, you are in the test, and you want go by some hotkey to the source the test is for.

no no, coverage is too much for vba :)

but i would take as much as possible form JetBrains IDE's/Plugins

Hosch250

@Mat'sMug You think RD has a lot of issues? Check out Roslyn: github.com/dotnet/roslyn/issues

Mat's Mug

except RD isn't a frakkin' compiler ;-)

@ya_dimon but we can't just assume a test method maps to a given class/method just by its name. The only way to navigate from code to tests is by knowing which test methods are calling which code (which, now that I think of it, isn't impossible either)

and the other way around is sort of already implemented - if a test method is calling a given member, using "Find usages" on that member should locate the test(s) that is/are calling it

ya_dimon

the AccUnit used just name convention. You defined it yourself. Its like "%modulename%_test"
that is the simple solution.

same by methods like "%methodname%_somethingelse"

same could be used to know that are the tests, without @TestMethod

but i dont know if you want convention over configuration.

Mat's Mug

filtering usages into a "find tests" feature should be next to trivial

@ya_dimon RD makes no assumption on how user wants to write their code, as long as it's testable in some way that one or more assertions can be made

I find linking a test to a method just by name is pretty frail.

especially when one of your features is a rename refactoring ;-)

I'd rather link a test to a method by actual code references - if a test is calling Class1.DoSomething and Class2.DoSomethingElse, it shouldn't matter that it's a badly written test, "Find called code" (or whatever) should just let you navigate to either, regardless of the name of the test method

similarly, if I want to find all tests for Class1.DoSomething, I only need to find all test methods that actually call Class1.DoSomething - I don't need a naming convention

also, %methodname%_somethingelse doesn't handle EventProvider_EventName possible conflicts - which are indeed all on the user, but still, an unnecessary edge case.

FreeMan

didjall see this?

stackoverflow.com/10m

RubberDuck

16:43

@Mat'sMug code highlighting probably won't ever be, but with a good enough parser, we might get the stats at least.

> To the almost two million of you that have chosen to share your knowledge here:

THANK YOU

Mat's Mug

scroll down

RubberDuck

Swag

Duga

[rubberduck-vba/Rubberduck] yadimon opened issue #751: Feature request: generate test by module/function

> it would be nice to let the RD to generate new test module by selected one. (maybe on right click on a module as option too)
Maybe with already predefined test names, depends on module methods.

[rubberduck-vba/Rubberduck] yadimon opened issue #752: feature request: link tests with sources

> would be nice to have a link between the generated tests with sources by annotation or name convention.

an method/module could have some annotation like `'@test %testModuleName%`
or a test could have `@testFor` annotation.

or the test module could have name like `%moduleToTest%_test`
methods in the module `%methodToTest%_somthing`

(in both cases problem with "rename" refactoring feature)

than maybe some hotkey for "navigate to test"/"navigate to source" from the source/test

ya_dimon

ok, i wrote test feature request.

nice add-in btw!

i ll try to use, test and contibute to it. But the parsing errors should be corrected first :/

Cu guys!

Mat's Mug

Thanks!

Stack Exchange

17:00

Without @StackExchange, @rubberduckvba wouldn't even exist. #SOreadytohelp #VBA and #CSharp devs http://stackoverflow.com/users/1188513 I have 41 of the 10M

Tweeted by rubberduckvba on August 20, 2015 at 4:47 PM

EBrown

19:18

Hey @Mat'sMug I was just looking at this issue...and I wondered if the grammar wasn't bugged.

Mat's Mug

19:31

@EBrown it's very possible. I already fixed quite a number of bugs in that grammar already, it's highly possible that there are unknown ones, on top of the known limitations

> * Known limitations:
*
* 1. Preprocessor statements (#if, #else, ...) must not interfere with regular
* statements.
*
* 2. Comments are skipped.
*
*
* Change log:
*
* v1.4 Rubberduck
* - renamed to VBA; goal is to support VBA, and a shorter name is more practical.
* - added moduleDeclarations rule, moved moduleOptions there; options can now be
* located anywhere in declarations section, without breaking the parser.
* - added support for Option Compare Database.
* - added support for numbered lines (amended lineLabel rule).

EBrown

Well it's been a while since I've done any grammar work, but it looks like this bit is the issue: INTEGERLITERAL WS TO WS valueStmt (WS? ',' WS? valueStmt)*

Not going to do a Clone/Fix/PR just to test one grammar thing, though.

Mat's Mug

it's ok, I'll handle it. Regenerating the parser is kind of a bitch anyway

at one point I'll PR the RD grammar file back into the repo I took it from

EBrown

I can't even remember what that grammar is called anymore, to be quite honest.

But if my interpretation is correct, then that line is causing the issue with # To #, # To #. It's not seeing the second # To # as valid, methinks.

Mat's Mug

EBNF

EBrown

(Just thought I'd throw a suggestion in there, though.)

Mat's Mug

19:35

@EBrown sounds like it

EBrown

I would consider breaking the chaining off into a separate tree.

Mat's Mug

make a separate rule for each "syntax"?

might be worth it, at least for cleanliness' sake

EBrown

Well, the question I would leave to that then is: can you have multiple case types on the same line?

I.e.: Case Is < 15, 25 To 35, 65

If that is valid, then yes.

Mat's Mug

never occurred to me, but most probably yes

EBrown

I'm not a VBA expert, but I would test to make sure that is valid first.

If so, then rewrite the grammar to treat each syntax the exact same.

Or diverge it to me and I'll do it. :)

Mat's Mug

19:42

@EBrown it's valid

and it breaks the parser, obviously

EBrown

@Mat'sMug So in that case you should probably break sC_Cond into two different rules: sC_Type, and sC_Selection et al.

Something like:

sC_Type :
    ELSE                                                            # caseCondElse
    | sC_Selection (WS? ',' WS? sC_Selection)*
;

sC_Selection :
    IS WS? comparisonOperator WS? valueStmt                       # caseCondIs
    | valueStmt*                            # caseCondValue
    | INTEGERLITERAL WS TO WS INTEGERLITERAL    # caseCondTo
;

Mat's Mug

added up-for-grabs to the issue ;-)

Duga

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

> This:

| INTEGERLITERAL WS TO WS valueStmt (WS? ',' WS? valueStmt)* # caseCondTo

Is the problematic part. In fact, even with that sub-rule fixed, this valid code would still break the parser:

Select Case foo
Case Is < 15, 25 To 35, 65

End Select

We need to rewrite the whole `sC_Cond` rule, and break it down properly.

[rubberduck-vba/Rubberduck] retailcoder added label up-for-grabs to issue #750: Parser error: Select Case, multi Case

EBrown

That grammar might do the trick, just replace sC_Cond with sC_Type.

Mat's Mug

feel free to comment on the issue!

EBrown

19:48

Wilco.

Now, IIRC, Case 35 60 is invalid, no?

They have to be comma-separated.

Ismael Miguel

Greetings

Mat's Mug

@EBrown Indeed

hi @IsmaelMiguel

Ismael Miguel

@Mat'sMug Finally I'm giving use to my coclock. I bought it yesterday. Burnt-pizza free forever!

Mat's Mug

@IsmaelMiguel won't keep you from sitting on your pizza though ;-)

Ismael Miguel

Yes, it will

EBrown

19:55

@Mat'sMug Should I throw some support for parenthesis in there, or do you already have that with valueStmt?

Ismael Miguel

I hope

Mat's Mug

@EBrown valueStmt seems like it's supporting them already. the rule is self-recursive, so the order of rules in there is crucial (there was a bug in there, because somehow the original author thought it'd be a good idea to sort them alphabetically)

github.com/rubberduck-vba/Rubberduck/blob/next/…

EBrown

@Mat'sMug Aha, good to know. :)

I'll post this comment then.

Duga

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

> I would try breaking the grammar down a little further, as follows:

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

sC_Type :
ELSE # caseCondElse
| sC_Selection (WS? ',' WS? sC_Selection)* # caseCondSelection
;

sC_Selection :
IS WS? comparisonOperator WS? valueStmt # caseCondIs

Ismael Miguel

...

I'm in the wrong room...

EBrown

19:58

Wow, @Duga formats those code blocks awful poorly.

Ismael Miguel

-.-

Since I'm here, where are the files to translate?

EBrown

It's so weird commenting on other people's repo's.

Mat's Mug

ah, I mixed things up. it's the blockStmt rule that was sorted alphabetically

(still is, mostly, but at least implicitCallStmt_InBlock isn't in the middle of it)

EBrown

Indeed. That could cause some interesting results...lol

Mat's Mug

as it did :)

@IsmaelMiguel it's RubberduckUI.resx, under the Rubberduck.UI namespace :)

you would make a copy, and name it RubberduckUI.pt-PT.resx

Ismael Miguel

20:06

Alright

Sounds easy

There's also Language_DE at the very end

Mat's Mug

oh yeah, you need to have a Language_PT key if you want a localized language name in the language selector dropdown!

Ismael Miguel

I won't forget it

EBrown

@Mat'sMug You mind if I go through all the grammar, and do a quick spot-check of it?

Mat's Mug

and that's the part I need to tweak.. make that key Language_pt-PT and I'll do what needs to be done for it to work

@EBrown of course not!

EBrown

@Mat'sMug Alright, I'll give it a good once-over at least.

Ismael Miguel

20:09

I regret buying the coclock

Mat's Mug

feel free to poke at the other vba-grammar issues, too

EBrown

If I spot any other potential issues I'll...well...create an issue for them. :)

I'll likely skip the "assigned" issues.

Ismael Miguel

Remember what I said about being burnt pizza-free? I was wrong

Which elements are to translate?

Mat's Mug

@IsmaelMiguel the <value> elements under each <data> item

  <data name="CancelButtonText" xml:space="preserve">
    <value>Cancel</value>
  </data>

Ismael Miguel

Alright, I'm on it

Mat's Mug

20:16

It's important that the name attribute remains the same

Ismael Miguel

I know

If the name is different, it won't work

That's basic XML

Mat's Mug

it's also XML-encoded:

  <data name="RefactorMenu_Rename" xml:space="preserve">
    <value>&amp;Rename</value>
  </data>

& in menu captions determines which letter gets the Alt-shortcut underline

Ismael Miguel

I know

I was writting that

Mat's Mug

if you have any question about any ambiguous key, don't hesitate to ask. Some of them are used in very specific places

EBrown

@Mat'sMug Can you clarify this issue for me?

Ismael Miguel

20:19

  <data name="RenameDialog_InstructionsLabelText" xml:space="preserve">
    <value>Por favor, especifique um novo nome para {0} '{1}'.</value>
    <comment>0: DeclarationType; 1: IdentifierName</comment>
  </data>

What goes on {0} and {1}?

EBrown

Does he mean, #If DEBUG et al?

Mat's Mug

@IsmaelMiguel the declaration type in 0 and the identifier name in 1, like the comment says. e.g. "class" and "Foobar"

or "constant" and "FOO"

or "user-defined type" and "Bar"

Ismael Miguel

F*ck

Mat's Mug

see the DeclarationType_Xxxxx keys

what?

Ismael Miguel

Where else are those used?

Mat's Mug

20:22

those....?

Ismael Miguel

Yeah

"DeclarationType_*"

Those

Mat's Mug

in some code inspection results, and in some refactoring messageboxes

@EBrown yup

@IsmaelMiguel is it a problem?

Ismael Miguel

Basically, this isn't ready for genderalized words

For example

Mat's Mug

worked fine in French...

hold on

Ismael Miguel

In pt-pt, we say "a variável" (female), and in pt-br they say "o identificador" (male)

Mat's Mug

20:25

but bt-BR will have its own resource file

EBrown

So, you would make one for pt-PT and one for pt-BR. They would be different.

Ismael Miguel

I know

Mat's Mug

note: "identifier" != "variable"

Ismael Miguel

It's pt-pt

Unless you mean something else

Mat's Mug

  <data name="DeclarationType_Variable" xml:space="preserve">
    <value>la variable</value>
  </data>

^^ French translation puts the "la" in the value itself

"le module", "la variable"

Ismael Miguel

20:28

Do you use possessive?

Mat's Mug

@IsmaelMiguel in DeclarationType_Xxxx? No.

Ismael Miguel

And outside?

Mat's Mug

feel free to look at RubberduckUI.fr for reference

I don't see the point... using "sa" here:

  <data name="ExtractMethod_InstructionsText" xml:space="preserve">
    <value>Veuillez spécifier le nom de la méthode, le type de sa valeur de retour et les paramètres (si applicable).</value>
  </data>

(its return type)

Ismael Miguel

I don't write French

EBrown

@Mat'sMug Is Rubberduck.Parsing it's own standalone?

Mat's Mug

20:31

@EBrown yes

EBrown

Alright, I might rewrite some grammar stuff then.

Mat's Mug

feel free to implement whatever I've assigned myself to, I'm quite busy refactoring the Hell out of the main dll

EBrown

@Mat'sMug This is my issue search filter atm: is:open label:vba-grammar no:assignee

Mat's Mug

is:open label:vba-grammar would work just as well ;-)

EBrown

Nah, I'll start with unassigned. :P

Mat's Mug

20:34

there's enough for everyone lol

EBrown

Lol

Well I might take a look at some of these vba-grammar ones in general then, and see what I can/can't do to fix them.

Mat's Mug

#417 is serious shit

EBrown

I'll probably just comment fixes on the issue associated with it.

@Mat'sMug I just saw that, and I have a few ideas.

Mat's Mug

that's awesome

EBrown

How opposed are you guys to me rewriting the majority of the grammar, in general?

(Assuming it works after I rewrite it. :P)

Mat's Mug

20:37

thing is, renaming existing rules means breaking anything that uses the ANTLR-generated code, more particularly the resolver code and some of the refactorings and inspections

EBrown

@Mat'sMug Well we can probably keep most (or all) of the names then.

I just noticed some inconsistencies, that could likely be cleaner.

Mat's Mug

the original grammar

make sure you update the change log at the top ;-)

EBrown

@Mat'sMug Heh, assuming I make any direct changes. ;)

Like I said, I'll probably post the relevant grammar changes to the issue associated with them.

Mat's Mug

that's more than welcome as well

EBrown

That way we don't have to mess around with PR's.

And that way the actual code doesn't have my name on it, so when the build breaks, I'm not to blame. ;)

Mat's Mug

20:41

lol

2

Ismael Miguel

"Configure inspection severity." --> Is this meant to mean agressivity?

Mat's Mug

"inspection severity" ranges from "do not show" to "error", so not really

Ismael Miguel

If I translate it to Portuguese in those words, it will sound like it is beating the living hell out of someone

Mat's Mug

it determines what icon shows up for an inspection result (or whether the inspection runs at all)

Ismael Miguel

I'm trying to come up with something

I will skip it

Mat's Mug

20:48

there are a few that are pretty hard to translate

Ismael Miguel

This is one of them

And the type names

EBrown

@Mat'sMug I see what you mean about the names now. The VBAParser file has things like public const int RULE_startRule = 0...

Hosch250

@IsmaelMiguel Hehe, that's pretty funny.

Ismael Miguel

@Hosch250 It sounds even funnier in my head

Mat's Mug

@EBrown yeah. ANTLR generates classes and methods for each grammar rule

EBrown

20:51

@Mat'sMug As fast as I'm sure it is, that has got to be a debugging headache.

I love this: github.com/rubberduck-vba/Rubberduck/blob/next/…

Auto-generation for the win.

Hosch250

@Mat'sMug Is there a way to write unit tests for the grammar to make sure it actually works?

EBrown

Oh crap, @Mat'sMug there's another pretty big bug in the Select Case grammar.

Double values are entirely valid in the To syntax tree.

That is, Case 1.50 To 2.50 will fail your grammar check (if I'm reading it correctly) when it should pass.

Unless VBA doesn't consider those valid, but VB.NET (and I think even VB6) does.

Transcript for