« first day (1326 days earlier)      last day (1854 days later) » 

7:00 PM
@Vogel612 Had a gathering for my g-ma's 90th bday party over the weekend, the one that's diabetic. She ate plenty of the chocolate my cousin brought. Seems my dad's not the only one with a morbid sense of humor.
 
Are they trying to kill her, or does she just like candy so much she doesn't care about the results?
 
Microsoft devs using #MSAccess have a ideas for much-needed improvements. Go vote for your favorites. Data import, integration. @Office365 @OfficeDev https://access.uservoice.com/forums/319956-access-desktop-application/filters/top
 
@Hosch250 The latter. She knows she shouldn't but still eats it. All things considered she's in really good health.
 
@IvenBach LOL, that's not surprising.
I'm beginning to be convinced that people who worry about their health destroy it more than people who just live.
 
Hosch can you test out portal.azure.com to make sure it's not just me getting a bad request message.
@Hosch250 Moderation should be followed first.
 
7:05 PM
Oh, I'm all for moderation, as long as you don't worry about it.
Yeah, I got in just fine.
 
I'll restart then, it's been a week+
 
@IvenBach Restart what?
 
Computer.
@Mat'sMug I'm anxious for the upcoming magic RD is going to allow. Red squigglies, anti-HN, code folding, and the rest is mind blowing for me. #MustHaveNao!
Syncing with RD, 25 skipped tests sound correct?
 
7:35 PM
that's what I get
 
[rubberduck-vba/Rubberduck] build for commit 3cac2cd4 on unknown branch: No report found to compare against
 
Figuring out the correct quick fix is going to be non-trivial.
 
> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3695?src=pr&el=h1) Report
> :exclamation: No coverage uploaded for pull request base (`next@3ce0920`). [Click here to learn what that means](https://docs.codecov.io/docs/error-reference#section-missing-base-commit).
> The diff coverage is `98.69%`.

[![Impacted file tree graph](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3695/graphs/tree.svg?width=650&height=150&src=pr&token=0zIVyLTgaw)](https://codecov.io/gh/rubberduck-
> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3695?src=pr&el=h1) Report
> :exclamation: No coverage uploaded for pull request base (`next@3ce0920`). [Click here to learn what that means](https://docs.codecov.io/docs/error-reference#section-missing-base-commit).
> The diff coverage is `98.69%`.

[![Impacted file tree graph](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3695/graphs/tree.svg?width=650&height=150&src=pr&token=0zIVyLTgaw)](https://codecov.io/gh/rubberduck-
 
@IvenBach if it was simple, another VBIDE add-in would already have done it :)
 
Only the tough ducks get lunch?
 
7:41 PM
...
it's non-trivial, sure. but not impossible
pretty sure we have all the framework to make it happen by now :)
@IvenBach TBH that should be a very fun one to tackle
 
@IvenBach Do that in a separate PR.
Fix up the tests, and that one is good to go.
 
@Mat'sMug You have a perverse sense of fun.
 
[rubberduck-vba/Rubberduck] build for commit 3a929ac4 on unknown branch: AppVeyor build succeeded
 
@IvenBach LOL, whenever something hard comes up at work, my first response is "That'll be fun."
 
@IvenBach what fun is a piece of code that's so trivial it "just works" as soon as you put in the final semicolon?
 
7:49 PM
"There are two ways to write error-free programs; only the third works." - Alan J. Perlis
 
gotta enjoy a good game of whack-a-mole once in a while...
 
If I have the knowledge to successfully achieve that goal, sure. I question being able to successfully tackle this one since I already know it's going to deal with updating the implicit variant inspection too.
 
Rome wasn't built in one day
 
True.
 
actually the implicit variant inspection may or may not need to be touched, depending on what you do with the resolver :)
 
7:56 PM
I didn't know the resolver needed to be involved.
#BlindDuckStrugglin
 
it's the resolver that's creating declarations and deciding "hmm, that one is implicitly Variant"
I'd find where that's happening and see if there wouldn't be a way to type the declarations as per any Def[Type] instructions in the module
and then the implicit variant inspection would magically stop firing results for them
without changing a single line of code in it
and the rest of RD would see these declarations with their correct type
 
Sounds simple enough... Famous last words.
 
it's not simple
it's fun :)
worst case, you break 1.6K tests
 
^
 
so what?
 
7:59 PM
Fun relates to relative skill.
 
no
go ahead and break things
that's how you learn
 
I guarantee that I will.
 
I guarantee that you will
I would, too
and then you debug, make a small tweak, re-run the tests, and boom, 1.5K tests fixed
so you narrow it down to the remaining 100
rinse & repeat
 
I hope I have enough of a skillset to understand the code that's there. I'll only find out by trying.
My duckling already knows I'm a failure, but not a quitter/giver-upper.
 
Eh, you should.
@Mat'sMug You know what's the best? When you do waterfall so fast it is agile.
 
8:05 PM
lol I guess
 
That's what we do at work.
 
@Hosch250 I've never done data driven tests. I read your comment but didn't fully grasp the idea.
I'm grabbing some lunch. My brain's not functioning at the level it should.
 
8:21 PM
@IvenBach me neither. think of it as a parameterized test, where you have some metadata feeding the parameter to the test runner.
 
        [TestCase("Bool")]
        [TestCase("Byte")]
        [Category("Inspections")]
        public void DefType(string type)
        {
            const string inputCode =
$@"Def{type} A
Public Function aFoo()
End Function";

            var vbe = MockVbeBuilder.BuildFromSingleStandardModule(inputCode, out _);
            using (var state = MockParser.CreateAndParse(vbe.Object))
            {
                var inspection = new DefTypeStatementInspection(state);
                var inspector = InspectionsHelper.GetInspector(inspection);
That will run twice.
 
once for [TestCase("Bool")], once for [TestCase("Byte")]
 
Yup.
 
so easy it's a shame we haven't written a single one like that
 
I've used these before, but I don't actually remember where. It might have been my checkers game.
@Mat'sMug MS Test didn't support it.
 
8:23 PM
well this explains that
 
Spotify needs a new ad person.
The only company I've heard ads from on Spotify and not put on my blacklist is Dollar Shave.
They have the funniest ads.
Our ticketing system keeps throwing me out because we don't have enough licenses.
We are working on more, but this is the crappiest ticking system.
We can't even have it open in two tabs because the (@#)$@*#%)!#$ system thinks it is two sessions.
Never, ever use Helix.
The only reason we don't throw it over is half the company likes it for some inexplicable reason.
It's as bad as TFS.
You send a feature to Stage for testing, and it gets rejected. you can't just set the status to Dev for more testing once that goes in, you have to follow the workflow, which is back to stage...
It's so bad that between TFS and Helix, I'd look for somewhere else if our codebase wasn't on the mend.
Hi, @AnnaVel.
 
Hi :)
 
Any luck with whatever you were working on?
 
@Hosch250 that was what I thought. I can see why that’s preferred.
I like it. Simplifies the ammonia unit of code written for tests
 
Oh it is going well :) It is the push/sync/pull request process that is killing me
We have TFS at work so git is a pain
 
8:38 PM
Nooooo.
TFS is murder.
Git is love. Git is life.
 
Haha, I really should be more considerate to people's convictions when I pass judgement like that :)
 
Don't worry, we have TFS at work too.
Git will certainly take getting used to, but it's so much more powerful and fast.
 
What do you do at work?
 
ASP.NET MVC.
 
Oh, how nice :) This seems so complicated to me, I work desktop only
Ok, corrections are made, build is building successfully, tests are passing
I am going to brave the check in process after I get some sleep
 
8:50 PM
@AnnaVel oh, Team Frustration Server! My deepest sympathies :)
 
@Hosch250 How did you learn about setting tests up like this?
 
Let me think. Was it the web?
@AnnaVel I wish!
I was hired for backend, but I do way more front end.
@IvenBach TBH, I don't know where I first learned about it. I think Mat or Vogel mentioned it once.
But NUnit has really good docs.
Right now, we have merge conflicts in TFS, so we are having the most trouble doing a hotfix.
Git would just let us apply a stupid cherrypick and get the darn thing done with.
 
9:10 PM
@Hosch250 I think that's a really cool capability. That will definitely cut down on the LoC.
@Hosch250 Is it a best practice to keep DefType_AllDefTypeAreFound and DefType_AllDefTypeAreIgnored separate tests? I can see how they could be combined but that's starting to feel like abusing the use of TestCaseAttribute
 
@IvenBach Sure.
Oh, best practice?
I'd have to look.
 
The only difference is '@Ignore DefTypeStatement and the result count.
 
Oh, yeah, keep them separate.
 
At least a have some sense about when to use it.
 
Use it when the tests have the same purpose AND they are similar enough to share data.
 
9:20 PM
Don't follow DRY as an absolute, use sense where it's merited.
 
Especially in tests.
 
9:32 PM
@IvenBach only a Sith thinks in absolutes
no coding principle is a law
 
@Mat'sMug Then I'm a Sith.
 
I knew it!
 
That means Mug is his master.
 
Dumb but every time I see word Sith, I can't help but not think of Slit throat
 
A principle is a law given certain conditions.
 
9:34 PM
which isn't that far off describing Siths.
 
@IvenBach I own no man as my master.
 
What?
 
The only absolute I'd consider remotely claiming as true is absolute zero.
I heard that even that was suspect.
 
How does one owns master?!? If one did, then that "master" isn't really a master.
 
@this Synonym for acknowledge.
 
9:35 PM
ah less sense => more sense.
IOW, a woman is your master, got it.
<wg>
 
> to acknowledge to be true, valid, or as claimed
@this I meant man in the old sense of the word.
 
@this Any married man ought to acknowledge that. At the very least consider himself an equal with his wife.
 
Ambiguity is the engine of poetry, humor and politics, as someone told me.
 
@IvenBach Fortunately, I'm not married.
 
I know. But I think most the rest of us are.
 
9:37 PM
But if I do, she'll certainly be my equal. I could marry noone who I didn't view as my equal and who didn't view herself as my equal.
 
^ That's a great start.
 
But it won't just be equal. We'll be partners (and not in the sexual sense of the word).
 
^ ... not just in ...
 
Point conceded.
But, I have high ideals for marriage and a limited interest in it, so...
 
9:55 PM
BTW, I thought Bluetooth was line-of-sight.
What kind of line-of-sight is that?
Because just a little bit ago, I went to the kitchen and forgot to turn my earbuds off.
That's on the other side of a conference room and the bathrooms, so several walls. And I was still connected perfectly fine.
 
> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3695?src=pr&el=h1) Report
> :exclamation: No coverage uploaded for pull request base (`next@3ce0920`). [Click here to learn what that means](https://docs.codecov.io/docs/error-reference#section-missing-base-commit).
> The diff coverage is `98.69%`.

[![Impacted file tree graph](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3695/graphs/tree.svg?width=650&height=150&token=0zIVyLTgaw&src=pr)](https://codecov.io/gh/rubberduck-
 
@Hosch250 ^ How do those tests look?
 
Good.
Now, once those tests run, I'll show you something interesting.
You'll probably already have noticed it if you had a chance to run locally.
 
I test locally F5/Unit before pushing, usually.
I'm interested, but might not have time. In between setting up a cluster to run some engineering calcs.
 
Won't take any time at all. I'll just post it in the chat when it loads.
 
10:08 PM
Ok.
I just noticed in the output window
[1/22/2018 2:02:16 PM Warning] line 2:0 extraneous input 'Def{type}' expecting {<EOF>, ':', ATTRIBUTE, BEGIN, CONST, DECLARE, DEFBOOL, DEFBYTE, DEFDATE, DEFDBL, DEFCUR, DEFINT, DEFLNG, DEFLNGLNG, DEFLNGPTR, DEFOBJ, DEFSNG, DEFSTR, DEFVAR, DIM, ENUM, EVENT, FRIEND, FUNCTION, GLOBAL, IMPLEMENTS, OPTION_BASE, OPTION_EXPLICIT, OPTION_COMPARE, OPTION_PRIVATE_MODULE, PRIVATE, PROPERTY_GET, PROPERTY_LET, PROPERTY_SET, PUBLIC, REM, STATIC, SUB, TYPE, VERSION, NEWLINE, ''', WS, LINE_CONTINUATION}
 
10:21 PM
[rubberduck-vba/Rubberduck] build for commit 2599f557 on unknown branch: AppVeyor build succeeded
 
> RubberduckTests.Inspections.DefTypeStatementInspectionTests.DefType_SingleResult‌​Ignored("Byte")
That's how it displays the tests so you know which is which.
 
[rubberduck-vba/Rubberduck] IvenBach pushed commit 6b7dbe96 to next: Remove unused code
[rubberduck-vba/Rubberduck] IvenBach pushed commit a063a07b to next: Unit tests
[rubberduck-vba/Rubberduck] IvenBach pushed commit e300f481 to next: Update resource file
[rubberduck-vba/Rubberduck] IvenBach pushed commit b981d678 to next: Add inspection
[rubberduck-vba/Rubberduck] IvenBach pushed commit 3cac2cd4 to next: Merge branch 'next' of github.com/rubberduck-vba/rubberduck into Iss388_Obsolete_DefType_statement
[rubberduck-vba/Rubberduck] IvenBach pushed commit 3a929ac4 to next: Merge branch 'next' of github.com/rubberduck-vba/rubberduck into Iss388_Obsolete_DefType_statement
[rubberduck-vba/Rubberduck] IvenBach pushed commit 6bd42d2e to next: Parameterize unit test
[rubberduck-vba/Rubberduck] IvenBach pushed commit 2599f557 to next: Clarify resource string
Merge pull request #3695 from IvenBach/Iss388_Obsolete_DefType_statement

Iss388 obsolete def type statement
 
@Hosch250 That one I caught. I was confused for a bit thinking 'I didn't name it that...' then realized the parameter.
I'm not dumb, just slow. Really slow.
 
I knew you'd see it. Mostly posted it for everyone else's benefit.
 
@Mat'sMug - here's a quick'n'dirty sample using AddressOf: gist.github.com/bclothier/c0f76c252b2641f2ef2cf520c43c5642
 
10:35 PM
@Hosch250 Thanks for the explicit check. I have missed those things before.
 
@Duga hmm i wonder if we are getting slower and slower to build.
 
11:17 PM
Having to remember how you got something to work months ago isn't the easiest. Especially when it was your first time doing it. #GoDocumentationWriting...
 
11:40 PM
But it feels good when you figure it out.
 

« first day (1326 days earlier)      last day (1854 days later) »