ooh, double-yearling on DBA.SE now

what do you think the argument to GetLowerBound means?

GetLowerBound(0) returns 1 because the LBound of the first (0th) dimension is 1

GetLowerBound(1) returns 1 because the LBound of the second (1) dimension is also 1

nooooooooooooo

?

I'm so dumb

I meant UpperBound there

haha

I'm still glad we talked about all of this, I learned two good new habits with Debug and also hard casting.

But man still 20 minutes of talking because this guy right here put Lower instead of Upper

eh, others might chew my head off for using var everywhere

Not me, I already like it.

Also, not sure if 2017 has this, it probably does,

but you were telling me System.Diagnostics namespace.

Ctrl+.+ENTER auto-inserts it?

that's been a R# feature for as long as I remember :)

2019 suggests any needed Namespaces with a lightbulb as long as I'm using a valid function from an included? imported? namespace.

yep

Nice, I might do that, sometimes the light bulb tries to get out of my way and it can be hard to pin it down for the click.

works in vanilla-VS2015 here too

@puzzlepiece87 Ctrl+.

#NoClick

:)

The main keyboard shortcut for VS that I've mastered so far is Ctrl+K, Ctrl+D to make things pretty.

I'll add Ctrl+.

VS is pretty great, I like it.

TBH that hotkey is messing with me big time... with R# it's Alt+Enter (or Ctrl+Enter, I always get mixed up now)

@puzzlepiece87 Ctrl+Shift+B has to be the most-used VS hotkey

I'm guessing that's build.

Right now I've been using F5 for that instead.

Yeah, that's build.

But I know if you just want to make sure you don't have compile errors left you'd build instead of run.

I'm finally done with compile errors on this program I dragged out of Excel. 100,000 taken care of.

Now I'm just debugging my translations.

R# tells you that without needing to build :)

(and Roslyn is getting better at it, to be fair)

Let me see what I have.

(One second)

[zspitz/ANTLR4ParseTreeVisualizer] zspitz created draft release 0.1.0.72

now that would be interesting

[zspitz/ANTLR4ParseTreeVisualizer] zspitz released release 0.1.0.72

[zspitz/ANTLR4ParseTreeVisualizer] zspitz published release 0.1.0.72

[zspitz/ANTLR4ParseTreeVisualizer] zspitz created draft release 0.1.0.73

[zspitz/ANTLR4ParseTreeVisualizer] zspitz published release 0.1.0.73

[zspitz/ANTLR4ParseTreeVisualizer] zspitz released release 0.1.0.73

@MathieuGuindon I would definitely be interested in new ownership.

Ooh my SO rep bump kicked in, made a whole almost 1.5K!

TTGH

@MathieuGuindon Thanks again for the help!

yw

@puzzlepiece87 I'm pretty deep in funucking my previous code. Mug got you squared away?

@Past-Iven: Your attempts at unit testing were admirable and well intentioned. We've learned a few things since then. Make sure your unit tests don't hinder refactoring.

@VBA Why you have no code folding? Be like VS.

@IvenBach Yup. I wasted a bunch of oxygen asking "Why is var uBound0 = Array.GetLowerBound(0); not working?"

facepalm

I've been there before. Reading things that weren't actually there, until I actually paid attention and read Every Single Character As Written.

@MathieuGuindon Your final gift to me was helping show me that our company didn't lock down the VS installer and I can update on my own.

TFW the bug is in code you were previously told not to update.

@puzzlepiece87 What did you expect that to result?

I claim no confidence in other peoples code. Nor mine without unit tests.

I don't trust my own code until I see it producing the correct result. And even then I'm hesitant.

^ So many times over.

That combined with a couple of "Huh, that actually worked?"

Based on the current inputs it's giving me correct outputs. Can't guarantee that for all inputs.

Input validation. If you haven't tested the input, refuse to calculate the output.

There was a time when I thought TDD would save me. Gave a false sense of security instead.

Give it to me or FreeMan and we'll find a way to break it.

This is a perfect example of why if you write tests at all, you have to write a lot of tests.

22 tests on a very simple function, still cases that were missed.

Fails for 2, 3 and 4 as input.

@Mast The upper bound, but instead I requested the lower bound by accident.

Anyone using the ANTLR debugging visualizer with 2017, could you test the 2019 DLL in the latest release with VS 2017?

@puzzlepiece87 Important difference. Luckily there's the function call telling which bound it picks.

(I am so going to regret all the time I spent making sure there were two builds...)

NB the latest release uses C# 8 -- particularly nullable reference types -- with .NET Framework.

@ZevSpitz Wouldn't it make more sense to use the 2019 debugger?

@Mast What's the current status of RD? Can it be opened in VS 2019?

@IvenBach thecodelesscode.com/case/116

@ZevSpitz I don't know. Anyone?

RD cannot be built in VS2019 yet.

@M.Doerner :+1:

We will merge the PR to change that after the upcoming release.

Afterwards, RD can no longer be built on VS2017.

@IvenBach Some weeks ago, I went and read them all.

I need to do that now I've a better understanding of coding.

@Mast Until RD moves over to 2019, I think the debugging visualizer needs to work with VS 2017.

Fair enough, I thought they were over already. My bad.

I think all that is missing for the release are some translations and a PR in which I will fix the unintended consequences of am enhancement I introduced.

@M.Doerner Are the .csproj files in the new SDK format? Or in the old format?

They are in the new format.

That is why we have to use a different SDK to make the WPF built work with the VS2017 version of msbuilt.

Great, not only did Excel crash, the entire VM went with it.

The latest version of the normal SDK finally works for WPF bultds, but cannot be used from VS2019.

And the alternative SDK was not upgraded to work in VS2019, because that workaround is no longer necessary.