RELOAD!

[bruglesco/FleetCommand] 2 commits. 31 additions. 7 deletions.

[rubberduck-vba/Rubberduck] 3 opened issues. 16 issue comments.

[Minesweeper] Games Played: 122, Bombs Used: 99, Moves Performed: 16169, New Users: 22

Hey @IvenBach

Yes?

You gitting good?

-.- I git by, just barely...

lol

With what I pulled earlier today with git rm . I will definitely say Nope.

There's still a lot I have to learn with git.

uhm, you're ahead of me. I use TortioseGit, and many of the menu options confuse me. I know I should CLI, but I just can't...

You're #DoingItRight, I'm #DoingItWrong

How many years did you use Excel to manually calculate formulas?
How many years did you use VBA without knowing how to use the Object Browser?
How many years did you try to learn to code while utterly failing?

I'll probably claim the top spot for all of those.

Me: Too many

10yr old Iven stat planned from lvl 1-99 for a game that way. #NoFormulasMA!

Meh, I solved business problems. Did I solve compsci problems? No. But who cares...

glad I can't see my VBA from 20 years ago

Glad I can't see mine from 3 years ago...

I actually miss VBA. Simpler times.

You coding in .Net now?

If I stay employed where I'm at it'll only be VBA.

How I wish I could get RD to compile on my work box.

@IvenBach Yeah, C# with "framework of the week". Integrating with Python, C, others...

@IvenBach I gotta say, that's a mystery. RD compiles on a clean Win10, for sure. Your work box is odd.

PEBCAK

@IvenBach No. your machine is borked.

Somethings definitely not right with it.

Can you install a Win10 WM on top?

Unlikely.

Can you have an accident with a cup of coffee? ;-)

j/k!

lol

this, Mug, Hosch, Vogel, et al. They've all said the same.

I'd ask to see your build logs, but if ^ couldn't diagnose it, I know I can't :-(

if it turned out to be something simple as missing a switch, that would be freaking hilarious

Ok, maybe not hilarious right now...

heh, yeh..

ish

@IvenBach you're able to build your WPF app on the work box, right?

--compileOnEveryonesPCExceptIvenBachWorkMachine = true

Anything that relies on a resource I have to comment out and build before I uncomment and have it work.

@mansellan That's how most everything feels like.

There is a reason, and it's not "because you"

You just haven't found it yet.

@IvenBach what?

sooo..... if you comment out the resources on your RD, it builds, too?

^^

I haven't tried that cause y'know there's like more than a few scattered all over.

@IvenBach can you elaborate?

Working on an MVCE with my work box right now.

I imagined but it would be a positive confirmatin that there's something wrong w/ the resources or wahtever

At least now I can #PhoneAFriendz and #CanHazAHopez for explaining it with MVCE.

You're finding an MVCE for the RD problems on your work box?

@TweetingDuck Bt;dt, it works. MS tryna get everyone on Win10, while still getting license fees from OEMs.

@this mkay github.com/IvenBach/MVCEResourceIssue will build on my home machine following a clone. No issue.

I clone on my work machine and I get an error.

Severity	Code	Description	Project	File	Line	Suppression State
Error		The tag 'Foo' does not exist in XML namespace 'clr-namespace:MVCEResourceIssue.Model'. Line 11 Position 10.	MVCEResourceIssue	C:\Users\ivenbach\source\repos\MVCEResourceIssue\MVCEResourceIssue\MainWindow.xaml	11

That's what prompted...

If I use the Name attribute it throws another build error which is pretty much what I'm getting in RD.

Are you sure that type Foo is in namespace MVCEResouceIssue.Model?

That's the right window namespace MVCEResourceIssue.Model AFAIUI.

<Grid>
    <Button Name="bar" />
</Grid>

^ In MainWindow.xaml produces the error on bulid:

Severity	Code	Description	Project	File	Line	Suppression State
Error		Unknown build error, 'Could not load file or assembly 'file:///C:\Users\ivenbach\source\repos\MVCEResourceIssue\MVCEResourceIssue\obj\Debug\MVCEResourceIssue.exe' or one of its dependencies. The system cannot find the file specified. Line 14 Position 17.'	MVCEResourceIssue	C:\Users\ivenbach\source\repos\MVCEResourceIssue\MVCEResourceIssue\MainWindow.xaml	14

You come to help rubberduck my build issue again Mug?

Ima XAML hopeless, but I can't see the 'Key' attribute being bound anywhere. Does it need to be a property in the Foo class?

I think the key is a xaml attribute. If I use another property like <Button Foreground="Black"/> I can build without issue.

Hmm - the error isn't same one on RD, I think

at least I don't remember this particular error about a tag not existing - it was about a file not found.

I'm stupid w/ XAML but is it because it's foo, not` Foo` (case sensitivity)?

maybe not - since you have model:Foo

yeah x:key is assigning it a key I thought

It could therefore be anything unique

@mansellan <UserControl.Resources> is essentially a dictionary; every entry (child tag) is an item; x:Key is, well, the dictionary key ;-)

Case sensitivity isn't an issue. My come box is building without an issue.

Ah I see you don't have public for your Foo

I think you need to make your Foo be public.

^ good point

yay padlock icons!

My non-repro test didn't use public

I didn't look at the icons; I was looking at the class definition on the right side.

still weird though... IIRC the default is internal

...which should work

@this yeah saw it after lol

^^

OK, GTK that internal is OK to use

I wouldn't bet on it working fine if it were in another assembly though

Even with explicit public class Foo it has same error.

hey, is your work computer's drives unusual?

roaming profile? network drive?

Not where I have the repo on the C:\...\Rubberduck.

and the `C:` isn't some kind of weird virtualized disk, right?

No. It's SSD.

it's actually a physical drive in your box.

I had already thought about that as well.

Ok.

Correct. SSD inside of the computer.

Other mapped network drives are coming from a server.

Thanks for asking that.

are you able to login using a local login?

(i'm assuming it's a domain-joined computer)

I'm using domain login.

do they allow local login?

(usually in form of YOURCOMPUTERNAME\administrator)

No. Only for laptops when others are out of the office.

Yeah, it would be insecure otherwise

Did you ever try deleting and recreating your profile?

As in windows profile?

yeah

it's a common troubleshooting step that domain admins do

No.

esp with Outlook for some reasons

you could ask your domain admins to do that or at least create a temporary profile just to rule it out

the fact that your MCVE is based on linking to another file, and the RD build has something similar, makes me think there's something funny with how it allows to find? use? create? link? the files

e.g. it's not allowing it to be used the way the compiler really wants it to be used.

@MathieuGuindon I'm just curious - have you had to create a collection of WithEvents'd variables and listen to events from any one of them?

It feels like this is the same issue with RD cause everything says something along the lines of XAMl issues. Or it's a non-build error that I'm guessing stems because of the XAML issues.

in your MCVE, you don't even see the BAML or *g.c.s files created, right?

No.

The only errors I see are regarding the XAML error with a specific line or the unknown build error pointing to an element that's using the Name or x:Name property.

@IvenBach FWIW, for non-trivial projects (RD certainly qualifies), the XAML is usually the very last to build. Anything failing before it (even a single error) can mask the problem in a thousand XAML issues. Is there anything other than XAML or CSC in your error list?

@mansellan That's all there is. In the file column CSC or FooBar.xaml

ugh, out of ideas. Sorry...

Don't be. It's had me flummoxed for a while but I think I'm starting to understand enough to tackle it.

For those who likes cargo cults, this might be an interesting read.

@this yeah?

dynamic controls, right? wrapped in a class?

actually, the controls being dynamic or not make no difference, come to think of it

@MathieuGuindon not necessarily controls; just to aggregate all events from controls into one place and handle it centrally or something like that. Up to now, I've been doing this with 2 classes; one to wrap the controls and take events, then pass it to the other class (of which there is only one instance). I was wondering if there's a straight way of just using only one class (with a private collection variable) instead.

that's pretty much how I'd have done it I think

2 classes, I mean

ok. just wanted to see if I can cut down on the boilerplate. Doesn't sound like.

> Computer programmers referred to cargo cult programmers, novices who ritualistically paste in lots of unnecessary code because they understand what the code does but not how it works.

That sounds a whole lot like a section of SO right there.

@IvenBach Did you ever try to build it from a network drive?

My uneducated guess is that you are missing permissions to access the location of the intermediate build results.

[rubberduck-vba/Rubberduck] daFreeMan created comment on issue #4926: '@Ignore ImplicitByRefModifier doesn't work with a line continuation

> @MDoerner I'm not quite sure I buy that logic - in the simple code I provided at the end, the 2nd @Ignore does work. i.e. there is no inspection generated on Function bar.

I get an Assignment is not used inspection on the 2nd line of this:

  Dim adoRecord As ADODB.Recordset
  Set adoRecord = New ADODB.Recordset
  Set adoRecord = adoCommand.Execute

I can't imagine that I actually need that 2nd line - I think it's a leftover from previous work or copy/pasta error, but wanted to double check before I delete it and create unexplained chaos in my code.

@FreeMan it's actually the first one you don't need.

since Execute will create a recordset for you.

By 2nd line, I meant physical line count in that set of 3, I was referring to the set adoRecord = New ADODB.Recordset. I think that's the one you mean, too?

yeah, I was counting the Set, not physical line.

Potato, potahtoe

.... Dan Quayle?

mostly

? How could you be "mostly" Dan Quayle? Either you are, or aren't.

:D

mostly quoting

actually, it's like being sorta pregnant

lol

and I fear I've dated myself with that reference. :-D

then we're on a date together.

O_O

I've been cleaning up code as I've practiced gitting this week. It seems I've managed to clean up a bit too much, but I'm not exactly sure what I've deleted.

I've been committing fairly regularly to my local repo, but I've not pushed anything upstream.

I know what method & module I've borked, what's the best way to get the code back?

I mean, I could go old skool, open up a backup copy of the xlsx and copy/pasta the method, but I wanna learn me something new!

(and no, it's not the line that @this said was safe to delete...)

@Duga the relevant part here is that the parameter declaration begins in the line after the annotation

@FreeMan git log can take a path argument

this allows you to see which commits touched the module

you can then look at these commits in specific (e.g. git show) to find out what was changed

@Vogel612 ah! I see your point!

@Vogel612 "path" as in directory? All my code is in one directory, so I'm not sure that will help, unless "path" includes the file name?

yes

well ... to be exact: no, yes

path as in path

looking behind the curtain: directories are just special files

paths and files are completely unrelated

[rubberduck-vba/Rubberduck] daFreeMan edited comment on issue #4926: '@Ignore ImplicitByRefModifier doesn't work with a line continuation

in this case then, it sounds like the path argument won't help me. :(

you misunderstand

yay!

a path can refer to a file, but it can also "lead into nothingness"

it can also refer to a directory, but that's just a file under the curtains of most filesystems

oh and paths are just chains of fragments

that's why using strings to represent paths is generally speaking a bad idea

ah, so git log filename does show me the commits in which this file was touched.

filename is that <path> argument there

Note to self: git bash <-- means remember you're using *nix here, not Windoze

therefore, apply *nix logic & thinking beanie

same logic applies to windows argument parsing, at least in PS

well, mostly

argument parsing is just convention

windows just doesn't have a proper convention

/arg:value /arg value /arg=value -arg value -arg=value --arg value --arg=value

^^ seen all of that already

windows has conventions, they're just unconventional.

lol no. Everybody just outright ignores them

similar to Linux - it's user friendly, it's just picky about who its friends are

MSBuild has at least three different ways to specify arguments

just saying

ugh... that's at least 30 years old...

YAY!!! I've found what I broke, and I know what commit it's in!!!

notes that reading diffs can be challenging...

There were multiple files in this diff, how do I pull back just the one I'm interested in?

@Feeds lol

@Feeds NSFE

^ha!

@Vogel612 I've seen all those in *nix, too...

@this Huh?

Not Safe For Eating

^ my take...

^ that

Oh, LOL.

I saw NSF, and auto-saw W...

@FreeMan I could just copy the one missing line from the diff and paste it into the IDE, but that doesn't feel right

@FreeMan That's what I'd do.

Source control is mostly there for logging the history.

Unless you want to go back, manually edit the commit (which is worse--you don't want to change history; the one exception is cleaning up a local development branch to make it represent the logical history of development instead of random commits when the dev runs out of time for a day), then re-apply everything and force-push to any other repos. And that doesn't work if more than one dev is working on it.

@Hosch250 hrm... OK. I'm good with that, it just felt like there should be something more... That long process sounds like the kinda thing you do if you're CDO or have really, REALLY messed things up.

Yeah, you don't want to do that.

I'd just put it back in and say "fix X"

[rubberduck-vba/Rubberduck] daFreeMan created comment on issue #4916: Permit MoveCloserToUsage refactoring to work on multi-line selection

> I'm really shocked that I hadn't submitted a request for this a long time ago. This is an excellent idea for those coming from cargo cult legacy code (of their own or others) to save significant time. As it stands, each must be moved one at a time with a full parse & inspection run in between, or moved manually then a single manual reparse. Either way, a time consuming (and error prone, in the 2nd instance) process.

What we have here is a #failuretopayattention

@Hosch250 It was windy, the W got blown over

Ha!

In light of my error, it seems that it's a GoodThing™ to commit very often after making very small changes.

makes picking through the resulting mess much easier.

@FreeMan yea, you may want the changes to still compile, though.

also there's git bisect which uses exactly that philosophy of committing to find which commit introduced an issue

yeah, that certainly seems reasonable. I wouldn't commit after every character or word typed...

A change to a single method seems like that might be too granular, too, especially if there are a not-insignificant number of changes coming, but certainly after changes to a single module.

commit messages should be like code comments - explain the why not the what, right?

I don't think so - you need the what, too.

It's useful to detail at the high level what changed to achieve the goal

^that does make sense

hunh, gitk is kinda handy

@Vogel612 so we're going to be using multiple settings file now? We just need to update the settings file to change the default and presto, right?

@this yeap, that should work

though not for all settings

why not all?

because not all settings maintain their default in a settings file

eg. ReferenceSettings don't

Right.

But that's not really a "default" setting anyway

Um... it is, though.

The ReferenceSettingsView allows adjusting that

so it's configurable and we're providing a default

i probably need to look at that closer - if it can be changed by user, it can't be a default setting

huh?

since default setting are kept separate from those in use.

Yes, and they are

github.com/rubberduck-vba/Rubberduck/pull/4914/…

I should actually update that to use a ValueDefault instead of DefaultSettings

oh well ..

I guess I'm confused because if we are providing defaults, they should come from a settings file -- which can be serialized into a XML file... right?

no. that's not how it works

the settings file provides defaults in the form of an xml string

which is automagically parsed, assuming we configured the settings wizard correctly

Right.

So we provide a serialized XML, which is then deserialized to obtain the defaults.

When reading the settings, user configuration trumps default configuration (which is handled by the PersistanceProvider)

There's a handful of settings where that doesn't work (e.g. HotkeySettings)

or rather where the setup makes it a bit harder to work stuff out correctly

And there's settings where the defaults aren't read from an XML-string, but are maintained in code

those two are equivalent, though

Yes - if my memory serves, we persist the original default and any users changes (if applicable) into the rubberduck.config XML file. From that point now on, we are using that to provide the current setting.

basically, yes

though again: if the user doesn't configure it, we don't serialize it to rubberduck.config, IIRC

And we have the "Reset" functionality

The point being that the original defaults are never written to; they have to come from a separate place in case where the config file doesn't exist.

You're right we only serialize it if user made a change.

ahh, now I understand what you're getting at :)

Hence why I was confused why reference settings don't get their own settings file and thought that was because we have no defaults at all?

oh wow....

spectrum.ieee.org/aerospace/aviation/…

> Like someone with narcissistic personality disorder, MCAS gaslights the pilots. And it turns out badly for everyone. “Raise the nose, HAL.” “I’m sorry, Dave, I’m afraid I can’t do that.”

Yeah, pretty much that.

Good news! Found out it's not just my user login that can't build.
Bad news! Another profile couldn't build on my machine.

Installing on another machine at IT suggestion. Woot!

[rubberduck-vba/Rubberduck] CSusgithub starred us

Day cut short because of sick duckling. But I found childhood happiness nathanfriend.io/inspirograph!!!

@IvenBach That. Is Totally. Awesome!!!

to get it right

Slow day today.

As in, I have a lot of work to do, but the systems are slow, so I'm barely getting anything done.

sigh

that good, eh?

I've been trying to set up a second instance of our site on my local since 9:30.

Except they made some DB model changes and didn't tell our EF model about it.

So the site seeders are crashing and I have to go through one by one and get them working again.

Finally done.

And I didn't get one of the seeders working, at that.

That was weird. Outlook just crashed on me.

Trying to drop an Excel attachment on an email.

Even more odd, the excel window moved from one monitor to the other when Outlook died.

Did any here play zork?

not zork but HHGtG, yes

(and WishBringer)

going to be a fun read!

not sure I played any of those, but if I go back to that ZIL manual, I'll be in the office all night!

closes tab, saves marriage

[rubberduck-vba/Rubberduck] MDoerner removed label PR-Status: WIP from pull request #4924: Remove dymanic calls from resolver

I feel like I'm missing something here -- isn't that redundant?

Oh, I forgot do comment on my PR that I have no idea what this method was really trying to achieve.

I wonder if it's actually supposed to be var paramList = proc.argsCall;?

since it was originally dynamic, we had no assurance that argsCall is valid, I guess.

@this you able to help me understand the binding for github.com/rubberduck-vba/Rubberduck/blob/next/Rubberduck.Core/…?

I can't figure out where the DataContext is coming from.

@this There has not been any context with an argCall method for a long time.

IIRC, argCall was a subrule of callStmt that the SLL parser really did not like too much.