On a more uplifting note... my conversion of my simple test sub to figure out how to call this API to an actually useful Interface/Implementation was a success!
I think I've asked this before, but don't recall the answer: If I'm working inside the class, is it kosher to reference the Private this.x variables defined within the class or should I use the Public Property Get/Let/Sets? I know that it doesn't matter functionally, just wondering what's the "right" way.
Skipping the extra function calls involved in using the Get/Let/Set seems to make sense, but...
@FreeMan if the getters do nothing but get the private field (as they should), I don't see a reason to not use the private instance fields directly. But YMMV =)
Conventionally, for dumb getter/setters, it's probably better to use the internal access this.x rather than x/me..
I have had one or two scenarios where the getter wasn't dumb and the internal access had to use the public member instead of this.x. But honestly that never felt right to me.
> **Justification** The wording of the message box can be misinterpreted, leaving the user thinking they are presented with release notes for a newer version that was just installed. Moreover, clicking "no" makes the next load prompt again, which can get annoying.
**Description** Let's clarify the message box to: - include the currently-running version# - indicate how it can be configured to stop nagging
Let's also track the last time the prompt was displayed, and introduce a settin
> Linking #2941 -- that would be an good opportunity to create the custom implementation of messagebox that Rubberduck should be using instead of lame old WinForms. We get the better UX by having action directly labeled in buttons and ability to toggle setting without having to close the dialog and open up settings.
> @bclothier indeed! "don't ask me again" would automatically adjust the setting to no longer run the version check. Also an opportunity to pull the version-check settings into their own "Upgrade" settings page.
Excel is great for lookups, but (assuming the data is living in a worksheet) generating unique IDs per record is the job of a database engine, not a worksheet - what happens when a row is deleted? Are IDs reused? Worksheets aren't made to ensure the integrity of relational data. If you really want/need the IDs generated in a worksheet I'd strongly recommend generating GUIDs instead of using row#-based or otherwise sequential IDs. — Mathieu Guindon14 mins ago
I was editing that chunk into an existing answer after some random upvote, and then figured "hey that would make great 'about me' content on my profile"
> relabelling as enhancement, because strictly speaking this is not "wrong" behaviour, just crappy behaviour. FWIW I'd expect the preview box itself to be readonly and therefore skipped when tabbing. That would make it unreachable via keyboard navigation, which carries accessibility implications (not that the VBE is accessible...)
also: with the exception of indexers (i.e. parameterized getters, which should raise index out of bounds given an invalid index argument), generally go out of your way to make a getter never raise any errors.
@Duga weeeell... if the cursor goes somewhere that it can't get out of, that's kinda buggy. Frankly, you can label it fred for all I care, it's on the radar for fixing enhancing.
gah... I have a procedure to call stored procedures. I felt pretty clever when I wrote it. Now I'm looking at it like, "carp - what id10t designed this horrific interface???"
Once I actually have the current crisis resolved, I'll get to tweaking that one.
It looks like I'm the preferred candidate right now. 2nd Phone interview Tue. If I'm a good boy duck and does well Wed is in person with pending offer. #Breathe
4
@FreeMan I been occupied with putting out work fires.
FWIW, i'm re-reading this book. A good read if you want to get a good theory.
fairly short, targeting those who are already familiar with SQL but want to understand the relational theory. CJ Date worked with EF Codd and is a mathematician himself, so that's a good resource.
twice the size and 8 times the mass of Earth IIRC. so sure, possibly life-supporting (habitable zone, water vapour in the atmosphere), but 8g sounds a wee bit too much gravity for me.
OTOH it makes a nice excuse to make exoskeletons
it's >100 LY away though, so.. don't hold your breath for a visit
that's funny because in the 80s Biff was often associated with Buffy and the argyle-sweater-over-the-shoulders, drinking-tea-with-your-pinky-in-the-air country club set.
well, at least until BTTF came out. But even after that...
This is irritatingly frustrating and I'm gonna blame it on not getting my nap in.
When I provide some sample data in SSMS and execute the code that became my sproc, I can shred my input XML string and insert into my table. However, when I take that sample data and pass it to the actual sproc, I get nada
I've debugged through it and I'm not getting any errors...
never seen BTVS. Took a sec to even figure out the acronym... The judge in Caddyshack is what Biff generally brings to mind for me - pompous, sniveling, selfrightous, self important, spinless country club type.
@FreeMan I know what types you're talking about. Those are supposed to have big and pretentious names like Bartholomew William Quinn III or something like that.
IMO, what they really need is instead of a full-blown "union", they need several standardized trade/movement laws. Much like the US states had to standardize trade/contract laws.
(Movement here was set by the federal gov after some states [looking at you, CA] tried to ban interstate immigration.)
That way, it wouldn't be as much an issue with things like banning selling apples by the pound.
@mansellan my understanding is that brexit referendum passed by a hair, after a campaign of, essentially, misinformation and fear-mongering... Kind of like how "no" won twice in Quebec.
am I way off?
(could be - I follow US politics a lot more closely than UK's)
> #5121 might ease this. If we implement client/server, the server can be any language. We could implement it in Net Core 3*, which has some low-level memory constructs - disposable ref structs, Span and Memory and others. These can help avoid allocations and copy
@MathieuGuindon well, it was 52/48, but it was also the biggest voter turnout in British electoral history, so that slim margin was 1.3 million votes. They also explicitly said that the result would be implemented... And yes, there were lies. Plenty of them, on both sides. Obviously, as politicians were involved :-)
> #5121 might ease this. If we implement client/server, the server can be any language. We could implement it in Net Core 3*, which has some low-level memory constructs - disposable ref structs, Span and Memory and others. These can help avoid allocations and copy
> #5121 might ease this. If we implement client/server, the server can be any language. We could implement it in Net Core 3*, which has some low-level memory constructs - disposable ref structs, Span and Memory and others. These can help avoid allocations and copy
> #5121 might ease this. If we implement client/server, the server can be any language. We could implement it in Net Core 3*, which has some low-level memory constructs - disposable ref structs, Span and Memory and others. These can help avoid allocations and copy
@this was thinking, the framework bump for mocking loses us Vista. Wonder if it's worth going a little more recent anyway, and lose Win8 (original) too? Might be better to get OS drops out of the way in one hit rather than piecemeal, and Win7, 8.1 and 10 are fully supported on all frameworks through to current....
So maybe go to, say, 4.7?
Don't want to go too new because #corps...
4.6.1 is the last version available on Win8 original....
So if we go to it, the very next bump cause another loss
My point really is that when we bump, we'll have to announce an OS loss. Better to have one such annoucement for Vista and 8.0, rather than one for Vista and potentially another for 8.0 some time later?
I would think that for VB6 apps that are still actively maintained, they're going to be running on some old ass machine or at least in a VM frozen in the time
I don't mind dropping support for dead EOL'd OSs (well, after my work laptop is upgraded to Win10 please =), but yeah that'll need an announcement. At one point or another, we'll have to drop them anyway
back to the point, I suspect that for VB6 apps still actively maintained (is that a thing?), we might find ourselves dropping support simply because they run in a time-frozen VM as opposed to installing VB6 on Win 10.
actually that begs yet another question - if you're actively maintaining VB6 apps, you'll need to make sure it's going to run on newer OSs --- otherwise what's the point of distributing it... right?
Worth noting too, that if and when we make the jump to Net Core, this problem evaporates as you can package a versioned CLR with your app. No more relying on having a given version on the user's machine.
Its a private CLR for your app, not a machine-wide install.
@this Heh yeah, it's a trade-off. You can still use machine-wide installs, but you don't have to. I suspect there'll be 2 versions of apps in the future, one with a private CLR and one without. Maybe even an installer thing.
Also, they're really going all-in on making the CLR modular. It'll be able to minify the private CLR to just the bits your app uses. They kinda have to do that for client-side Blazor, where the webpage has to DL the runtime on first page load...
There's some very cool stuff happening in Core right now...
VBA Rubberducking
VBA Rubberducking