« first day (1932 days earlier)      last day (1248 days later) » 
00:00 - 21:0021:00 - 00:00

Duga
Duga
12:00 AM
RELOAD!
[rubberduck-vba/Rubberduck] 7 commits. 2 closed issues. 10 issue comments. 176 additions. 204 deletions.
[Minesweeper] Games Played: 94, Bombs Used: 78, Moves Performed: 14161, New Users: 9
 
IvenBach
IvenBach
:barf: Just tried with dictionary replacement. No-go. Will start on late-binding tomorrow.
 
Mathieu Guindon
Mathieu Guindon
no go?
the dictionary drop-in is early-bound by default IIRC
 
Duga
Duga
> Parse tree filtering, just landed:

![Parse tree filtering](https://raw.githubusercontent.com/zspitz/ANTLR4ParseTreeVisualizer/master/parse-tree-filtering.gif)

Currently, the token list filter and the tree filter are independent from one another.

Should I also allow filtering the tree by token type as well?
 
IvenBach
IvenBach
Still getting reference errors. I have a lot of other non-default references too.
 
Mathieu Guindon
Mathieu Guindon
not sure if there's a precompiler toggle
yeah delete the Scripting typelib reference
now the Dictionary refs in your code point to the drop-in class, and making this Dictionary class all late-bound is the only modification you should have to do (maybe run the attributes by Rubberduck ;-)
 
IvenBach
IvenBach
12:09 AM
There's FSO that's also using it. It appears to have been the Dictionary that was the first result.
 
Mathieu Guindon
Mathieu Guindon
same thing: wrap it up
 
IvenBach
IvenBach
And I'll have to do that for the other 8 references too....
 
Mathieu Guindon
Mathieu Guindon
not if they're not a problem
 
IvenBach
IvenBach
That's a tomorrow task.
My caring jar is empty right now. Home time and homework already emptied it.</iven>
Thanks for the idears Mug.
 
Mathieu Guindon
Mathieu Guindon
later!
 
Hosch250
Hosch250
12:11 AM
See ya.
 
Mathieu Guindon
Mathieu Guindon
hey @Hosch250
 
Hosch250
Hosch250
What's up, duck?
</daffy>
How's it going?
 
Mathieu Guindon
Mathieu Guindon
seen @ZevSpitz's visualizer?
it's getting pretty damn nifty
 
Hosch250
Hosch250
Yeah, it's really cool.
I'm thinking about rewriting Manning's website for them :P
Their checkout is great, but man, their search system and product display system is bad.
Especially the search.
The product display is just an alphabetic list, which makes it hard to filter by relevant topics.
If I was to redo the search, I'd make it search by book title + description + chapter titles, author name, or raw content.
 
Zev Spitz
Zev Spitz
@MathieuGuindon @Hosch250 Are you using the visualizer regularly? What else can I improve?
Mathieu, you said you wanted to set the root of the treeview to a specific node; that's next.
 
Hosch250
Hosch250
12:19 AM
Sorry, I'm not.
I can't really develop RD at the moment.
It would work with any ANTLR grammar, though, right?
 
Zev Spitz
Zev Spitz
@Hosch250 Yes.
 
Hosch250
Hosch250
Nice. I have some of those, but I've not been working on them much.
I'll install .NET Core 3.0 when it comes out Monday and try to update RD
Maybe it can be one of my Hacktoberfest things.
And then we can go to C# 8 (and leave the nullable reference types feature off--for now).
:P
Then you can enable it project by project.
But you still get the index/range feature, and the default interface methods, and a couple other things.
 
Mathieu Guindon
Mathieu Guindon
12:34 AM
@ZevSpitz I tried it a couple of versions ago, but there was an exception and I forgot to note anything about it... I'll definitely use it a lot, as soon as I have a build that traverses the executable nodes :)
 
Mathieu Guindon
Mathieu Guindon
12:55 AM
and I think I just might have exactly that .... now
...well, anytime now I mean. swings sword
 
Hosch250
Hosch250
They failed to get the Brits to surrender and/or destroy the RAF as an effective force. In reality they don't seem to have had any real plans for an invasion. (And if they never had any way to land troops anyway, then why did the Brits bother with defending the coast...) — Tomas By 8 hours ago
Because they armed the coast during the Battle of Britain before they knew they'd won it...
You can't just put up defenses without time and thought.
Sometimes, people don't think before they post. (This applies to me too.)
 
IvenBach
IvenBach
1:41 AM
^ We all have our Pineapple moments.
 
IvenBach
IvenBach
2:00 AM
@MathieuGuindon @this @Vogel612 @Hosch250 any all the other ducks at the pond: dev.to/dvddpl/what-makes-a-10x-developer--1k0f is an article about you all.
Not individually, but what you represent to ducklings like me @puzzlepiece87 @FreeMan and the rest.
 
 
1 hour later…
Duga
Duga
3:18 AM
> These changes make the the inspection xml-docs much more flexible. - hasresult attributes (all attributes actually) are now matched case-insensitively, and using StartsWith, so hasresults and hasResults would both match. - Introducing support for a module node (which itself supports name and type attributes, respectively for the module name and component type e.g. "Class Module" or "UserForm" - content isn't validated at this point). Inspection examples that need multiple...
modules can now do this: xml <example hasresult="true"> <module name="Module1" type="Standard"> <![CDATA[ Option Explicit ... ]]> </module> <module name="Class1" type="Class"> <![CDATA[ Option Explicit ... ]]> </module> </example> Otherwise, the CDATA code block can also remain directly under the <example> tag. Although theoretically supported, inspection xml-doc examples should either all use a <module> tag, or none of them should. No need for a <module> tag if the module type...
is irrelevant or if there's only one module, however when the example is only applicable in a specific module type, consider including a <module> tag with a e.g. type="Class" attribute, even if there's only one single module involved. Now, does anyone knows enough CSS to make the module name & type look pretty?
[rubberduck-vba/RubberduckWeb] retailcoder pushed commit 541f4f7c to master: added support for modules and multiple code examples; fixed hostapp & refs, xml-doc attributes match case insenstive with StartsWith.
[rubberduck-vba/RubberduckWeb] retailcoder pushed commit 7ffb9c65 to master: fixed annotation comment
Merge pull request #72 from retailcoder/master

Inspection Details fixes
 
Mathieu Guindon
Mathieu Guindon
also fixes the reference and hostapp attributes: rubberduckvba.com/Inspections/Details/SheetAccessedUsingString
and now we can do this: 
<example hasresult="true">
<module name="Module1" type="Standard">
<![CDATA[
Option Explicit
...
]]>
</module>
<module name="Class1" type="Class">
<![CDATA[
Option Explicit
...
]]>
</module>
</example>
(the type attribute isn't used for now)
ttgtb
 
Duga
Duga
 
Duga
Duga
3:43 AM
 
IvenBach
IvenBach
Whoops. Couldn't resist clicking the checkBox...
 
 
1 hour later…
Duga
Duga
5:03 AM
> past-Iven would have wanted regions. current-Iven, working mostly in VBA-Dev-Landia and after the SoftwareEngineering link above, feels like they would be abused more than they help.

> I haven't missed it one bit. Especially after I found Ctrl-M-O in Visual Studio,

Having worked with small/simple code bases I feel lost/overwhelmed in any larger code base when seeing everything. Stepping up to the RD, which I still naively call a 'large' code base was tough. I agree with @mansellan that '
 
Duga
Duga
5:14 AM
> past-Iven would have wanted regions. current-Iven, working mostly in VBA-Dev-Landia and after the SoftwareEngineering link above, feels like they would be abused more than they help.

> I haven't missed it one bit. Especially after I found Ctrl-M-O in Visual Studio,

Having worked with small/simple code bases I feel lost/overwhelmed in any larger code base when seeing everything. Stepping up to the RD, which I still naively call a 'large' code base, was tough. I agree with @mansellan that
 
 
7 hours later…
FreeMan
FreeMan
12:17 PM
@IvenBach well put
 
 
2 hours later…
Duga
Duga
2:06 PM
 
Duga
Duga
2:24 PM
[rubberduck-vba/Rubberduck] build for commit 02da237e on unknown branch: AppVeyor build succeeded
[rubberduck-vba/Rubberduck] BZngr pushed commit 02da237e to next: Fixes xml-doc typo
Merge pull request #5154 from BZngr/4882_UCI_MsgRewording

UnreachableCaseInspection xml-doc typo correction
 
Duga
Duga
2:41 PM
[rubberduck-vba/Rubberduck] build for commit 11f458ad on next: AppVeyor build succeeded
 
this
this
somehow I just noticed now, after all those time.... the email title is exceedingly redundant....
[rubberduck-vba/Rubberduck] Pre-release Prerelease-v2.4.1.4952 - Rubberduck v2.4.1.4952-pre
 
Mathieu Guindon
Mathieu Guindon
so... we stop identifying prerelease builds as such, and leave the GitHub flag in charge?
 
this
this
I don't know. i think that problem exists only on the email subject
it isn't as bad on the releases page
having the pre makes sense, esp for the git log
it was just funny to me that the email subject kept emphasizing the "pre-release" due to how it was formatting and I never noticed until just now, that's all.
 
Mathieu Guindon
Mathieu Guindon
3:03 PM
:)
 
FreeMan
FreeMan
3:33 PM
Whelp, I called support to follow up on my support ticket. She's getting in touch with the support person who is supposed to be working on it to give me some follow up.
We'll see how that goes. :/
 
this
this
how very supportive!
 
IvenBach
IvenBach
Soon you'll be getting a call from their support line to support them with supporting your supporter. Support-all-around.
 
FreeMan
FreeMan
Frankly, I don't really care much who I hear from so long as I hear from someone.
#ThanksForYourSupport
oh, it would be good if they can come up with an answer, too.
oooohhh....
> Thank you for calling us at AppointmentPlus today. It was nice speaking with you.

To review, we discussed ticket 265959 and that you would like a follow-up. I will look into this further and advise the person handling your ticket to reach out to you.
I got follow up! To the phone call itself... :/
 
this
this
3:51 PM
Remember Monsters Inc? Their slogan was We scare because we care
For whatever reasons, that whole folllowup thing reminded me of that.
 
FreeMan
FreeMan
Interesting... Would you like to talk about that?
 
this
this
If I did, will you send me a follow up?
 
mansellan
mansellan
TFW git reflog saves your ass...
what did we do before git. ah yes, I remember. we lost code and screwed up merges.
 
this
this
well it re-flogs.
 
mansellan
mansellan
lol
 
this
this
4:02 PM
I know it's supposed to be "ref log" but I just cannot read it that way.
 
FreeMan
FreeMan
@this I will indeed!
 
this
this
Then, no, I don't want to talk about it....
 
FreeMan
FreeMan
@this thanks for that clarification. All I could think of was ye olde version of "beatings will continue until morale improves". Flog, reflog, reflog some more, morale improves!
3
 
this
this
IKR?
 
Hosch250
Hosch250
Yeah, people are savages.
 
FreeMan
FreeMan
4:08 PM
ooohhhhh... my ticket's "last activity" stamp is now "today at 11:38". Somebody's looked at it! No response still, but that's progress! Right?
 
Hosch250
Hosch250
You know the background of the word "decimation" is a military punishment for a cowardly/disobedient unit that had 1/10 of their members killed by lot (and the officers basically never made it out alive--either they were killed for initiating the retreat/mutiny, or they were killed by mutinous soldiers).
 
FreeMan
FreeMan
Which is why you can't "decimate" to the last man. You have to leave 90% of him or it isn't deci-anything.
 
Duga
Duga
> Currently, there are five inspections dealing with default member accesses: `ValueRequiredInspection`, `ProcedureRequiredInspection`, `DefaultMemberRequiredInspection` and `ObjectVariableRequiresSetInspection` dealing with failed default member resolutions and `ImplicitDefaultMemberAssignment` dealing with assignments to non-indexed default members.

I would like to propose to extend the coverage regarding successful default member accesses, but to give up the distinction between accesses an
 
IvenBach
IvenBach
4:29 PM
@this thank you again for hammering the idea of fully qualifying references. Made this janky workaround tolerable.
 
FreeMan
FreeMan
"last touched" timestamp updated again, but still haven't heard anything...
 
this
this
@IvenBach I'm still waiting for someone to chime in "oh, thank God for the implicit conversion!" (really, implicit anything).... Still batting 0.000 so far.
 
Hosch250
Hosch250
@this I love it.
decimal m = 0;
Instead of decimal m = 0m;
 
this
this
is that really implicit, though?
 
Mathieu Guindon
Mathieu Guindon
interesting
Decimation (Roman army)
Decimation (Latin: decimatio; decem = "ten") was a form of Roman military discipline in which every tenth man in a group was executed by his cohorts. The discipline was used by senior commanders in the Roman Army to punish units or large groups guilty of capital offences, such as cowardice, mutiny, desertion, and insubordination, and for pacification of rebellious legions. The word decimation is derived from Latin meaning "removal of a tenth". The procedure was a pragmatic attempt to balance the need to punish serious offences with the realities of managing a large group of offenders. ==...
 
Hosch250
Hosch250
4:34 PM
Yeah, it's an implicit conversion from an integer literal to a decimal value.
 
this
this
see, you have decimal right there
if you did var m = 0;, it'd be an int
 
Hosch250
Hosch250
Yep.
But if I do decimal m = 0; m = 2;...
 
this
this
so that's not surprising
 
Hosch250
Hosch250
And 2.0 is a double literal.
(And I dislike doubles for basically everything I work with, but then, I mostly use floating points for money and money-like values.)
 
Mathieu Guindon
Mathieu Guindon
    public static implicit operator Decimal(int value) {
        return new Decimal(value);
    }
 
Hosch250
Hosch250
4:37 PM
I've only ever written an implicit cast a couple times, and they were basically between dto/viewmodel types written in different languages.
I had my data structure coming from F# with its Option<T> and stuff, and I wrote an implicit conversion to the C# representation.
And then I decided it didn't actually work out too well anyway.
I changed them all to ToModel and FromModel methods.
 
this
this
@Hosch250 sorry that didn't make sense. Why would you even want to use floating points for money?
 
Hosch250
Hosch250
@this I don't. I use decimal.
 
this
this
that's not floating points.
 
Hosch250
Hosch250
Oh, sorry, I said the wrong thing.
I mean non-integer, basically.
 
this
this
@MathieuGuindon Granted but can you cite an example where one will go WTF?
@Hosch250 Ok, that makes more sense. Decimals are fixed points, I think.
 
Mathieu Guindon
Mathieu Guindon
4:40 PM
I'm not seeing any WTF situation here..
 
this
this
Whereas it abounds in VBA or JS or PHP....
see, even though Hosch's example be "implicit", it's not what I really think of as implicit because it's already right there in the code. It does exactly what you tell to do.
VBA, JS, or PHP, OTOH might just decide to give you "11" just for giggles
and I've seen people get all mucked up because they did something like SomeMoney / SomeInteger in SQL and get back a integer and go all WTH
as far as languages go, C# is by pretty far most explicit, even with the var and implicit operations.
 
Hosch250
Hosch250
FWIW, that will happen in C# too.
NVM, no.
That's just int/int I'm thinking of.
 
IvenBach
IvenBach
Decimation is certainly one way to incentivize an army to stay in line.
 
Hosch250
Hosch250
4:55 PM
No, only part of an army.
Typically what they do is trick them into surrendering their arms while other armed groups are drawn up around them (in, say, parade formation to ostensibly watch the punishment).
Basically, you have to have enough of your army prepared to fight the group to get them to submit to it and hope the lot falls "right".
 
this
this
would be kind of difficult to command the army "Fall upon your own sword!" and expect everyone to obey it.
 
Hosch250
Hosch250
^
Good way to get yourself killed.
 
this
this
Right. One officer dead, the whole army come home with a story. Who's going to find out?
 
Hosch250
Hosch250
Yep. Happens a lot in combat--especially when you aren't in a winning situation so morale is low.
Capture an AK-47 and kill an unpopular officer. Looks just like enemy fire.
 
mansellan
mansellan
@Hosch250 Implicit operators are quite useful for ValueTuples
 
Hosch250
Hosch250
5:00 PM
Typically if they are super stupid, like requiring salutes in combat situations (could reveal your position) or clean uniforms (dirt hides you better--and you don't have time to keep clean...).
 
this
this
You know, I honestly never understood the ways of army in the Napoleonic era.
everyone had their distinctive hats
and of course, everyone has to be dressed up pretty, never mind that they'll get killed, but they'll die fashionably!
 
mansellan
mansellan
@this I don't see the problem... The order doesn't say the sword can't be laying flat on the ground when fallen upon...
 
this
this
LOL
 
Hosch250
Hosch250
Well, that actually had its own uses too.
When you had black powder with limited range, it actually makes sense.
You had no radios. You moved based on visuals.
 
this
this
wasn't that what the drummer boys were for? Relaying orders?
 
Hosch250
Hosch250
5:04 PM
Partly, yes.
Color helped make sure you didn't walk into an enemy column, and it didn't matter anyway, because you were practically at point-blank range before you fired.
And they were actually trained to not aim before firing in most cases.
 
this
this
Seriously?
 
Hosch250
Hosch250
Fire as a volley, not at a specific target.
 
this
this
Seems a waste of single bullet they'll have
 
mansellan
mansellan
Quick-sniped.
 
Hosch250
Hosch250
And don't fire until the enemy fires--then you can run a couple steps and get closer.
And then your volley will be more devastating.
 
this
this
5:06 PM
and after that, it's basically win by bayonets
 
Hosch250
Hosch250
@this It was because the odds were they'd get close enough to miss anyone else, but not close enough to actually hit.
Which is partly why Bunker Hill was so devastating. The Americans were entrenched, and aimed before each shot, and waited until they were super close relying on the British not shooting first.
 
this
this
ah, hence "don't fire until you see the whites of their eyes"
 
Hosch250
Hosch250
Then the British line was so badly hit, instead of charging in before they could reload, they fell back to regroup.
Yes.
 
Duga
Duga
5:18 PM
> Just to see if this would work. I did not follow completely on the part about splitting by usual, recursive and unbound....

| Inspection | Default Severity | Notes |
| --- | --- | --- |
| `UseOfBangNotation` | Warning |
| `ImplicitDefaultMemberAccess` | Warning |
| `IndexedDefaultMemberAccess` | Hint |
| `ProcedureDefaultMemberAccess` | Warning |
| `RecursiveDefaultMemberAccess` | Warning | May be shown with other inspection(s) |
| `UnboundDefaultMemberAccess` | Warning | May be show
> Just to see if this would work. I did not follow completely on the part about splitting by usual, recursive and unbound....

| Inspection | Default Severity | Notes |
| --- | --- | --- |
| `UseOfBangNotation` | Warning |
| `ImplicitDefaultMemberAccess` | Warning |
| `IndexedDefaultMemberAccess` | Hint |
| `ProcedureDefaultMemberAccess` | Warning |
| `RecursiveDefaultMemberAccess` | Warning | May be shown with other inspection(s) |
| `UnboundDefaultMemberAccess` | Warning | May be show
> I think the distinction is useful and warranted, and the more inspections we have, and the more granular they are, the better IMO.

I'm not sure I understand though - does that mean we would be removing an inspection that explicitly calls out a missing `Set` keyword?
> First, the second split would yield `ImplicitDefaultMemberAccess`, `ImplicitRecursiveDefaultMemberAccess` and `ImplicitUnboundDefaultMemberAccess`, etc.

Second, the `ProcedureDefaultMemberAccess` is for procedure coercions, i.e. if the default member is used because you used an object variable in a call statement instead of a procedure, function of property.
> First, the second split would yield `ImplicitDefaultMemberAccess`, `ImplicitRecursiveDefaultMemberAccess` and `ImplicitUnboundDefaultMemberAccess`, etc.

Second, the `ProcedureDefaultMemberAccess` is for procedure coercions, i.e. if the default member is used because you used an object variable in a call statement instead of a procedure, function of property.

Third, the `ImplicitDefaultMemberAssignmentInspection` does not advise to use `Set`. That is the job of `ObjectVariableRequiresSetI
> I think in this particular case getting 2 inspections make more sense than getting 1 very specific inspection, IMO.

I also think `ProcedureDefaultMemberAccess` may need some clarification because that's not what I think of when I see code like `foo(bar)` (where `bar` is an object variable with a default member, and `foo` takes a value type for the parameter). `ImplicitDefaultMemberAccessInProcedureCall`, maybe? Too wordy, though.
> I think you gat me wrong regarding procedure coersion. What I mean is what happens to `bar` in the following example.
```vba
Public Sub Foo()
Dim bar As Class1
Set bar = New Class1
bar
End Sub
```
The last reference to `bar` is a call to the default member of `Class1`.
> I think you get me wrong regarding procedure coersion. What I mean is what happens to `bar` in the following example.
```vba
Public Sub Foo()
Dim bar As Class1
Set bar = New Class1
bar
End Sub
```
The last reference to `bar` is a call to the default member of `Class1`.
> I think you get me wrong regarding procedure coercion. What I mean is what happens to `bar` in the following example.
```vba
Public Sub Foo()
Dim bar As Class1
Set bar = New Class1
bar
End Sub
```
The last reference to `bar` is a call to the default member of `Class1`.
 
M.Doerner
M.Doerner
5:42 PM
@Duga Sorry for spamming by fixing typos one by one.
 
this
this
@Duga Ok, that's originally what I thought it was, which goes to the original point I was trying to make. For non-user objects, whether it's a foo.Value or foo.GetValue() hardly matters to the user; they don't have a choice either way and the QF would be the same... right?
It would only matter if it was their own defined class and they can control the default member assignment since assigning a procedure as a default member should be a huge smell.... right?
 
M.Doerner
M.Doerner
The quickfix is always the same, except for the unbound ones, which don't have one.
For procedure coercion, it is irrelevant whether the default member is a procedure, property or function.
My opinion is that using it will be an error in 99% of all cases it is used.
 
this
this
So you want the procedure coerecion to have an error as the default severity, while the equivalent implicti default meber access be warning?
 
Hosch250
Hosch250
6:30 PM
Ugh, SO...
Or rather, Business Insider...
> Programmers love Google more than Apple, but dread Microsoft according to 100,000 developers
> It finds that Google's Go programming language is a hair more "loved" than Apple's Swift, but that Microsoft's Visual Basic 6 is still the "most dreaded."
Way to write a misleading title...
 
this
this
.... WTF?
 
Hosch250
Hosch250
Then of course there's:
> Millions of programmers rely on this site to do their jobs, and its CEO has big plans to make it even more helpful
I guess BI loves clickbait.
I won't be clicking any more of their links, I guess.
 
this
this
It's kind like holding a beauty pagenant and having a 60 years old run against a bunch of 20-somethings.
 
Hosch250
Hosch250
Also, according to SO's salary calculator, I'm worth roughly $120k/yr.
I'm not quite so sure...
(But then, maybe I'm over-rating myself, since their standards are supposedly super strict, and my team isn't?)
 
Mathieu Guindon
Mathieu Guindon
@Hosch250 that's if you worked at SO
 
Hosch250
Hosch250
6:43 PM
Yeah, I can't see myself there. I think the company has turned from a star into a cash cow.
I like working at small stars. They are intense and fun.
 
Mathieu Guindon
Mathieu Guindon
@Hosch250 well, Craver's gotta pay all these LEGOs!
 
Hosch250
Hosch250
LOL.
Nothing against cash cows, but they tend to rest on their laurels.
It's perfectly a valid business strategy, but it's just not my personality.
 
Duga
Duga
> Closes #3962. The issue was actually corrected by a previous PR (probably #4127) and a unit test linked to the issue already existed. However, while reviewing the test, I realized that it could possibly generate a false-positive (passing) result. So, this PR simply identifies that #3962 is resolved and includes only a modification to the unit test.
 
Duga
Duga
7:09 PM
[rubberduck-vba/Rubberduck] build for commit 92c86ca6 on unknown branch: AppVeyor build succeeded
Improved AdditionString test

Prior test would have masked an incorrect result
Merge pull request #5156 from BZngr/3962_UCI_StringMath

3962_UnreachableCaseInspection
 
this
this
Unrelated: Chris McClellan commented about using zero bug process few days ago. It made me wonder - should RD be doing that, too?
 
Vogel612
Vogel612
with the amount of heisenbugs that COM interaction with the VBE entails ... I doubt it's doable
 
Mathieu Guindon
Mathieu Guindon
@this only 169 open
 
Duga
Duga
 
Feeds
Feeds
 
Duga
Duga
7:24 PM
 
this
this
@MathieuGuindon yeah, hence the question
 
Duga
Duga
[rubberduck-vba/Rubberduck] build for commit 236492e4 on next: AppVeyor build succeeded
 
this
this
also, with hack-o-fest (can only speel right now) coming up, it might make sense to make it a bug bash.
 
Mathieu Guindon
Mathieu Guindon
I don't think zero bug is going to happen, but yeah we should absolutely make a bug-bash-fest
 
Duga
Duga
> 1) We cannot reparse willy-nilly. It's just too slow for many of our users still (unfortunately).
2) It should not be added multiple times. This should be addressed.
3) We can retrigger the inspections, or maybe create a system to invalidate inspection results (maybe the inspection should get a new member `IsResultInvalid(Result, Data)`).
 
Vogel612
Vogel612
7:34 PM
@Duga @Hosch250 we already do have a mechanism for invalidating inspection results, IIRC
 
Hosch250
Hosch250
Do we?
That's news. Where have I been...
Basically, though, he's right. It should give some indication of having worked, but a reparse isn't needed.
 
Duga
Duga
> 1) We cannot reparse willy-nilly. It's just too slow for many of our users still (unfortunately), and (shouldn't) be needed to update any results returned over this name.
2) It should not be added multiple times. This should be addressed.
3) We can retrigger the inspections, or maybe create a system to invalidate inspection results (maybe the inspection should get a new member `IsResultInvalid(Result, Data)`).
 
this
this
@MathieuGuindon I take the point that some bugs may be very hard to fix. If not zero bug, what is a good goal to measure the quality of the product?
e.g. looking at # of open bugs = 169 ≠ poor quality, apparently.
 
Duga
Duga
> Is this still occurrring? No repro here with .4942
 
Mathieu Guindon
Mathieu Guindon
@this it's probably closer to 150 (if not fewer) if we close the stale ones and the dupes. also, if there's only 1 bug and it's "blows up on startup", # of open bugs isn't really an indication of anything
 
this
this
7:47 PM
Well, yes. However, as I understood it, the whole point of ZBP is that no product should ship with known bugs, effectively prioritizing bug fixes over any enhancements
 
Mathieu Guindon
Mathieu Guindon
is implementing CPA to fix an inspection false positive a bug fix or an enhancement?
 
this
this
But the heisenbugs makes me wonder how other projects handle a similar situation.
See, exactly.
 
Hosch250
Hosch250
I tried to do ZBP over my internship.
And over other time periods.
 
Duga
Duga
 
Hosch250
Hosch250
It can be kind of draining, because you keep fixing bugs that are being found, while other people who aren't on board with it are adding fun new features.
 
Duga
Duga
7:49 PM
> No repo in .4950 either.
 
FreeMan
FreeMan
BugCount--
3
 
this
this
@Hosch250 Yeah, there's that. As Mat pointed out earlier, introduciing a rearchitecture would easily make several bugs irrelevant.
It belatedly occurs to me that Extract Method perhaps even get easier with enhanced CPA
 
Hosch250
Hosch250
That's what I've repeatedly said.
 
Mathieu Guindon
Mathieu Guindon
@this totally
 
Hosch250
Hosch250
IMO, there are some cases that can't be handled without that.
 
this
this
7:50 PM
I should ask Chris about those type of projects. Would he ZBP the RD project?
 
FreeMan
FreeMan
@this while simultaneously unintentionally introducing a new collection of bugs neatly mounted and categorized
 
Mathieu Guindon
Mathieu Guindon
FWIW I'm building CPA with the CPA issues-list in mind, and EM is definitely part of it
 
this
this
@FreeMan oh yeah. I'm expecting that mocking framework will create 500 new bugs. :D
and that'll be in spite of the fact taht I've already written 400+ unit tests.
 
Hosch250
Hosch250
TBH, though, at this point...
 
Mathieu Guindon
Mathieu Guindon
dang
 
FreeMan
FreeMan
7:51 PM
will they be neatly mounted on little pins in a display case with carefully hand-typed labels identifying each one?
 
Hosch250
Hosch250
What would it take to break RD up into a couple projects?
Say, unit testing in one.
 
this
this
?
 
Mathieu Guindon
Mathieu Guindon
@Hosch250 it's 19 projects atm
 
this
this
^
 
Hosch250
Hosch250
No, not that.
 
this
this
7:52 PM
well, 22, actually
 
Hosch250
Hosch250
The "other" projects.
 
Mathieu Guindon
Mathieu Guindon
the "github" projects?
 
this
this
not sure what you mean by projects....
 
Hosch250
Hosch250
More "products".
 
FreeMan
FreeMan
Dad got his BS, MS & PhD in biology - I spent waaaaay too many years hanging around university biology departments
 
Mathieu Guindon
Mathieu Guindon
7:52 PM
"features" then
 
Hosch250
Hosch250
For example, unit testing doesn't really need any of the other stuff.
The grammar, inspections, etc.
Eh, I'd say not worth it.
It all ties in so well together.
Since everything needs the code pane wrappers, and almost everything needs the grammar.
Even the test pane discovers tests with the grammar, I think.
 
Mathieu Guindon
Mathieu Guindon
technically the test pane/toolwindow doesn't discover anything, but yeah
 
IvenBach
IvenBach
@Hosch250 at least you aren’t fixing bugs that were already fixed because you can’t git merge foo/bar into and Excel workbook...
 
Hosch250
Hosch250
LOL, sorry.
TBH, I think it was a mistake dropping SC when we did.
I had it really close. It worked for me...
And it even gave sensible error messages most of the time, if you knew Git.
You could initialize a local repo from a remote (I even did it with RD to test, since Git doesn't care what the repo contains).
You could create a local repo, then push it to an empty remote (tested with GitHub and BitBucket).
You could do push, pull, merge, commit, etc.
 
Mathieu Guindon
Mathieu Guindon
meh. we hardly had 100 opened bugs, and more than half were SC
 
Hosch250
Hosch250
8:06 PM
Most were probably duds.
And some probably were simply due to lack of training.
 
Mathieu Guindon
Mathieu Guindon
roadmap wiki: "There are well over 250 open issues and as many awesome ideas to turn into awesome tools for the VBA community!"
bwahahaha
@Hosch250 yeah... ish. if software blows up because you're using it wrong, the problem isn't with the user...
 
Hosch250
Hosch250
Agree, but the core functionality was there and tested.
The cleanup would've been light work.
OTOH, we could start another project with just it.
We could use what we've learned from the safe code panes and add/remove listeners, and take the logic from the history.
I'd re-build the UI anyway, that was a tiny bit of a mess.
 
mansellan
mansellan
3 sleeps till Core 3!!!
 
Hosch250
Hosch250
And what sleeps they better be.
I feel like I could faint, I'm so tired :( Worried about driving home.
And yes, it's my fault.
I'm a Reddit-aholic.
Which is funny, because I do the exact opposite. I have Amazon up on my broswer as I shop retail stores, so I can read reviews and get technica details bot on tge package. Having fleeced Amazon for all this data they gave away for free, I then buy the item retail. I want the item, not a wait. — Harper 3 hours ago
Ohhh, I should write an app for that.
Basically look for an item on Amazon/Walmart/whatever and show reviews and ratings.
And maybe have a "click here to buy" at the bottom so Amazon doesn't make me stop :P
And maybe they'd let me actually use their APIs if I did that, instead of web scraping.
 
IvenBach
IvenBach
Enjoy the video Mug.
 
mansellan
mansellan
8:20 PM
@Hosch250 heh, he's not heard of Amazon Now :-)
 
Hosch250
Hosch250
LOL, but why wait 30 minutes?
 
mansellan
mansellan
(not that I can enjoy it, i'm an hour's train ride from London...)
 
Hosch250
Hosch250
And deal with that shady, tax-dodging, identity-stealing company?
 
mansellan
mansellan
heh yeah, there's that...
 
Mathieu Guindon
Mathieu Guindon
@IvenBach lol
 
IvenBach
IvenBach
8:30 PM
Hopefully that starts your weekend off right.
 
mansellan
mansellan
well nuts, I missed the cutoff for enrolment in this semester (I can speak murican!)
guess my sabbatical will be a little longer than planned
 
Hosch250
Hosch250
@mansellan :(
Good luck, I guess.
Also, pro-tip. Hacktoberfest doesn't have good validation on their form.
 
mansellan
mansellan
meh, my head's not in the right place for study atm anyway
 
Hosch250
Hosch250
My phone was acting up, so I wasn't completing sure it properly submitted the form last year.
I re-did it on my computer, and I got 2 shirts :D
 
IvenBach
IvenBach
@MathieuGuindon Next round of RD gear should include a baseball hat with a yellow brim for a ducks bill.
 
Mathieu Guindon
Mathieu Guindon
8:36 PM
baseball hats were too expensive last time though
 
Hosch250
Hosch250
I already get enough weird looks about my Hacktoberfest shirts, LOL.
"What's Hacktoberfest? Is that the thing they do in Vegas?"
 
Vogel612
Vogel612
I actually see the latest hacktoberfest shirt semi-regularly at my uni
 
Hosch250
Hosch250
"No, it's a programming thing where DigitalOcean, a web hosting company, <eyes glaze> partners with GitHub, a code hosting company, to <looks away> advertise their product."
@Vogel612 Ah, yeah, programmers at uni would know.
Not random joes, though :D
The stuff in <> are people's typical reactions.
 
mansellan
mansellan
This looks pretty nice, I can see myself using that at work...
Enhances the simple Net Core IoC container with assembly-scanning abilities
 
Hosch250
Hosch250
@mansellan That's awesome.
 
mansellan
mansellan
8:46 PM
As soon as 3 lands, I'm gonna push for work to align with the Microsoft abstractions, especially in IoC, Configuration and Hosting. They're very good indeed - arguably not the absolute best in class, but there's big gain in keeping to the standard recipes if they're up to the job.
 
Hosch250
Hosch250
Yeah, I love them.
They are a bit different than 2.2, though, as I quickly learned updating a PoC project.
 
mansellan
mansellan
Atm, we're using DryIoc, but wanna move to MS
 
M.Doerner
M.Doerner
@Vogel612 The t-shirt from two years ago definetly looks better, though.
 
Mathieu Guindon
Mathieu Guindon
yeah I like the blue one better too
 
M.Doerner
M.Doerner
@Hosch250 SC had serious problems because of limitations of the git provider we used. E.g. it did not work in a quite common corporate situation, remotes on a network drive.
 
Hosch250
Hosch250
8:51 PM
@M.Doerner It's the same one VS uses.
Not sure why ours didn't work in that situation.
 
M.Doerner
M.Doerner
@this I wanted to make the procedure one and the unbound a warning, the implicit one a suggestion and the indexed one a hint.
 
00:00 - 21:0021:00 - 00:00

« first day (1932 days earlier)      last day (1248 days later) »