[rubberduck-vba/Rubberduck] IvenBach pushed commit 182db688 to next: Clear resource string to flag for translation

[rubberduck-vba/Rubberduck] IvenBach pushed commit 69e5d1e2 to next

[rubberduck-vba/Rubberduck] IvenBach pushed commit f7add1a9 to next: Merge remote-tracking branch 'upstream/next' into Issue2777_CleanedUpCommits

[rubberduck-vba/Rubberduck] web-flow pushed commit 7b8a285f to next

[rubberduck-vba/Rubberduck] retailcoder closed issue #2777: Code Inspections: Grouping By Location - State what source is

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #2777: Code Inspections: Grouping By Location - State what source is

> Closed in #5263

RELOAD!

[rubberduck-vba/Rubberduck] 73 commits. 4 closed issues. 2 issue comments. 22178 additions. 18226 deletions.

[Zomis/GetMotivated-vue] 2 commits. 1 addition. 3 deletions.

[Minesweeper] New Users: 13, Games Played: 47, Bombs Used: 33, Moves Performed: 5966

[rubberduck-vba/Rubberduck] rubberduck-vba-releasebot created draft prerelease Prerelease-v2.4.1.5194

[rubberduck-vba/Rubberduck] rubberduck-vba-releasebot prereleased draft prerelease Prerelease-v2.4.1.5194

[rubberduck-vba/Rubberduck] rubberduck-vba-releasebot published prerelease Prerelease-v2.4.1.5194

[rubberduck-vba/Rubberduck] rubberduck-vba-releasebot created tag Prerelease-v2.4.1.5194

[rubberduck-vba/Rubberduck] build for commit 7b8a285f on next: AppVeyor build succeeded

@Duga Two questions: first, why merge before the formatting is fixed? Socond, do we also want the component type on the location group header? To the second, I never got any answer and it is not included in the PR.

[rubberduck-vba/Rubberduck] MDoerner created pull request #5276: Replace tabs by spaces in XAML to merge MDoerner/Rubberduck/CorrectFormatting into rubberduck-vba/Rubberduck/next

> …where they have been introduced in PR #5263

[rubberduck-vba/Rubberduck] build for commit ffe8d09d on unknown branch: AppVeyor build succeeded

[rubberduck-vba/Rubberduck] mansellan created comment on pull request #5274: Follow up on SC style import commands

> I've not had a chance yet to fully review this PR, but it seems that the import assumes that certain component types must always provide binary files alongside the primary (declarations and code) file. This is not the case - binary files will only be present where there is a need for them based on the properties of their child controls. If a component does not have at least one child control which has a binary-content property set, a binary file will not be produced on export and will not be i

@M.Doerner sorry to be the bearer of bad news...

@M.Doerner I can double-check, but I would definitely assume not, presumably any control which cannot serialise to text will emit binary. I vaguely recall reading an article about how component serialisation works, perhaps in the VBA SDK. Will try to dig out a link.

[rubberduck-vba/Rubberduck] MDoerner created comment on pull request #5274: Follow up on SC style import commands

> The assumption is not that some components always need binary files (except for user forms, which always have a binary, even when empty). The assumption is that there are components that might need binary files. These types are excluded from update unless an implementation of IRequiredBinaryFileExtractor exists that supports them.

[rubberduck-vba/Rubberduck] mansellan created comment on pull request #5274: Follow up on SC style import commands

> Ah ok cool, that was my misunderstanding from a quick scan of the code in a web browser.

[rubberduck-vba/Rubberduck] retailcoder merged pull request #5276: Replace tabs by spaces in XAML from MDoerner/Rubberduck/CorrectFormatting into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] MDoerner pushed commit ffe8d09d to next: Replace tabs by speces in XAML where they have been introduced in PR #5263

[rubberduck-vba/Rubberduck] web-flow pushed commit 98051418 to next

[rubberduck-vba/Rubberduck] rubberduck-vba-releasebot created draft prerelease Prerelease-v2.4.1.5196

[rubberduck-vba/Rubberduck] rubberduck-vba-releasebot prereleased draft prerelease Prerelease-v2.4.1.5196

[rubberduck-vba/Rubberduck] rubberduck-vba-releasebot published prerelease Prerelease-v2.4.1.5196

[rubberduck-vba/Rubberduck] rubberduck-vba-releasebot created tag Prerelease-v2.4.1.5196

[rubberduck-vba/Rubberduck] build for commit 98051418 on next: AppVeyor build succeeded

sup

Late night. Most other ducks have settle down for the night @Jelly

@MathieuGuindon I feel for you too bro! We do encourage positive psychological safety in programming practices culture you know. :) so thumbs up as @M.Doerner use mediation for 5 to 20 mins to relax and unwind. Perhaps under the Getting things done philosophy of David Allen’s system shift one task of RD a week instead of several in one day. Take care.

[rubberduck-vba/Rubberduck] Irubataru created comment on issue #5265: Vertical scrollbar in code inspections should not include the top and bottom information

> Probably not a surprise, but the test explorer has the same issue.

[rubberduck-vba/Rubberduck] Irubataru created comment on issue #5053: Excel freezes when running test during refresh

> Although I am sure you already know I just wanted to report that it works here as well. Thanks! 😄

@MathieuGuindon you're becoming a manager here, you know? From someone struggling with that beast since about a decade: the trick is to celebrate the small victories and forgive yourself for things beyond your control

Also remember that not only coding is work :)

[rubberduck-vba/Rubberduck] daFreeMan created comment on issue #5269: Code Inspections not highlighting code on double click

@MathieuGuindon You know, that's one of the reasons why I switched jobs.

Wasn't ready to settle down and just do nothing yet.

On the other hand, it sounds like you might be getting SAD a bit early (seasonal depression).

@Hosch250 Or, as Iven pointed out, SAD as in hanging out in SO too much. ;)

I get that yearly, but it starts in January for me, typically. One thing I find useful is to set a small specific goal I know I can reach.

Then, once I get started and complete it, I feel better and tackle something else.

And a series of successes can do wonders emotionally.

Good luck solving it, either way.

@Duga I think the issue was in .5096 and was resolved in .5097 "Fix untyped function inspection". So, weird as it looked, I believe all is good! (not sure where to look in the release notes to see what PR(s) that build included and what issues they fixed, but I do remember something about that now that I've dug into it.)

keeps screen grabs handy for a bit, just in case...

[rubberduck-vba/Rubberduck] MDoerner created comment on issue #5269: Code Inspections not highlighting code on double click

UGH! another silent excel crash on exit. :/

@FreeMan On the releases page, there is a link to the build on AppVeyor. There, it tells you the title of the PR.

I don't think RD has anything to do with the crash. When Excel comes back up, it gives me the "you just double-clicked excel.exe, what would you like to do now" start screen", not the "hey I puked on you, here are the remains of all the files you had open when I did". So I think Excel has pretty well shut down, it just can't fully take care of itself...

I think in the last version of Excel, the start screen is always the same.

It only tells you about the remains of previously open files when you open another file.

@M.Doerner Ugh... I was looking at this: Enabled horizontal scrollbar in the Code Inspection window. Closes #5…, not this: Merge pull request #5231 from glowingrunes/5135-HorizontalScrollBarCodeInspection. Forest/trees, needs moar coffee, #ReadingFail, etc., etc., etc.

@M.Doerner that's my point, though. Previously, when excel crashed with files open, it would relaunch itself and reopen the files that were open when it crashed. I presume the behavior is the same now, and it's not reopening files for me, so I'm assuming (with all inherent danger) that there were no files left open when it managed to die.

I'm also presuming/assuming/hoping that RD has managed to completely exit by that point, so I'm not pointing a finger at/blaming the duck at all for these crashes.

Of course, I've rarely got Excel open without the duck open, so it'll be hard (for me, at least) to prove the negative.

[rubberduck-vba/Rubberduck] bclothier commented on Rubberduck.CodeAnalysis/Inspections/Abstract/QuickFixBase.cs of pull request #5274: Follow up on SC style import commands

@Duga this might end up beyond the PR's scope but the JunkDrawer does not seem that it'll go away anytime soon. We may need to clearly define the project's scope, especially WRT extensions which can devolve into unfun game of hide-and-seek to find the extension's namespace.

what's wrong with putting enum extension methods in the same namespace as the enum being extended?

i.e. if you can use the enum, you can use its extension methods

I presumed because it's used by more than one projects, and enumerables aren't our defined types.

that is, it's from System.Collection.whatever, and we're extending on it. Where do we put it?

oh wait it's the vbext_ct_ComponentType enum?

yeah, then it's ...not so simple :)

After previous discussions, it seems the best way to launch an SSIS package from my VBA code is to call a stored proc that will launch the package. I'm good with that. Any tips on how to check on the results of the SSIS package so I can log the results?

From what i see, only the EnumerableExtensions was moved.

But several files didn't get the old using removed, so it makes me question whether it really needs its own namespace.

@FreeMan pretty sure Ola's script will help you out

@this Ola... wracks brain... nope, nuttin'

ola.hallengren.com

^ bookmarked. reading...

can't remember where he mentions running a job

but I thought I saw he had a script for that somewhere in the site

and it handles the logging, too

if nothing else, you can use his backup script as an inspiration for how to get the logging done

ah, gotcha. I was looking for a way to query the results & pass it back as the output of the SP, then logging it with my VBA. I guess if I can log it in a logging table then read that back, that'll work, too.

[rubberduck-vba/Rubberduck] MDoerner commented on Rubberduck.CodeAnalysis/Inspections/Abstract/QuickFixBase.cs of pull request #5274: Follow up on SC style import commands

@this I moved the extensions because you cannot use them from VBEditor if they are in Parsing.

@MathieuGuindon i'm thinking that for the function array post it's not replicatable for SO due to possible specifics on a single machine... goign to vote to close as such, but still not sure

@Cyril this part irks me the most:

the function OP is showing us isn't using the num parameter, so whatever you pass the function, shouldn't make any difference whatsoever

i told him "@PulkitJain In reading the other comments, and specifically seeing John Coleman's comments on possible corruption or even a phantom bug, I would wonder if on a rudimentary point, if you were to fully declare and test the fully-declared code on the specific computer, would it run? Whether or not it runs does not determine why the current issue is happening, but would at least give a path forward"

similar to your point, we're not seeing the actual code, and thus can only guess... at the code changes AND the system settings/parameters

also, there're better choices for returning a column label from index

they're returning a hard-coded string literal and then claiming there's some logic to it

I'm this close to downvote and consider it a troll post

i'll stick with the vote to close as typo/not reproducible... the simplest solution is to declare all variables, which should correct your pain points... i think you and i both hit on that in the comments to the main post

F it, did it

LOL

if nothing else from that post, i learned something from vityata's post... didn't know you could redim if you haven't previously dimensioned

unsure if that's related to not using option explicit or it's actually solid

no, it's legit

It may be legit but it ought to be illegitmiate.

^

I can't recall, did we get that fixed in RD? this used to confuse the resolver..

...a bastard of coding

I think so, yes.

checks

@Hosch250 it's november... did the new gig start?

no, it's no longer confused.

but.... I am not sure about the # of references. It seems to be counting the y and y(x) as 2 separate references

that is, if I dbl-click the first reference, it highlights only the y, double-click the 2nd reference, it highlights the y(x) instead.

So we get 3 extra references.

@MathieuGuindon <whispers>Stay within the safety net of CR. Better questions. Posters usually know enough to explain properly.</whispers>

@IvenBach but.. I'm slipping!

I guess probably because the expression y(x) has a subexpression for y (the LHS), and both expressions get picked up.

@MathieuGuindon If it helps you keep your sanity slip all the way down to where plebs like me are. You'll have company that way. It's lonely at the top from what I hear.

@this could this be a problem with multiple rewrites to the same token? can you rename y?

Feel better knowing the fact I've given 19 answers but only have 13 points. Less that 1 point per answer.

does nothing now

will post log after mtg

verified the non-redim'd y does rename OK

@IvenBach I notice less upvoting on certain tags... i believe someone was explaining "if it's excel or VBA tagged, [the answer] should already be out there... you're just helping someone see it how they want to"

@Cyril there's a lot of truth to that. it's pretty rare to find a VBA question for which an existing answer doesn't apply at all.

and more than half the time, it's something Rubberduck would have warned them about.

there are some places that will not allow you to have any additional IDE tools/addons, so you're stuck with what microsoft provided... even with that in mind, i feel that the majority of posters are first time folks who have improperly posed inquiries, or really do need hand holding, which SO is the inappropriate forum. hell, look at my first few questions ever asked (which i think you answered a few of them) where i just don't know the correct language.

from that perspective, SO is a learning site, like w3school is for html, just with fewer gross perspectives, inherently making it harder for a first timer to learn

sorry, i'm a block of text kind of person... my wife hates texting with me

lol

@Cyril it is, and that's sad because it means it failed its mission, and is basically becoming just another Facebook Excel/VBA group page, or subreddit, or Quora ...except with half-decent answerable questions at least half the time.

"online help" forums kill me. every single one of them is full of "I need XYZ, can someone do it for me because I'm not willing to make the effort to learn anything to do it myself"

at least on SO these get shot down pretty quick

doesn't help that Microsoft insist on using their crappy NIH forum software either.

@MathieuGuindon not to belittle the mission, but hasn't that changed over time? i believe the mission was redefined when Joel left?

...despite what Meta wants to believe

at this point I don't know/care what the mission is... I'm just answering questions where I can see an OP that wants to learn stuff. If the person just wants quick copy-pasta crap code they won't understand, I'm leaving it to others.

copy

it's just.. the broken-record syndrome hits at one point or another

that's a large part of why I mostly stopped answering on CR

don't get me wrong, I love CR

related to CR, that's still considered "under development" right? it's been, what, 4 years since it was started?

CR is fully-grown :)

launched in 2011 IIRC

we got it out of beta in 2015

let me see if i can find teh post on SE I made about it, but I was under the impression that 2015 was the "birth" and the fact that it was under development was specifically why mods refused to put it as a "site more appropriate" in the reasons to close section, as of last year

you're talking about migration paths

CR graduated from beta mid-2015 (we had our first elections in June of that year, that's when I was elected)

regardless of my pedanticism, CR has been helpful the few times i've been there. learning some basic "streamlining" tips helped immensely

before that, a SO->CR migration path was deemed inappropriate because the site was beta. since then it's still inappropriate, but only because SO people have no idea what's on-topic on CR and if there was a direct migration path we'd end up with tons of crap and rejected migrations on CR

copy

@Duga is scraping all SO comments that mention CR, she's posting them in The 2nd Monitor

@this You get both references for the array itself and the array access. This is intentional.

@MathieuGuindon to outpedant a bit here: There was an intentional technical limitation in place to disallow migration paths from graduated to beta sites.

It used to be that the list of migration targets for users was automatically filled with the top migration targets

If the toolwindow is supposed to show only the references to the array itself, it has to filter to IsArrayAccess == false.

and mods can migrate to any site they want

@this Completely unrelated, thanks for writing that article on using OASIS-SVN together with git. It was rather helpful.

Glad to know it was helpful!

RE; the redim issue - I'm not sure it's that.... because if I add Dim Y, the reference drop to 5

comment out the Dim y, then references goes to 8

and I confirm that I can rename when there's Dim y there

the refcount on the RD toolbar might be confusing, too

it reports the same number

...well, sorry if my question put you all into a rabbit hole... the bastard-code redim arr(x) appears to be a mess in general, even without the RD references

@Cyril yeah.. the thing is, if it's legal VBA code to treat ReDim statements as basically "executable declarations", then RD has to treat them as such, too

I'll open an issue

IIRC With blocks had a similar issue at one point

(every . dot would show up as an identifier reference)

(which is correct and needed for RD to do its thing, but confusing for the user)

@this What happens if you declare it as an array?

Ok, will incorporate into my issue. still writing it up

I think the array access is not recognized on a Variant because it could be a default member call.

[rubberduck-vba/Rubberduck] bclothier opened issue #5277: Array & undeclared Redim'd variables' refcounts and renaming are incorrect

[rubberduck-vba/Rubberduck] bclothier added label bug to issue #5277: Array & undeclared Redim'd variables' refcounts and renaming are incorrect

@M.Doerner you were correct about it being a problem w/ array. The symptoms is present even there so I included it in the bug report.

@MathieuGuindon ah, the vanity of chasing unicorn farts and internet points...

;)

@MathieuGuindon and the other half is something Rubberduck will warn them about shortly!

@FreeMan i mean... internet points rule the world; it's like a social-rating system... just hope i can still get a flight if my number isn't high enough

@M.Doerner I don't seem to have that one bookmarked. Care to reshare the love?

if you google his name, it's the first hit (i had to look, too)

accessexperts.com/blog/2019/07/02/…

@Cyril first link I got was an ad. 2nd was for his LinkedIn profile.

That is the link.

What annoys me a bit is that there is no good way to handle views.

We have got tons of them on our Access applications, but apparently the Access export handles them in a really stupid way.

@FreeMan my bad... i forget that google can show different results

If you build them via the designer the export will be completely different from when you insert or modify the SQL directly.

if you don't like the vanilla SQL editor for Access queries, you could look at this cheapo tool. However, that can complicate import/export because metadata are stored somewhere else

fieldeffect.info/wp/access-sql-editor-documentation

The editor is not really the problem.

That editor will reformat the SQL

It is just a PITA for SC that the export is not the same as for the entirely equivalent SQL.

but an Access query includes more than just SQL.

like, parentheses. lots of parentheses.

:)

that goes into SQL. :)

I know that the view actually contains the layout.

I'm referring to metadata stuff like window layout, caption, format, whatever. That also gets get exported.

That is not what I am complaining about.

My problem is that the query part is not encoded consistently.

Wouldn't that happen whenever you open the query, and make changes, though?

If you keep it in SQL view, it won't do that but if you switch to design view and do something, it'll rewrite the entire SQL. Is that what you mean?

If you always insert the sql directly, you get a somewhat sensible diff. If you always use the designer, it also makes sense. Just mixing them produces a stupid diff.

Yes, that's true. That said, I believe if you use the addin you don't have to deal with that SQL/Design views inconistencies.

The problem comes down to who gets to rewrite the SQL.

in SQL view, you're the one rewriting it. In Design view, Access gets to do it. And obviously you two won't agree.

I will have to check again, but AFAIR, when I exported a query written via the designer, it literally encoded my choices in the designer and no SQL.

@MathieuGuindon :waves: I still want to learn. Anxiously awaiting your book :poke:

If the query comes from the designer, you get something like the following.

Begin InputTables
    Name ="SystemInfo"
    Name ="SignatureInfo"
    Name ="SignatureDiscoveries"
End
Begin OutputColumns
    Expression ="SystemInfo.Region"
    Expression ="SystemInfo.Constellation"
    Expression ="SystemInfo.System"
    Expression ="SignatureInfo.Signature"
    Expression ="SignatureInfo.Type"
    Expression ="SignatureInfo.Site"
End
Begin Joins
    LeftTable ="SignatureInfo"
    RightTable ="SignatureDiscoveries"
    Expression ="SignatureInfo.Signature = SignatureDiscoveries.Signature"

If you go to the SQL view and format a bit, this changes to the following.

@Cyril especially when "google" = DuckDuckGo

:)

dbMemo "SQL" ="SELECT DISTINCT SystemInfo.Region, SystemInfo.Constellation, SystemInfo.System, "
    "SignatureInfo.Signature, SignatureInfo.Type, SignatureInfo.Site\015\012FROM Syst"
    "emInfo INNER JOIN (SignatureInfo INNER JOIN SignatureDiscoveries ON SignatureInf"
    "o.Signature = SignatureDiscoveries.Signature) ON SystemInfo.System = SignatureDi"
    "scoveries.System;\015\012"

When you go back to the designer and save, you get the original one back.

These two different representations are really not helpful for SC.

holy carp, Batman! That's a lot of files in use elsewhere!

that's pretty much everything I've got open.

Excel & Word, but not Access, Notepad++, git bash and Postman. Those are the only things I have open that Tortise doesn't want to git its webby little feet into!

it's a Windows shell extension

therefore it's not surprising that there's lot of uses because applications will also use the shell and therefore shell extensions implicitly

that makes sense....

guess I needed to reboot my machine anyway. It's been about 2 weeks...

brb!

Hi guys. Anyone here using SolidWorks?

[rubberduck-vba/Rubberduck] MDoerner created comment on issue #5277: Array & undeclared Redim'd variables' refcounts and renaming are incorrect

interesting

makes me wonder...

> the higher ups said no

why? Install in the userland, no problem.

would it have been "yes" if we had a "RD Prime" USD $99/year subscription plan?

^ :-D

(IDK if they actually can lockdown userland installs. AFAIK, it can be done but is strongly recommended against)

what's the SLA on that?

@this it's pretty easy, actually... force an allowlist for all executables

IDK, we charge your company and give you the freeware, and then we get to make t-shirts and stickers for everyone every year

maintaining that allowlist is the annoying part about it :)

@Vogel612 Ah, #TIL

@FreeMan i haven't used duckduckgo since i had opera 9... at this point, i forget people don't use chrome with default google, though I have extensions to limit what can show up (if it's a promoted/advertised site it's blocked)

wow now that's locked down

@MathieuGuindon depends on the IT policy... i had to get a special admission to download notepad++...

wow

I've heard but never actually seen a workplace where bringing a USB thumb drive to work would get you booted

But I can see that being a thing

i downloaded it in the hour window i was permitted and gave a copy to our local IT guys since a few of them use it, and i've previously gotten others to use it due to the language identifiers

@this I can understand that kind of policy if your employer is, IDK, some government agency. But if your company sells t-shirts, that kind of policy is nuts.

NOBODY MUST STEAL OUR TRADE SECRET!!!11!

@this i've been at those before; the global IT monitors what is installed on everyone's computers and if it's not permitted they straight up uninstall things and you get a reprimand letter. if you get a second letter, fired on the spot

Meep!

lol

Reminds me when I was a total n00b and I used ssh for first time.

weirdest one was that i had IT lock down my registry when i was an 'admin" on my laptop... group policy would actively revert regedit changes even

I forgot to sudo, whatever, and I was told Not in sudoer list. The incident will be reported.

Reported to who?!?

please let the logged command be sudo make me a sandwich

and my favorite thing that company did... local IT was redacted the "privilege" of having access to power query by global IT so if they want to put any pushes or anything, they have to put in a ticket to IT to get escalated

er, powershell***

We actually cannot access USB drives at work.

I work for cowboys that let me be local admin and don't review any http logs :)

@MathieuGuindon Hi, Yosemite Sam!

yeee-haw!

It makes sense that we are not allowed to. Having a virus infested system would be a GDPR issue for us.

GDPR is interesting in general. supposed to provide more rites but what i've historically seen is companies don't hcnage, they just use jargon and layers of pages to hide info to still make "accessible" the information. as an american who regularly uses EU regulated sites, it's a weird thing to see... might be normal to the rest of the world

kind of like how 73% of houses in my development fly at least 1 flag, which is seen as weird to a lot of people (particularly my german friends)

sorry for the tangent... i get the USB thing. i don't think i've seen it that far, where they aren't recognized, but it makes sense

don't think that's any different. If they can circumvent, they will. There's a truth behind Reagan's saying WRT governmental mandates.

@Cyril meh... google gets more than enough of my life via my Android phone. They don't need all my search queries, too. And I'm migrating away from @gmail to @protonmail.com

is the search as good, though? That's the best selling point WRT gmail; email are much more searchable than other clients I've used.

@this You can get canned for it here. But we're a healthcare provider and HIPPA regs around PII/PHI are somewhat... strict. Violations can be really bad for the company.

@this are you referring to expanding intelligence agencies powers to gain access without allowing recourse for said ability?

@Cyril No. “Government’s view of the economy could be summed up in a few short phrases: If it moves, tax it. If it keeps moving, regulate it. And if it stops moving, subsidize it.”

@this I don't do a huge amount of searching in my email, so I can't say

@FreeMan Let me rephrase that slightly - if they can get away without getting burned, they will. You can mandate by letters but not by spirit.

Now that you mention it... I worked at a place once... They didn't have a lot for me to do, unfortunately, so I got bored. I got my wrist slapped for some general web-browsing, so I set up an SSH server at home and had a portable SSH client on my very first ever USB drive (all 32MB of it!!) and did my browsing via a very early portable Firefox release (also on said drive). They seemed to like that even less. Put a big target on my back come layoff time (due to the lack of work...).

@FreeMan just teamviewer or logmein to your home computer and play games...

had a colleague show me that once; logmein was paid for by the company, so he'd set-up meetings with himself at home and get the instant screen, rather than the 3-5 sec delay

~.~ when I hear about the thing pineapples do.

That just sounds so wrong.

^ 'That's just wrong.' Fixed it for you.

?

Logging into a home computer from work to play games = Wrong.

@Cyril If either of those existed back then, I was blissfully unaware of them.

gonna have to second that, @this ^

@IvenBach i was just confused about the pineapple comment... my first thought was "that they eat whoever eats them?"

@IvenBach it does put a damper on productivity. he did that crap every frickin day. boy was obsessed with warcraft

@this I fear we also have a bug renaming default members.

Thinking about it, I think basically all refactorings are broken for default members.

hm, the VBE might silently drop attributes when refactoring; that is part of the benefit of using annotations - should we try to recover them automatically or it's something else?

That is not the problem.

We now have the default member references.

For those, the refactorings will not work properly.

E.g. rename is not supposed to change their contexts at all.

If I'm not mistaken, we probably need to provide some type of markers to easily filter references depending on what we need it for.

Reorder parameter will not find the argument list at the same place since the entire expression is references, etc.

e.g. IsRefactorable to indicate that the reference is appropriate to use for rewriting

or something like that.

Well, we have IsArrayAccess and IsDefaultMemberAccess with the subproperties IsIndexedDefaultMemberAccess and IsNonIndexedDefaultMemberAccess.

I guess that would be also a different set from what we may want to show in the references, too.

Right - may be better to have a DeclarationFinder.GetRefactorableReferences(...)

then we can set up the filtering logic there

The references to refactor really depend on the refactoring.

I was thinking about presenting only a filtered list of declarations/references that are safe to rewrite

Rename should ignore all default member accesses, but reorder parameters has to reorder the parameters of indexed default member accesses.

then the refactoring could filter them further

The downside, though that means double filters which is usually a big performance hit

(right?)

hm, that's something we'll want to fix before the release, no?

I am right at it.

The array access thing is already fixed, unless there are more places we have to consider.

The default member thing will be a bit more work.

Now, how do I setup a default member in the rename test setup?

ouch

That thing is a bit unfriendly.

OK, that will not work as intended anyway.

We will loose the default member attribute.

It is a known gap in the member attribute recovery.

If the name of the member changes, we have no idea where to put the attribute.

Should we still allow to rename a default member?

I guess, I will open an issue for it.

@M.Doerner sure - that's exactly why "missing attribute" inspection was made!

(VBE is perfectly capable of losing member attributes all by itself, RD refactorings just make that slightly "easier" ;-)

We actually recover all member attributes except those for renamed members.

that's pretty slick ...I suppose we could try to recover attributes dropped on rename too

Unfortunately, the recovery kills the undo stack.

but I'm not considering this a showstopper at all - so long as the "missing attribute" inspection is doing its job

I think to avoid the drop on rename, we will have to use an attributes pass rename and actually deal with the attribute, too.

The most annoying situation would actually be a member attribute on an interface member implementation.

That is especially complicated since renaming the interface will execute the rename under parser suspension, but an attributes rewrite also suspends the parser, either leading to an exception or a deadlock.

letting "missing attribute" fire a result is fine too :)

not ideal, granted, but "good enough"

now if there was an attribute but no annotation (and user didn't put the annotations in)........ tough luck yeah

Hm, the test setup is a bit borked.

We do not remove the attribute.

Interesting, the test passed.

But in reality, it is broken.

Ah, the default member was private.

[rubberduck-vba/Rubberduck] MDoerner opened issue #5278: Stop loosing member attributes for renamed members

[rubberduck-vba/Rubberduck] MDoerner added label enhancement to issue #5278: Stop loosing member attributes for renamed members

[rubberduck-vba/Rubberduck] MDoerner created comment on issue #3929: Prompt before wiping member attributes

> Since we generally recover member attributes and the gap for renamed members is covered y issue #5278, can we close this issue.

[rubberduck-vba/Rubberduck] bclothier edited issue #5278: Stop losing member attributes for renamed members

[rubberduck-vba/Rubberduck] MDoerner edited comment on issue #3929: Prompt before wiping member attributes

> Since we generally recover member attributes and the gap for renamed members is covered y issue #5278, can we close this issue?

[rubberduck-vba/Rubberduck] retailcoder closed issue #3929: Prompt before wiping member attributes

[rubberduck-vba/Rubberduck] retailcoder created comment on issue #3929: Prompt before wiping member attributes

> Yup :)

@Duga Without the similar issue feature on issue creation, I would never have seen that related issue.

I simply do not look at enhancement issues.

There are still enough bug and technical dept issues and issues I opened myself to work on.

Hmm, good point. Eventually may need to do some cleanup via the enhancement issues.

I usually try to clean up bugs or some misc. labels (e.g. support) but I haven't actually went through the enhancements.

Speaking of which.... what do we do with the thanks issue exactly?

We have 2 thanks issue open

I guess we can keep them for a month or so and then close them.

Maybe, we can actually automate that.

actually 3. Per the proposal to close in a month, only one would qualify.

Yep, we got two rather recently.

I just checked, it can be automated via the REST API, if we want to actually do it.

So far, I think it is not worth it, though.

Regarding my hunt for refactorings broken for default members and array accesses, I will ignore EM.

start to thinking those probably should be outright deleted

(but that's another PR anyway)

so yes agreed w/ ignoring EM

[rubberduck-vba/Rubberduck] MDoerner assigned MDoerner to issue #5277: Array & undeclared Redim'd variables' refcounts and renaming are incorrect

[rubberduck-vba/Rubberduck] MDoerner created pull request #5279: Fixes around array access references to merge MDoerner/Rubberduck/FixesAroundArrayAccessReferences into rubberduck-vba/Rubberduck/next

[rubberduck-vba/Rubberduck] MDoerner added label PR-Status: WIP to pull request #5279: Fixes around array access references

Hm, what shall we do about change signature on default members used in index access expressions?

@Cyril This is my second week.

[rubberduck-vba/Rubberduck] build for commit 6cea837f on unknown branch: AppVeyor build succeeded

@Hosch250 sup man

@MathieuGuindon Regarding thyour comment from yesterday, I think you actually do quite a lot for RD. You are definitely our head of marketing.

Without that, a lot less people would know about RD.

=)