> The github page shows 439 open issues and all kinds of caveats in the updated version installation notes. I'd be afraid to try this app on anything but a virtual machine with a database I didn't care about.
@ThunderFrame my actual wording is currently as follows:
> Dismissing Rubberduck solely on the basis of the release notes caveats and the number of open GitHub issues, strikes me as ignorant of how GitHub, issue tracking, and long-term project management works.
> Hi there, I manage the Rubberduck open-source project, and frankly I find jwhite's assessment rather harsh.
The answer is right here on Stack Overflow - Rubberduck needs to parse the code before it can let you use most of its [rich] features, and then when you make code changes, you need to hit [refresh] to keep its parse trees up-to-date with what's in the VBE. You'll find the [refresh] command on every single toolwindow, as well as on Rubberduck's CommandBar (the parser state button would say "pending"). This discoverability issue has been addressed since then, and the Code Explorer to…
I've posted a reply on your forum, and seeing "Thank you for posting! Your post will not be visible until a moderator has approved it for posting." - how long do I have to wait before I know if my reply will be published or not? I'm replying to a [solved] thread that is damaging the reputation of my project out of ignorance and misinformation.
Cheers, Mathieu Guindon Microsoft MVP (Excel)
the damage is real: at least one user has uninstalled RD based on that one forum user's assessment
I mean, you can't have googled "rubberduck" and not have found us
Didn't mean for this to come off rude or anything - I'm not a registered trademark, just an open-source project. But anyone googling "Rubberduck" will quite likely find me before they find you... Do I get a hosting plan? (half-kidding)
Didn't mean for this to come off rude or anything - I'm not a registered trademark, just an open-source project. But anyone googling "Rubberduck" will quite likely find me before they find you... Do I get a hosting plan? (half-kidding)
You can do this pretty easily with a regular expression in the form of something like:
(^Jon\s)|(\sJon\s)|(\sJon$)
I'd wrap it in a function to allow building the pattern dynamically from user input. This is just an example - you'd either need to do some more escaping beyond just the . or (prob...
RE: searching - I think it's biased. I understood(?) that if I'm flemish, I'll get Flemish sites on rubber ducks before I even see any other rubberducks.
RE: @comintern, FWIW, I pinged him on GH referencing his PRs, too, so....
Judging from what I reviewed, he was using lot of tools, including something similar to Spy++ but in retrospect, I think it could be been misleading if we were all going by the window messages.
> FWIW, my vote would be #2. But do note that this is specific for this inspection. For other inspections, this does not make sense so the proposal has to be a bit more generic and thus allow the inspections themselves to describe what they will accept as ignorable.
> When the workflow involves cleaning up bad / broken files, it's very helpful to not have to have both CE and PE open until enough cleanup can be done to get the project to parse. Otherwise, it does not feel like I can totally let PE go, leaving me with more cluttered IDE. That said, I'm content to make this up for grabs.
@ThunderFrame The great thing about subclassing is that it works. Until you violate the assumptions that the original developers made about the object that's being subclassed.
> Once we have the ITypelib for each project, we should be able to have a fully populated Code Explorer without necessarily needing to reparse the source. I guess the ITypeLib might be missing the Private/Firend members.
> Once we have the ITypelib for each project, we should be able to have a fully populated Code Explorer without necessarily needing to reparse the source. I guess the ITypeLib might be missing the Private/Friend members.
Private => nobody else can call it; Public => anyone can call it; Friend => only members within same project can call it; Fiend => only members outside the project can call it....
@ThunderFrame it's very helpful to have that hand pointing them out....
Haven't heard from you in a long while, I really hope all is well. I'm writing a draft article on RD News about how we're finally fixing the crash on exit with the help of Wayne Phillips (vbWatchDog), and I was typing this line:
We lost a core contributor along the way. Perhaps he burned out, perhaps he got hit by a bus, to this date we still don't know.
...I went to lookup your "last seen" date on your SO profile, and was rather shocked (in a very very good way!) to read "Dec.29, 2017".
> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3664?src=pr&el=h1) Report > :exclamation: No coverage uploaded for pull request base (`next@4a6db52`). [Click here to learn what that means](https://docs.codecov.io/docs/error-reference#section-missing-base-commit). > The diff coverage is `91.93%`.
[](https://codecov.io/gh/rubberduck-
FWIW - this assumes he's already dead. He might have been burned out or had some major life events, neither which are mentioned as possibilities. I would adopt a more neutral stance with the last paragraph.
6/8 I emailed him asking for a life sign 7/29 I emailed him again emphasizing getting worried about him and informing him that I had to revoke his R# license (and transferred it to Max) then there was the RD swag google form email around that time too
@ThunderFrame VBE offers two versions of ITypeLib... one that offers the private stuff, and one that doesn't. And IIRC there is an easy way to switch between the two
Haven't heard from you in a long while, I really hope all is well. I'm writing a draft article on RD News about how we're finally fixing the crash on exit with the help of Wayne Phillips (vbWatchDog), and I was typing this line:
We lost a core contributor along the way. Perhaps he burned out, perhaps he got hit by a bus, to this date we still don't know.
...I went to lookup your "last seen" date on your SO profile, and was rather shocked (in a very very good way!) to read "Dec.29, 2017".
interesting! also, I hope Kent from the linkedin forum opens an issue for his crashes on open problem. I haven't noticed anyone mention such an issue before
also, what is the state of SCC in RD now? experimental?
"Rubberduck is a very active open-source COM add-in project that integrates with the Visual Basic Editor to enable the features every programmer wants to have in their IDE. From unit testing to source control, from code inspections to refactorings, programming in VBA will never be the same."
I worry that we're advertising SCC, but yet it's still experimental
> With a predeclared annotation becoming less necessary, RD needs a way of showing whether a class (or any userform/document module) is predeclared.
This could be done by either/both of: - Appending a "Predeclared" suffix to module/component selections in the status bar - Flagging predeclared modules/components in Code Explorer, possibly using an overlay or a new icon
It might also be worth identifying (with similar markers) modules/components that are *PublicNotCreatable*
> When adding `'@Folder` comments to modules and hitting refresh on the Code Explorer, I've been getting no changes to the folder layout
## Reproduce:
- Open a new instance of Excel , add a new module , add a routine to it ``` Sub DoStuff() End Sub ``` - Turn on the Rubberduck Code Explorer, and hit parse - Add `'@Folder("folder1")` and reparse again - Code explorer does not update
My guess was that the hash function outlined in the [parsing process page][1] was not taking co
> Just looked, I'm getting a lot of `Annotation 'Folder' is illegal in this context` errors (see image), some of which can be solved by moving the annotation to the first line of the module. Others like the example in the image (where the module is in the wrong folder) still give the error, and still won't update the Code Explorer, except on opening/closing the workbook.
You'll need to have at least an empty line between the folder annotation and the Sub like so:
``` '@Folder("Tests")
Public Sub DoStuff() End Sub ```
the VBE should then draw a demarcation line below the comment to mark the end of the module header. Folder annotations are only parsed correctly in the Module Header.
The second repro seems worrying, it should work in that case, the first is [status-by-design] though. Because there's no empty line be
> So a generic `@InspectionIgnore MissingMemberInspection Access.Control, Value, SelText, SelStart, SelLength; Excel.Control, Value, [..]`?
At that point we should consider moving these into a setting, IMO, since the ignore specifications are theoretically the same at every possible inspection site. As such ignoring per-site leads to quite a lot of overhead ...
git newbie question... after updating code on my branch in vs2017, how do I now update my git PR? I did a 'sync' in team explorer on my branch, but now what?
Yeah, as you saw, a sync will automatically sync the PR. Just keep pushing to the same branch (that's where PR tracks changes).
@M.Doerner In the RPS, we have both a dictionary for projects and a static field that I made non-static. I did not realize we had the dictionary but now seeing that, I am thinking that we should use dictionaries for both projects and components to cache the reference, do you agree?
Basically, I hand over the caching to a repository that caches, the projects collection, projects, components collections, components and code modules.
Regarding your comments on VBENativeServices, I was considering having the ComSafe have a list of SafeEventedComWrapper so that a single call to detach all events everywhere can be done, but without disposing the wrappers.
Hmm. I start to see why @comintern used static to avoid those. Because events are now per-instance, it might be better if the repository had the list of handlers so it can manage register/unregister of old/new wrappers on the behalf of RPS & SC viewmodel and whoever needs events?
howdy, Mat.
@Duga @Vogel612, hence why it has the feature-settings label. ;)
while troubleshooting a commenter on the blog post I realized just how horrible of an experience copying code from that blog is - I need to fix this, it's unreal.. every line has some hidden character that the VBE isn't understanding.
I think we should have something like an EventsProvider to handle the events in a central spot. It could take the repository as constructor argument and be injected via the IoC container as singleton.
@M.Doerner so the EventProvider would accept an instance of the .NET class and its delegate and specify which COM event it wants to be linked to, and it then manages the attach/detaching?
@this I was thinking of a ProjectAndComponentsEventsProvider that gets the VBProjects collection via its constructor packaged in the repository and manages the attaching and detaching of events from there. I itself would provide all the projects and components events on its interface.
> For clarification - the calls here all assumes only one name. In case of ComModule.cs's GetComFields, there's a loop where we call the GetNames - could we have called it only once by finding out how many fields there are then pass in an array of that many elements without a loop?
> @bclothier Unfortunately not, they each have there own member ID, so you have to call individually.
For clarity, the >1 use case is for argument names, where you pass in the member ID, and get the member name, and optionally all argument names in one scoop.
> Adding `Option Explicit` below, below with a space, above or above with a space all leave the same error message, and also fail to update the Code Explorer
Rubberduck v. 2.1.0.10753 Windows v.10.0.16299 Build 16299 Office v.1711 Build 8730.2165 Click-to-Run
> # [Codecov](https://codecov.io/gh/rubberduck-vba/Rubberduck/pull/3664?src=pr&el=h1) Report > :exclamation: No coverage uploaded for pull request base (`next@4a6db52`). [Click here to learn what that means](https://docs.codecov.io/docs/error-reference#section-missing-base-commit). > The diff coverage is `91.93%`.
[](https://codecov.io/gh/rubberduck-
> @Greedquest note that this issue has been fixed already; we haven't "green-released" v2.2 yet because we're currently working on putting the final few nails on the teardown issues' coffin, but feel free to upgrade to the latest pretelease build.
> Yes, the parsing all went smoothly, and as I say, when I do a full parse (close and open the file then re-parse) the folders in Code Explorer work as expected. But the inspection errors are still there, and updating the '@Folder argument then re-parsing doesn't change the folder structure any more.
> I cannot reproduce this issue from next. My "Illegal Annotation Inspection" severity is set to "Error", but I don't see any inspection results, even though I have at least one illegal annotation there. Soo ... something's fishy, but the other way round
on latest pre release, renaming a module causes leaks in the following objects: _VBComponents, _dispVBComponentsEvents, IConnectionPoint, IConnectionPointContainer
looks to be all events related, so I'll retest this once @this has done his changes.
@WaynePhillipsEA The way my PR is currently, I think it'll leak more. That's why @M.Doerner was earlier talking about holding a repository to help with managing events.
Were it written in C++, without all the COM interop code in the way, I'd imagine we'd have less of a problem managing the events because we can release them deterministically.
But there aren't that many C++ developers running around... ;)
VBA Rubberducking
VBA Rubberducking