> @IvenBach I haven't looked at the conflicts, do they look salvageable? If they become mergeable after the release, I'll merge them, but the "resource freeze" window is closing soon, and then every new resource key will probably conflict again.
is there an alternative to RegexStringValidator? VS tells me I need to add a reference to the assembly System.Configurations.ConfigurationManager but as far as I can tell assembly references cannot be added to ASP.Net projects.
note, there are a number of simpler conditions that don't necessarily need a regex, e.g. it's overkill to determine whether a required field was supplied
MSB3103 Invalid Resx file. The 'data' start tag on line 1374 position 4 does not match the end tag of 'date'. Line 1376, position 5. Rubberduck.Resources C:\Users\IvenBach\source\repos\Rubberduck\Rubberduck.Resources\RubberduckUI.cs.resx 1376
I pick WPF at work for user-facing desktop apps, otherwise I have a number of console apps, some scheduled... tbh most desktop apps I wrote there could be intranet apps... if we had an intranet server
*user-facing apps that don't "have" to be macro-enabled Excel workbooks
But most of my day is typically spent in SQL Server land
BUT, when you are in a loop, you might not want it to reuse the top N nodes leaving data in place as it does. You'll get weird cases of data not appearing to be updated.
So inside a loop, you put the @key attribute that gives Blazor a unique key for the node.
Now Blazor knows which nodes to drop/keep, instead of just doing the top N.
And yeah, it usually manifests as weird cache issues.
> The goal of the PR is to separate out the mechanics of going through a type library from the operating on the data returned by the type library because we have different needs that requires going through the type library:...
* We need it for our `Com***` declarations, which provides a stringified representation and is appropriate for caching. * We need it for future object browser to replace VBIDE's object browser and provide more information. * We need it for COM mocking. Currently the COM mocking relies on magical `Marshal.GetTypeFromITypeInfo` method which has shown itself to be unreliable. We need to be able to build .NET types from the data. * We may replace the dependency on OleWoo in the...
Rubberduck.Deployment.Build for the purpose of generating a valid IDL file. That has the added benefit of removing a dependency that is tied to a specific bitness. In order to achieve the goal, we provide iterators in form of both TypeLibWalker and TypeInfoWalker which handles all the enumerating, converting pointers into a valid data structure or values and returning those data for easier handling. Both types are also visitable, and can AcceptITypeLibVisitor(s) and...
ITypeInfoVisitor(s), respectively. The implementations are also required to indicate whether they can provide the other visitor to handle the visiting of a referenced type library from a type information or to visit the types within a type library. The user is free to visit only a type, or can choose to visit the entire type library. Furthermore, the user is able to direct whether the referenced types ought to be visited or not. It is hoped that offers enough flexibility in allowing...
different users to operate in their own ways upon the same type library. I'm posting this for review by others while developing tests. Feedback appreciated.
I searched and got the answer of shrug emoji so wanted thoughts from here
by "shrug emoji" I mean there were a ton of different opinions and I had zero coalescence at all
some thought always adhere to ISO others said use localized setting others said use whatever you want and let user change if they desire
some who advocated option a or b said letting user set that was overloading user, which I didn't necessarily buy
some said that the use of leading zeroes was helpful for 24 hour time to resolve potential ambiguity with non meridiem designated 12 hour time but that is not relevant to my use case
my inclination is to default no but let user decide
ISO does not really seem like a relevant concern for business users who may or may not have substantial technical computer skills
We need to clarify --- does the PR #5744 and PR #5692 overlap and need to be resolved? Both talk about improving handling of the identifier references.
Hi. Translation is almost complete. I've found a possible bug? I have in even if I set the language to English so... Can someone check that in your installed version? (I have +- week old)
but in the case of the per-user/per-admin and legacy workload, you are configuring how to install, and for whatever reason still unknown to me, it has to be a variable.
@this I think it wants a ; after end only when there aren't any more conditions - so end when there's an else if, and end; when the inner condition block is completed
you need a ; to terminate the if...else because there's no end if
but the content within the true and else if branch(es) must not be terminated with a ;
PASCAL IS FREAKING AWESOME.
I honestly don't know what they were thinking back then. "yeah, we shouldn't allow semi-colons to terminate the branch. It'll be confusing." Confusion ensures
I'm thinking that tB could do with some language-level support for enumerators... Apparantly it's possible for VB6 to implement a custom enumerator, but it involves some pretty nasty hacks and IDL...
Making EXCEL.EXE my debugging host app is trapping in a box where I only think about hosted-in-Excel scenarios, it didn't even occur to me that the Excel library might get unloaded between two runs...
@MathieuGuindon this is for the case else if CurPageID = wpReady then, which IINM is after the install.
@M.Doerner Not sure the inspection will cover the case but even without any Excel workbook being specifically referenced, it's possible to accidentally call an Excel Global method from Access host.
It's much more likely if you have Excel and Word both referenced as there is more overlap in the names between the 2 libraries (e.g. Range and Shape)
@SonGokussj4 @M.Doerner FYI the installer PR is introducing a new installer page with custom strings -- would you rather push commits to that PR's branch, or include the .iss tweaks in the translation PRs?
In my case, I avoid holding a explicit instance globally by using only procedure-level variables and a function to retrieve the instance. That way, when the procedure exits, it is taken care of.
could be a hint-level, [RequiredLibrary] inspection (Application.Quit isn't standard IIUC), so we could warn about Quit never being invoked for any not-the-host Application instance.
ref-counting is instant: when it gets to 0 it's freed; GC (in .NET at least) "generationizes" refs and decides on its own when things actually get freed
AFAIU, .Net actually does a graph search from the root node (well actually multiple if you dare to pin something in memory explicitly) and uses the generations to avoid doing the work over and over for long-lived instances.
@M.Doerner there also is a factor of moving objects to dedicated areas for generationality to improve the speed of memory access, because generally speaking most programs will create large numbers of small objects with small lifetimes. Having those in a hotter memory area improves performance
@MathieuGuindon Got a nagging bit of Imposter Syndrome, I'm sure I'm missing something. Only vaguely aware of what's happening on the COM side of things...
@mansellan combined with inheritance, unless I'm also missing something this would enable twinBASIC types to basically define every data structure VB-land has dreamed of for 30 years
hmm in order to test the walkers, I'll have to import my unmanaged memory allocator. But I don't need it within the project; only for tests. Should I add it anyway for future use?
VBA Rubberducking
VBA Rubberducking