Search results

Actually, an Access BE runs fine when on an ext4 partition with Samba authentication running on Linux. I did it for years. An Access BE doesn't run in place, it's opened by the FE.

Agreed. Cheers.

I understand well the distinction between compiled and interpreted code. No reason exists to suggest otherwise. I also don't conflate expansion with compilation. So, among other things, this means I do not call symbol-table isolation a design defect. Bash is indeed the Bourne (Again) SHell...

I take the point and understand why you might think that. Actually, I'm casting desperately for understanding for want of documentation providing it. My only expectations are those that someone else's documentation articulate. When someone else's language deviates from someone else's...

[facepalm] Of course. I stand corrected, and thanks. Good to know about that. I'm careful with scope so don't usually require them but things happen.

I finally got a readable presentation of your code and thanks for the ideas. Yes, all those things will work. This is just one step in a refactoring exercise that I describe in my response to @MajP 's comment. The data will migrate to a table in time but that has some dependencies and for...

Backslashes are the escape character in C and Bash. It operates to quote the immediately following character in a string. This usage is a bit of Access' C++ code making it past compilation.

Oh, God. Numeric strings and an ambiguous operator to boot. The result is actually correct. The question is whether it's numeric or a string. It should be a string and this should be numeric: Eval(("1"+"1")) but (should <> is)

Philipp, I discovered this morning that Eval() also fails with references to local functions. So, starting with a standard module named "stdBas": Option Compare Database Option Explicit Public Const MyConstant As String = "ConstantString" Function MyFunction() As String MyFunction =...

That's a big "if," actually. The reason? Several but I'll start with criteria expressions, which, when required and for all their pickiness, are isolated from expansion before being submitted at runtime for expansion and submission. So, Eval() fails because either (a) its argument isn't...

Yes, exactly. That's the design failure. My point is that Eval() would not fail if it instead were implemented essentially as a subclassed compiler to exclusively evaluate its argument.

Thanks for that. I've read it before and actually do exactly what it describes for a slightly different purpose. Mike uses a predeclared class module as the primary container but I've found a standard module will perform identically and is a bit more robust. I also actually had a close look...

Thanks. That's quite helpful and explains or confirms much. I actually have parsed the VBAL on a number of occasions over the years and so am familiar with it. In fact, I believe I have several editions of it in my reference directory. It's minimally descriptive and can't hold a candle to...

In Bash, this is called expansion. What you describe explains the failure. Eval() jumps the gun by expanding its argument before it executes. Wow. To me, this reveals a profound and fundamental design flaw of Eval() in that its very purpose would seem to depend on deferring expansion of the...

Quite helpful, this and your other comments, and thanks. Coincidentally, I've actually been diving deeper into VB.NET lately and had overlooked this. This reveals, among other things, that I don't understand expressions as well as I might. I have tried to get a handle on evaluation otherwise...

Pete, thanks for asking. The context is a standard module with the necessaries for getting code into a project, including modules, references, unregistered DLLs, etc., which I'm refactoring. For the time being, the module has among other things a number of module-level string constants...

No, the module has a number of library GUID strings as constants, used to add, test, or remove project references.

Thanks for both replies. Your code is illegible in my browser (I think there may be site settings for this) and it's late, but I'll work on getting something readable and revert.

Having read that documentation just now, I wouldn't necessarily say "misleading." "Silent" might be more descriptive. "False," "wrong," and "incorrect" also would work nicely. Regardless, any might well be amplified with the adverbs "woefully" or "profoundly." To the point, it's good to know...

They're string constants, actually, GUIDs.