[RESOLVED] VB6 QUESTION: LSP server implementation for VB6

[yereverluvinuncleber]

Why do we not have a LSP server implementation for VB6? (I know the answer to this question, it is more or less rhetorical and designed to test the water and stimulate debate).

Other languages that have almost zero footprint in the programming world seem to have an LSP server implementation. It seems to be becoming a right of passage to have your own LSP implementation for your own language. Without it, our languages' lack of a place in the pantheon of modern programming technologies is highlighted.

Even visualBasic.FRED does not yet have such an implementation and I understand the reason for that - but it would seem necessary to add VB6 to ensure the longevity of VB6's distinct dialect and increase its exposure to the world and to help regain a tiny part of its place under the sun.

https://docs.microsoft.com/en-us/vis...n?view=vs-2019

Other implementations of some sort of future VB6 replacement could benefit from this, it if existed. Users of VBscript/VBA as well as VB6ers are those that could potentially benefit from having such a resource. It at least shows there is some life in the old dog yet. If it had existed already I am sure it would have been useful for RADBasic/TwinBasic development.

Not having a LSP server implementation for VB6 does imply that the languages only future lies in the old IDE alone and there is more or less no potential for it outside.

IS this the sort of project that could unite this fractured community to help create?

[wqweto]

It has to be an LSP server for VB6 implemented in VB6 too.

TwinBasic is using ANTLR4 grammar, probably inherited from RubberDuck's VBx one with (numerous) fixes/enhancements as TB language evolved.

IMO, anything short of manually written (in VB6) recursive-descent parser would be sub-optimal for the reputation of the language and the community :-))

cheers,
</wqw>

[Niya]

I think sometimes you guys forget that VB6 hasn't been updated in 20+ years and outside of our little community here, no one even knows it exists and those who do know it exists don't care about it. The world is run by Python, Java, Javascript and C#.

[yereverluvinuncleber]

I think sometimes you guys forget that VB6 hasn't been updated in 20+ years and outside of our little community here, no one even knows it exists and those who do know it exists don't care about it. The world is run by Python, Java, Javascript and C#.

Don't bring negativity here please Niya, we have enough of that without you using your scattergun approach to denigrating VB6.

Note that the work done to implement a LSP server for VB6 would naturally benefit VB.NET too as a lot of the syntax is similar. It could form a basis for that.

[WaynePhillipsEA]

It has to be an LSP server for VB6 implemented in VB6 too.

TwinBasic is using ANTLR4 grammar, probably inherited from RubberDuck's VBx one with (numerous) fixes/enhancements as TB language evolved.

IMO, anything short of manually written (in VB6) recursive-descent parser would be sub-optimal for the reputation of the language and the community :-))

cheers,
</wqw>

Just FYI, twinBASIC has always used a brand new parser, written from scratch in C++. Nothing is used from RubberDuck, and we don't use ANTLR. Naturally, it does have its own LSP server too, which it exposes to VS Code.

[dilettante]

VB6 is not a lower-tier 3GL "dark world" (command line) scripting language of the sort VS Code was designed for.

Stop posting irrelevant threads here. That's what "Other Basic" is for.

[WaynePhillipsEA]

VB6 is not a lower-tier 3GL "dark world" (command line) scripting language of the sort VS Code was designed for.

Stop posting irrelevant threads here. That's what "Other Basic" is for.

If your comment was aimed at me... it was quite clear that I was correcting mis-information, and I will continue to do so.

[Niya]

Don't bring negativity here please Niya, we have enough of that without you using your scattergun approach to denigrating VB6.

Note that the work done to implement a LSP server for VB6 would naturally benefit VB.NET too as a lot of the syntax is similar. It could form a basis for that.

Some one has to keep it real. You guys keep wanting VB6 to be something it's not. VB6 is for all intents and purposes a dead language. Olaf has extended it somewhat with his powerful RichClient library but that only goes so far. All I'm saying is that at a certain point, you have to be realistic. If you're interested in all these new toys modern developers get with modern languages, you'd be better served by moving to those languages instead of irrationally sticking to VB6.

[Niya]

If your comment was aimed at me... it was quite clear that I was correcting mis-information, and I will continue to do so.

He was directing it at the OP I believe.

[WaynePhillipsEA]

He was directing it at the OP I believe.

I thought that may be the case. I was the only one who mentioned VS Code, so wasn't sure.

[Niya]

I thought that may be the case. I was the only one who mentioned VS Code, so wasn't sure.

Well the topic is about an LSP implementation for VB6 and where LSP is mentioned, VSCode also becomes relevant. Dilettante is basically saying something to the effect of VB6 not being suited to VSCode which is why a VB6 LSP implementation isn't really a worthwhile topic.

[Shaggy Hiker]

If your comment was aimed at me... it was quite clear that I was correcting mis-information, and I will continue to do so.

No, it wasn't directed at you. Don't worry about it.

[dilettante]

VB6 and classic VB in general was a 4GL (or maybe a "3.5th" generation language) much like its contemporary competitor PowerBuilder.

These are primarily "designer first" GUI-based development systems with code-behind. A model Delphi and .Net languages in an altered Visual Studio emulated with varying degrees of success. VB's precursor, Alan Cooper's Tripod didn't even have any code-behind language at all. Then came the Ruby UI designer/generator, which was soon paired with the EB engine which brought Basic into the picture for the first time.

Trying to work from the rear toward the front by focusing on writing yet another QuickBasic clone first is doomed. We already have FreeBASIC for that, and it is still stuck in the dark world over a decade later. If you want to create another one... well, more power to you.

VSCode is basically a fancy text editor that can call "make" and "debug" macros. It's great for that but completely lacks the plumbing to become a replacement for VB6's IDE.


I just can't see how VB programmers are going to have much to add to these unrelated efforts. If they decide to leave VB there are plenty of other choices available already.

[Shaggy Hiker]

All I'm saying is that at a certain point, you have to be realistic.

I don't agree with this. Ultimately, what you are getting to is, "at some point you have to follow the herd." Whether or not that is even true is debatable, but whether or not it is good advice is a much more difficult debate.

If you work for somebody, you often don't get to choose which language you will be using, or the tools you will be using. So, we can rule out anybody in that situation: It's not up to them, so they either change where they work or they live with what they are required to use.

For the rest, there are the self employed and the hobbyists. There's a theoretical category that is between those two, which I fall into much of the time, but I don't think it's a real category. I suppose the category would be those who work for somebody else, but have few, or no, requirements as to what tools they use. Those folks are essentially the same as hobbyists, or the self employed, as they follow the same rules.

For people in those categories, if they always just follow the herd, they are generally safe (as long as the herd isn't running off a cliff), but they may miss out on things. Also, in this field, the direction is always changing. Any language WILL be abandoned if everybody stops using it. Any language will NOT be abandoned if everybody keeps using it.

There is no facing reality outside of that. It's just guessing at which way the tide will run. You might be right, you might be wrong, and even if you are right, you might miss out on something good. Also, if EVERYBODY went with the tide, the world would become homogenized. There would be no Linux, no TwinBasic, no VB6, no C++, and so on. At one point, all were swimming against the tide. That's reality. Face reality: The future can only be guessed at. The collective guesses will have a big hand in shaping it.

[wqweto]

Just FYI, twinBASIC has always used a brand new parser, written from scratch in C++. Nothing is used from RubberDuck, and we don't use ANTLR. Naturally, it does have its own LSP server too, which it exposes to VS Code.

Thanks for the info!

TwinBasic effort never stops amazing me :-))

cheers,
</wqw>

[Niya]

I don't agree with this. Ultimately, what you are getting to is, "at some point you have to follow the herd." Whether or not that is even true is debatable, but whether or not it is good advice is a much more difficult debate.

If you work for somebody, you often don't get to choose which language you will be using, or the tools you will be using. So, we can rule out anybody in that situation: It's not up to them, so they either change where they work or they live with what they are required to use.

For the rest, there are the self employed and the hobbyists. There's a theoretical category that is between those two, which I fall into much of the time, but I don't think it's a real category. I suppose the category would be those who work for somebody else, but have few, or no, requirements as to what tools they use. Those folks are essentially the same as hobbyists, or the self employed, as they follow the same rules.

For people in those categories, if they always just follow the herd, they are generally safe (as long as the herd isn't running off a cliff), but they may miss out on things. Also, in this field, the direction is always changing. Any language WILL be abandoned if everybody stops using it. Any language will NOT be abandoned if everybody keeps using it.

There is no facing reality outside of that. It's just guessing at which way the tide will run. You might be right, you might be wrong, and even if you are right, you might miss out on something good. Also, if EVERYBODY went with the tide, the world would become homogenized. There would be no Linux, no TwinBasic, no VB6, no C++, and so on. At one point, all were swimming against the tide. That's reality. Face reality: The future can only be guessed at. The collective guesses will have a big hand in shaping it.

If someone wants to use VB6 or has to use it, I say go ahead and use it. But it hasn't been updated in 20+ years and only like 12 people still use it. There is really no point in wondering why it doesn't have the same modern amenities modern languages and frameworks enjoy. It should be obvious why. Enjoy it for what it is and if you see some things in other languages that you like, then use that language. VB6 is VB6 and it's not going to change. It's not VB.Net. It's not Python. It's not JavaScript. Microsoft abandoned it decades ago and so did the rest of the world. What remains is what you have to work with.

This idea of trying to fit a square peg in a round hole is unwise and a huge waste of time unless the effort is a hobby you enjoy like the efforts of Olaf or Krool with their respective libraries.

[dilettante]

only like 12 people still use it.

That must be why it ranks as #11 even on the highly biased Tiobe Index, hmm?

I actually think most of that is Office VBA, but Tiobe is a can of worms at best anyway.

[Niya]

TIOBE is regarded as unreliable at best and garbage at worst. Use PYPL instead. The Visual Basic family ranks 18th on that list.

That must be why it ranks as #11 even on the highly biased Tiobe Index, hmm?

As far as I'm concerned, only the top 5 matter for any serious discussion. Maybe even top 10 if you want to let loose a little. We here all love Visual Basic(VB6, VB.Net, VBA etc) for various reasons but the truth is, the world doesn't care about Visual Basic. We should just be thankful for whatever little scraps thrown our way and when we are ready to get serious, we should leave it and go live in the top 5.

[Niya]

BTW Wayne if you're still here, does the TwinBASIC compiler support modern x86 instruction sets like AVX, or even older SSE instruction sets?

[mansellan]

Bringing this back to the question asked in the OP...

LSP can't be retrofitted to the VB6 compiler. LSP is just just a protocol, a way to expose language comprehension in a standardised format to front-ends like VSCode. The VB6 compiler makes none of its internal data structures available, so there's nothing to expose.

So there are a couple of options. You can ignore the VB compiler entirely, and write your own system to lex, parse, and comprehend the VB source (which is exactly what Rubberduck does). Potentially you can then expose this parallel system as an LSP server, and indeed there's an issue to discuss that: https://github.com/rubberduck-vba/Ru...ck/issues/5176

Or you can reimplement your own compiler for VB (/compatible) source code, and expose an LSP server. TwinBASIC has done that.

[Shaggy Hiker]

That must be why it ranks as #11 even on the highly biased Tiobe Index, hmm?

I actually think most of that is Office VBA, but Tiobe is a can of worms at best anyway.

VB6 is NOT found on the Tiobe Index. What IS found is Visual Basic, but if you look at the methodology that Tiobe publishes, you will see that, while it includes VB6, it includes three other things, as well, one of which is VBA, and the other two end up being a mix of VB6 and VB.NET. They decided to assign 50% of those two to the Visual Basic category and 50% to VB.NET, but if you repeat their steps, it should have been 95% to VB.NET and 5% to this category, and that 5% is still not VB6. You will also see that VBA makes up half the ranking.

If you used their methodology and split out VB6, it would be down with the also ran languages.

And for anybody who is curious: Yes, it annoys the heck out of me that Tiobe publishes their methodology such that any person could repeat their work, and folks STILL think that Visual Basic category means VB6. It's been YEARS folks.

[Shaggy Hiker]

It takes two to tango...and you two always find a dance partner. Just stop.

[Schmidt]

...removed (thanks to the Mod who was stepping in).

Though have to say, that: "It takes two to tango..." -
is one of the most idiotic idioms I know of.

If somebody walks into your living-room and then defecates all over your table at dinner-time,
would you like it, when the judge is letting him off - and tells *you*: "It takes two to tango"?

Olaf

[Niya]

VB6 is NOT found on the Tiobe Index. What IS found is Visual Basic, but if you look at the methodology that Tiobe publishes, you will see that, while it includes VB6, it includes three other things, as well, one of which is VBA, and the other two end up being a mix of VB6 and VB.NET. They decided to assign 50% of those two to the Visual Basic category and 50% to VB.NET, but if you repeat their steps, it should have been 95% to VB.NET and 5% to this category, and that 5% is still not VB6. You will also see that VBA makes up half the ranking.

If you used their methodology and split out VB6, it would be down with the also ran languages.

And for anybody who is curious: Yes, it annoys the heck out of me that Tiobe publishes their methodology such that any person could repeat their work, and folks STILL think that Visual Basic category means VB6. It's been YEARS folks.

I don't know why people still use TIOBE. The general consensus is that it's garbage. PYPL is way more accurate.

[Niya]

It takes two to tango...and you two always find a dance partner. Just stop.

Did you mean to say find each other as dance partners? That would make more sense

[Shaggy Hiker]

I don't know why people still use TIOBE. The general consensus is that it's garbage. PYPL is way more accurate.

Tiobe is a company that does a variety of things. That index is nothing more than one way of processing data, and an interesting way, at that. It's simple, repeatable, applicable to a variety of questions, and has intriguing built-in sources of bias. PYPL will also have built-in sources of bias.

For those who are interested in methodology, Tiobe is interesting. For those interested in sampling systems and their biases, Tiobe is interesting and accessible. PYPL may be, as well, though I have no experience with that one, nor do I care that much.

However, for those looking for TRUTH!!! Well, neither one has THAT. So, saying one is better than the other is kind of pointless.

[Shaggy Hiker]

Sorry that my casually selected phrase caused such a kerfuffle.

[yereverluvinuncleber]

You can toast a kerfuffle but you have to catch it first.

Shall we get back on topic?

[mansellan]

Shall we get back on topic?

Heh, I did try