Vim Creator Bram Moolenaar Aiming To Improve Vim Performance With Vim9 Fork

[wizard69]

Interesting read in the linked article. However I really have to wonder if building yet another programming language, which appear will happen here, is a good thing. Would it not make more sense to replace a Vim Script with a more common scripting language or even a language with a modern REPL?

Bram has some good points such as the wisdom of supporting multiple scripting solutions not working out that well. So why not make a big move here to a brand new Vim supporting a modern scripting languauge with wide support and a good future. It looks like Lua is already a decent competitor and if versions are an issue ship Vim with the suitable Lua release. If Lua isn’t right there are other newer languages that have adopted REPL support that should be considered.

I bring this up because it looks like a Vim Script will change enough that many scripts will require rewrites. We all know what happened with Python when the idiots out there couldn’t manage to adopt the new Print function, I can see a similar Vim revolt. At least by making a completely new product you avoid some of these issues.

so what do you guys think. Should Vim be gutted to support a widely supportted scripting language for better long term support? Or should a Vim Script continue on? I actually believe Bram is in the right track as far as having one and only one scripting interface but there seems to be a better way in Lua rather than reworking Vim script.

by the way I’m mentioning Lua here due to the test results that indicate impressive performance already vs the other solutions. Lua also runs close to the Vim9 scripting performance. It wouldn’t be impossible for other REPLs from other languages to duplicate the results.

[wizard69]

Yeah, 'cause VIM is soooo slow. Even Windows 10 on a 2002 HDD boots faster than VIM!!! [/sarcasm]

After spending a bit thinking about this I have to say that Vim or even GVIM is very fast at loading. This in comparison to Geany let’s not even get into the joke of an Eclipse start up.

speed though is subjective! It does make you wonder if chasing speed benchmarks is even a good use of developer time on Vim! Slow just doesn’t come to mind with Vim. If speed is an issue maybe they need to consider other things like multi threading.

[uid313]

Nobody wants to bother learning an application-specific esoteric programming language. They should just use a popular scripting language that people know and like. Not having to waste time learning some weird language that nobody use.
I think it should support something popular and widely used like Python or JavaScript.

[kpedersen]

something popular and widely used like Python or JavaScript.

Python in its entirety is larger than Vim. Seems a little inelegant to choose that. For those types of user, there is Visual Studio code.

[Lycanthropist]

I'd vote for guile, since it's GNUs official extension language.

[uid313]

Python in its entirety is larger than Vim. Seems a little inelegant to choose that. For those types of user, there is Visual Studio code.

Good point. But Python is also used for scripting by other software such as X-Chat / HexChat, etc.
Also if Python is too big, then there is MicroPython.

Either way, if Python and JavaScript are too big, then there is Lua.

I'd vote for guile, since it's GNUs official extension language.

GNU should just abandon Guile. Nobody ever heard of it. Nobody cares about it. Nobody knows it. Nobody wants to learn it. Plus it is in the Lisp dialect, so nobody wants that. Better go with Lua then.

[jrdoane]

Yeah, 'cause VIM is soooo slow. Even Windows 10 on a 2002 HDD boots faster than VIM!!! [/sarcasm]

You must not use vim with many plugins. I use vim for my day job to write TypeScript and Clojure and recalculating syntax highlighting for Clojure can be brutal, particularly if you're including defined vars in the current namespace. I happen to know this from experience.

Plus it is in the Lisp dialect, so nobody wants that.

As a Clojure dev, I'll kindly suggest that you're wrong, but I'm one of this masochists that writes a Lisp in vim though.

[uid313]

As a Clojure dev, I'll kindly suggest that you're wrong, but I'm one of this masochists that writes a Lisp in vim though.

But you're one of those weirdos who code in Clojure. It is not even on the top 50 in the TIOBE list.

[timofonic]

Interesting read in the linked article. However I really have to wonder if building yet another programming language, which appear will happen here, is a good thing. Would it not make more sense to replace a Vim Script with a more common scripting language or even a language with a modern REPL?

Bram has some good points such as the wisdom of supporting multiple scripting solutions not working out that well. So why not make a big move here to a brand new Vim supporting a modern scripting languauge with wide support and a good future. It looks like Lua is already a decent competitor and if versions are an issue ship Vim with the suitable Lua release. If Lua isn’t right there are other newer languages that have adopted REPL support that should be considered.

I bring this up because it looks like a Vim Script will change enough that many scripts will require rewrites. We all know what happened with Python when the idiots out there couldn’t manage to adopt the new Print function, I can see a similar Vim revolt. At least by making a completely new product you avoid some of these issues.

so what do you guys think. Should Vim be gutted to support a widely supportted scripting language for better long term support? Or should a Vim Script continue on? I actually believe Bram is in the right track as far as having one and only one scripting interface but there seems to be a better way in Lua rather than reworking Vim script.

by the way I’m mentioning Lua here due to the test results that indicate impressive performance already vs the other solutions. Lua also runs close to the Vim9 scripting performance. It wouldn’t be impossible for other REPLs from other languages to duplicate the results.

You are perfectly describing Neovim, check it out

[ssokolow]

After spending a bit thinking about this I have to say that Vim or even GVIM is very fast at loading. This in comparison to Geany let’s not even get into the joke of an Eclipse start up.

speed though is subjective! It does make you wonder if chasing speed benchmarks is even a good use of developer time on Vim! Slow just doesn’t come to mind with Vim. If speed is an issue maybe they need to consider other things like multi threading.

It can feel very janky depending on the language you're editing. (PHP is somewhat infamous) Syntax highlighting recalculations still block the main thread in Vim 8.x, so any momentary hitch will prevent it from responding to user input.

I use pear-tree, not because I'm particularly attached to having my editor insert matching closing quotes/parens/etc., but because auto-balancing parens results in a noticeable UI responsiveness improvement due to how Vim does syntax highlighting.