I have to say that list has gotten waaay longer since last time I saw it.LSP really has been a massive success, and despite being made by MS, has been one of the main reasons I’ve been able to keep using Emacs and not migrated to other more IDE-like editors. MS deserves som serious credit here for being one of the few players willing to think in protocols these days. Protocols vs products makes a big difference.
|
|
|
|
| |
It's truly one of those I-can't-believe-nobody-thought-of-this-before ideas. And beyond the obvious benefits to developers' experience, it's also a huge boon to smaller language projects. If you build a new, open language and don't have the backing of a huge company, you definitely don't have the resources to integrate it directly with every editor someone might want to write it in. But you might be able to put together a single language-server.
|
|
| |
There was (and still is) YouCompleteMe for Vim, which is the idea of a code-completion server with editor plugins, so it's definitely a thing people had thought of.Microsoft really took the ball and ran with it, and I'm very grateful for that.
|
|
| |
I think Slime for Emacs was the first to think of it, a looong time ago (at least it was the first place I saw the idea in action).Edit to clarify -- Slime was never meant for general use across many languages. But Slime had the "smarts" in a separate process and communicated with the editor via the network.
|
|
| |
I know I'm making an LSP implementation for my personal language. A total must-have.
|
|
| |
It seems that Language Server Index Format (LSIF) https://code.visualstudio.com/blogs/2019/02/19/lsif is the successor to the Language Server Protocol (LSP). LSIF supports code navigation without having all source code files available on a local disk like LSP requires.
|
|
| |
Maybe "successor" isn't the right word, as per the linked article:> The goal of the Language Server Index Format is to augment the LSP protocol
|
|
| |
Correct! We designed LSIF to be a complement to LSP. LSIF is basically a way to persist language server state, so that LSP queries (e.g. what are the definitions for this symbol?) can be answered without needing an active LSP server running.Among other things, LSIF enables rich IDE navigations on remote codebases, where’s its much more scalable to perform those queries using indexed data vs. spinning up language servers. Here’s the blog post where we first announced the work we were doing, that led to the creation of LSIF: https://code.visualstudio.com/blogs/2018/12/04/rich-navigati....
|
|
| |
Thanks for making that clear.
|
|
| |
At Sourcegraph we are also running an identical website but for tracking LSIF indexers :) https://lsif.dev
|
|
|
|
|
|
| |
Quick piece of feedback: the big table with all the languages could use a color other than red (gray?) for "Not Applicable". For example, I assume that's the reason JSON is missing "Jump to def", "Find references", etc.
|
|
| |
THe SourceGraph server for Go is deprecated in favor of 'gopls'. This is documented on the sourcegraph github page.
|
|
|
|
| |
Sourcegraph CEO here. We fixed it and added gopls. Sorry for the mistake. It is not bias. We are very happy that Google made gopls. It is not our core product (universal code search is), and Google building gopls means we can focus more on our product. :)
|
|
| |
It's interesting that langserver.org is actually run by SourceGraph and it's still out of date.
|
|
| |
Given this, I wonder what the future holds for JetBrains. They've made a good business out of providing standalone solutions for the more popular languages.
|
|
| |
The same thing as always, making kick ass IDEs and actually working, robust productivity tools?A few rows down someone is praising "language servers" as some sort of Einstein discovery. I'd argue that was the easy part and they are rapidly learning about the difficulties involved, and the way is paved with auto-generated configuration pages wrapping json.
|
|
| |
Indeed, and so far for the languages I’ve used, the level of support and refactorings surpassed anything presented by open source tools.Additionally, they provide a uniform editing experience for a load of languages, which is one of the things I love the most. Where you’re writing go, python, java m, cpp or rust, the refactorings work similarly. (I am kind of a fanboy)
|
|
| |
I haven't seen a paid language server, but I would be willing to pay for a fully featured, plug-&-play server with installation instructions for common editors. Maybe they could pull some of their tools out of their IDEs and into a server?
|
|
|
|
| |
Another thing I really like about LSP is that formatting and linting are built-in concepts. Using something like https://github.com/iamcco/diagnostic-languageserver you can bridge existing command-line tools that only support stdio into your editor of choice.
|
|
|
|
| |
Where can I found a feature comparison of state of the art LSPs vs intellij idea features?
What can one do that the other cannot?
|
|
