I'm happily developing a language in ANTLR 4.4 / ANTLR4CS, and I'd like to start integrating my language into an IDE. The natural place for me feels like integrating it into Visual Studio. It occurs to me that someone might have figured out a general way to plug an ANTLR lexer into the Visual Studio syntax highlighting system, or ANTLR parsers into an MSBuild task such that errors appear in the Visual Studio error list.
Is there any kind of starting point, base package, plugin system, or similar which lets you integrate an ANTLR4CS language with Visual Studio's language support mechanisms?
Alternatively, has anyone had any success building, say, a Sublime Text build system / language def automatically from their grammar?
#Lucas Trzesniewski wrote:
I did this by reparsing the whole file every time, doing it on a line-by line basis was too much of a PITA for my needs (and our DSL files are small so it works fine).
I don't know if it'll be suitable for your needs, but you can implement a single file generator from a VS extension. The Generate method takes a IVsGeneratorProgress parameter which lets you report errors easily.
Related
I am working on a nodejs project in which all code is written in typescript. It follows a microservice pattern (and each microservice is an independent project) hence lot of projects needs to be opened and debugged at the same time.
I tried webstorm and visual studio (with NTVS) but not satisified with both of them. Webstorm ignores many build errors (very important as those will fail during CI) and not as good as visual studio in intellisense and ease of use. On the other hand, Visualstudio is not as fast as I want. Also I am tied to windows and the filename length problem in node_modules is very frequent and annoying.
Others (atom, VSCode) doesn't have typescript debug support (or am I missing something?).
Can you please share your experience? Also I was wondering which editor google developers might be using to develop angular 2 :)
I have been working on a TypeScript project for 1.5 years now and am quite happy with using Visual Studio because we used .NET for back-end stuffs. For myself I consider the speed as "Ok", it is not that good - but the other stuffs e.g. intellisense is what we really like. I also personally think that as a "son" of Microsoft, TypeScript will get the best support from its "cousin" - Visual Studio.
However if you are a Eclipse person, you may be interested in this TypEcs (http://typecsdev.com/). There was also a post by Steve Fenton talking about TypeScript IDEs you may be interested (https://www.stevefenton.co.uk/Content/Blog/Date/201409/Blog/Which-TypeScript-IDE/), he listed a few of the popular ones with a few sentences of comments, including Atom and the atom-typescript extension.
A bit late to the conversation.
I recently tried VisualStudio 2015, VisualStudio Code 0.9, Atom (with Atom-Typescript package). And quite comfortable with Atom.
Atom-TypeScript uses the latest version of TypeScript so you can use all new features like async/await, string interpolation etc. The package created by community so we don't need to wait for major IDE release to use new TypeScript features.
For debugging couldn't find one. But love to use node-inspector for its light weight. The problem is you need to debug it in javascript version not in typescript version. Not so sure but maybe if you can provide map file you can debug the typescript version? like we have on Google Chrome Inspector.
VSCode can debug in Node. No browser debugging yet.
Not a recommendation, but I suggest you do not buy JetBrains' products (IntelliJ/WebStorm/ReSharper) for TypeScript editing unless you have evaluated them properly.
They reused a lot of their own logic for other languages instead of taking full benefit of the TypeScript language service, which means they behave more like a fancy JavaScript editor that can make some clever guesses as opposed to a robust IDE for a statically typed langauge. (Their type inference is different from the compiler and may fool you into believing an untyped expression (implicit any) is well-typed. Also, all the refactoring are the same as the JS ones, so they do not use or preserve any type information. No type-information-based refactoring (like safe renaming and moving) are available either.)
As for debugging, you may have more milage debugging the generated JS. Not only is it well supported, it also reveals problems that occur in compiler-inserted code. The latter would be otherwise difficult to understand if you're only looking at the TypeScript source. Also, in the generated code, you can actually set break points in single line lambda function bodies.
I too use Visual Studio. I haven't had any problem with its speed; I find everything is as immediate as I expect it too.
However, I haven't been using "TypeScript" projects. I have been using generic "Web" projects with TypeScript files. Plus, I don't use Visual Studio's own building and compilation facilities; I use my own Gulp tasks for my project (including rebuilding, watching for saves, minification, testing, etc), and Task Runner Explorer to manage/run them from inside Visual Studio. I think it's the best of both worlds: the snappiness and features of Visual Studio, and the flexibility of running my own custom tasks independent of platform.
I have to admit I don't do any deep debugging (e.g. breakpoints, stepping) from the IDE. Does VS do that with TypeScript? If I catch an error during execution (in the browser), I always have my source maps indicating where the error is, so I go back to the code, attempt to fix it, save, and reload in the browser.
I am using VSCode to actively debug an angular/node/typescript application. It does work, but you will need to do some setup/configuration. You must tell VSCode how to debug your applications. For instance I have multiple options in my launch.json configuration to debug my gulp scripts, my node server.js file, and the angular/typescript application via chrome using the "Debugger for Chrome" extension installed in VSCode. It is all possible with some configuration. Finally, depending on your exact situation you may need to have and correctly configure the map files for your project from javascript to typescript.
I also had the problem of file paths too long in the node_modules folder in visual studio when opening a project with npm. This is due to Visual Studio trying to download all your packages - incorrectly so I might add. The only way to avoid this for my project was to mark the node_modules folder as a hidden folder in windows (top level only).
I want to add support for syntax highlighting for lex and yacc-files in Visual Studio 2010.
How can I do this?
Following the given link to Syntax Coloring leads to another, more pertinent page Implementing Syntax Coloring, where it is noted
Visual Studio does not specify a parser interface, and parser implementation is completely up to you. However, a default parser implementation is provided in the Visual Studio Language Package project. For managed code, the managed package framework (MPF) provides complete support for colorizing text.
Depending on what you want:
trivial, coloring only the C code portions of the lex/yacc files
much harder, coloring the patterns so that one can make sense of them
you could in principle make a parser using just lex (yacc isn't necessary). For yacc files, that's not so hard, but for lex there's the complication of regular expressions. vi-like-emacs does that and while the interface differs in detail, conceptually it's similar. Reading the lexers might give you some ideas how to apply that approach:
lex syntax filter
yacc syntax filter
There is documentation from Microsoft on integrating yacc/bison & flex/lex with Visual Studio. Although a URL only link is discouraged on SO, its best to go to the source for this kind of detailed information:
https://msdn.microsoft.com/en-us/library/aa730877(VS.80).aspx#vccustombr_topic3
It lays out all the steps necessary for integrating with the build tools. Syntax highlighting is covered elsewhere as described in Add a new language to Visual Studio 2010 with syntax highlighting and intellisense. In particular the guidance for syntax colouring can be found here: https://msdn.microsoft.com/en-us/library/bb166778(v=vs.100).aspx.
However, I'm not aware of anyone who has published extra colouring rules for the grammar specific components of flex and yacc. However, most of the body of flex and yacc files are written in C or C++ for which there are syntax highlighting rules that can be applied, and suit most peoples needs.
I want to build a language service for visual studio 2010. I was first trying to follow the tutorial and documentations from MSDN.
The problem is i don't succeed to make this works (i'll explain later my problem). So i digged into existing implementations, i found Ook! and lua . both of these projects doesn't use the tutorial or documentation i found on MSDN, but something based on MEF. Lua used this only with previous Visual Studio versions.
So i'm wondering if i'm using an obsolete method to create a language service (But the documentation aims Visual Studio 2010), or there is different ways to do this, which depends on needs.
In my case, i've got a language that doesn't need to be compiled into cli, but i want to have an editor that have colorization, syntax warnings and errors, intellisense ...
The problem i mentionned is that when launching exp instance, there is no text editor with my file extension, and visual studio begins to have many lags. The language service is registered using 3 attributes : ProvideServiceAttribute, ProvideLanguageServiceAttribute and ProvideLanguageServiceExtension. Also initialized in Package intialize method, like mentionned in Proffer the Language.... The package is loaded when i try to open the file with my extension, the language service is initialized.
So i don't get it why i does not work, could you please help me to understand how language service works, and what is the best way to implement it
Thanks
Good chance your IScanner implementation has an endless loop, happened to me.
Custom Compiler Warnings and
C#: Create custom warning in Visual Studio if certain method is used in source code
haven't helped as they deal with code that is under the author's control.
We are using a 3rd party suite of UI controls (DevExpress) in our software and I want to generate a warning when someone uses MessageBox.Show("blah"); instead of XtraMessageBox.Show("blah");
Is there a way to do that?
This sort of thing can be addressed relatively easily via a custom rule for FxCop/Visual Studio Code Analysis. If you are using Visual Studio Developer Edition, you will even see the rule failures displayed along-side your compilation warnings and errors in the IDE.
While there's no way you can do real custom compile-time error in .NET, there's a number of third-party tools (both free and commercial) that can inject their validation logic into the build process, usually after the compilation.
Here are three ways I know of to solve you problem:
Resharper 5.0($) will support custom rules / warnings.
In PostSharp(free) you can define OnMethodBoundary aspect, overwrite its CompileTimeValidate method and emit a [post]compile-time error from it.
NDepend can be integrated with your build process ($) to enforce coding policies like that
No there is no direct way. If you think about it you are looking for a compiler warning for some code that you don't even compile.
If you really want this you could use Reflection methods on YOUR compiled assembly to check if any methods/assemblies you don't want have been called. Cecil has a lot of the functionality you need. You could then make this part of your build process.
I am trying to figure out what the browse information (.sbr files) is used for but find only references how to create it. So what is it for?
Thanks
Dima
Read here (Visual C++ Team Blog: IntelliSense History, Part 1)
Capturing information about a C or C++
program’s structure has been around
for a very long time in Microsoft’s
products. Preceding even Visual C++
1.0, the compiler supported generating program information through .SBR and
.BSC files. (Note: The compiler in
Visual C++ 1.0 was already version 8,
so the command line tools had been
around a while already.) The SBR
files contain reference and definition
information for a single translation
unit that the compiler generates as it
compiles. These SBR files are
combined in a later step using the
BSCMAKE tool to generate a BSC file.
This file can then be used to look at
many different aspects of a program:
reference, definitions, caller-callee
graphs, macros, etc.
.sbr is pretty much Visual Studio's ctags - an index of symbols with backreferences to the source. When available, it's used by "Find Symbol" and other similar tools. It's more accurate than the built-in VS parser, because C++ can be tricky, and the real compiler can do a better job (though that is not quite true in VS2010 anymore).
At one time browse info drove the "Go to definition" engine, but that has been reworked in later version of Visual C++. Some third-party tools still use browse info (can't remember for sure, but I think one of Rational's tools does) to cross-reference code.
I always disable it, to shorten build times.