I am using C++ in vs code and would like to automatically pull in changing external include paths that for clang to find. Currently I have to edit the c_cpp_properties.json file as so:
{
"configurations": [
{
"name": "Linux",
"includePath": [
"${workspaceFolder}/**",
"/some/path/headers1/**",
"/some/path/headers2/**",
"/some/path/headers3/**",
...
"/some/path/headersN/**"
],
...
}
],
}
Can I instead do something like this?
{
"configurations": [
{
"name": "Linux",
"includePath": [
"${workspaceFolder}/**",
"${workspaceFolder}/headers.txt"
],
...
}
],
}
Where ${workspaceFolder}/headers.txt is:
/some/path/headers1/
/some/path/headers2/
/some/path/headers3/
OR
"/some/path/headers1/**",
"/some/path/headers2/**",
"/some/path/headers3/**"
to be compliant with the original json includePath format. So there a way for me to reference the ${workspaceFolder}/headers.txt in the c_cpp_properties.json so that I don't have to manually update the paths?
Related
I'm trying to write a TextMate grammar for a VS Code language extension. Take the following example
(lang=css attribute2=something-else)
"""
.css-class {
background: gray;
}
"""
The (...) part is an "attributes" section, and the """ ... """ is a code section. I'm trying to highlight everything in the code section according to the lang attribute.
The problem is, they are two distinct sections where one might be present without the other in other parts of the file. For example, you can have attributes without the code block.
In the grammars section of package.json I have
"embeddedLanguages": {
"meta.embedded.block.css": "css",
"meta.embedded.block.javascript": "javascript"
}
In the tmLanguage.json file I have both patterns in the repository property.
"attributes": {
"begin": "\\(",
"end": "\\)",
"captures": {
"0": {
"name": "punctuation.definition.annotation punctuation.section.group punctuation.section.parens"
}
},
"patterns": [
{
"begin": "[a-zA-Z_][a-zA-Z0-9_\\.-]*",
"beginCaptures": {
"0": {
"name": "entity.other.attribute-name"
}
},
"end": "(?=\\s*+[^=\\s])",
"patterns": [
{
"begin": "=",
"beginCaptures": {
"0": {
"name": "punctuation.separator.key-value"
}
},
"end": "(?<=[^\\s=])(?!\\s*=)|(?=/?>)",
"patterns": [
{
"match": "([^0-9-.\\s='\"][^\\s='\")]*)",
"name": "string.unquoted.html"
},
{
"match": "=",
"name": "invalid.illegal.unexpected-equals-sign"
},
{
"include": "#strings"
},
{
"include": "#number"
}
]
}
]
}
]
},
"fenced-code": {
"begin": "\\(.*lang=(css|javascript).*\\)\\s*(\"\"\")",
"beginCaptures": {
"1": {
"name": "string.quoted.triple"
}
},
"end": "\"\"\"",
"endCaptures": {
"0": {
"name": "string.quoted.triple"
}
},
"contentName": "meta.embedded.block.$1",
"patterns": [
{
"include": "source.css"
}
]
}
I have a third pattern not shown where I'm using these together by including them in its patterns array. They seem to be mutually exclusive though. I can have the attributes, and I can have a code block if I start the pattern at """, but if I start the code pattern with \\(.*lang=(css|javascript).*\\)\\s*(\"\"\") to capture the lang attribute, the attributes stop getting highlighting.
Is this even possible? I've never worked with TextMate grammars outside of a VS Code theme and VS Code doesn't seem to have deep documentation on the more "advanced" (I guess) things like this.
I tried using VS Code's HTML grammar for its uses of embedded code, but I don't think the HTML one needs to swap syntax based on something in a previous line.
Update
The fenced-code pattern I have below allows the attributes pattern highlighting while also having the code block with dynamic contentName property, i.e. "meta.embedded.block.$2" becomes meta.embedded.block.css when "css" is found in attributes.
"fenced-code": {
"begin": "(\\(.*lang=(css|javascript).*\\))\\s*(\"\"\")",
"beginCaptures": {
"1": {
"patterns": [
{
"include": "#attributes"
}
]
},
"3": {
"name": "string.quoted.triple"
}
},
"end": "\"\"\"",
"endCaptures": {
"0": {
"name": "string.quoted.triple"
}
},
"contentName": "meta.embedded.block.$2",
"patterns": [
{
"include": "source.css"
},
{
"include": "source.js"
}
]
}
However, there's two things wrong so far.
It only works when the opening """ is on the same line as the attributes section
fenced-code's patterns array doesn't seem to allow a dynamic include, i.e. "patterns": [{"include": "source.$2}]. I'm not sure if including the different languages as I did above will work
I would like to pass environment variable(like passing prod or dev) in step functions so that pyspark program can get it by using os.getenv().There are many programs using same environment variables so passing them as arguments to the programs will take some time.Tried below code
[
{
"Classification": "yarn-env",
"Properties": {},
"Configurations": [
{
"Classification": "export",
"Properties": {
"VARIABLE_NAME": VARIABLE_VALUE,
}
}
]
}
]
but it didn't work.Seems this code will work for setting yarn configurations.
This is a question similar to Why eslint consider class as variable in naming-convention rule?, but that one is pretty old and I see no consistency in the handling now.
When I statically import a class type then ESLint recognizes it as such and applies the class naming rule, for example:
import { PieGraphLayout } from import("../console.worker-types");
When I do this with a dynamic import, however, I get an error:
const { PieGraphLayout } = await import("../console.worker-types");
leads to:
Variable name PieGraphLayout must match one of the following formats: camelCase eslint#typescript-eslint/naming-convention
I have to suppress this warning, but would like to modify my ESLint rules instead, if possible. My current naming-convention rule is:
"#typescript-eslint/naming-convention": [
"error",
{
"selector": "default",
"format": [
"camelCase"
],
"filter": {
"regex": "^_",
"match": false
}
},
{
"selector": "class",
"format": [
"PascalCase"
]
},
{
"selector": "typeParameter",
"format": [
"PascalCase"
]
},
{
"selector": "enum",
"format": [
"PascalCase"
]
},
{
"selector": "enumMember",
"format": [
"PascalCase"
]
},
{
"selector": "typeAlias",
"format": [
"PascalCase"
]
},
{
"selector": "interface",
"format": [
"PascalCase"
],
"prefix": [
"I"
]
}
],
What needs to be changed so that ESLint no longer gives a warning for such dynamic imports?
You don't have defined formats for variable entities, and ESLint uses your default selector format, which is camelCase in the provided configuration.
To provide formats for a variable, please configure variable selector.
In your case, there should be something like
{
"selector": "variable",
"format": [
"PascalCase"
]
},
Here is a doc for all available selectors.
I have a CMakePresets.json file which makes use of inheritance and macro expansion.
Here is an excerpt, in reality I use multiple versions of "Foo":
{
"configurePresets": [
{
"name": "default",
"hidden": true,
"generator": "Unix Makefiles",
"binaryDir": "cmake-build-${presetName}",
"environment": {
"PATH": "/opt/foo/$env{FOO_VERSION}/bin:$penv{PATH}",
"LD_LIBRARY_PATH": "/opt/foo/$env{FOO_VERSION}.0/lib"
},
"cacheVariables": {
"FOO_VERSION": "$env{FOO_VERSION}",
}
},
{
"name": "debug-foo1",
"inherits": "default",
"environment": { "FOO_VERSION": "1" }
},
{
"name": "release-foo1",
"inherits": "debug-foo1",
"cacheVariables": { "CMAKE_BUILD_TYPE": "Release" }
}
],
"buildPresets": [
{
"name": "debug-foo1",
"configurePreset": "debug-foo1"
},
{
"name": "release-foo1",
"configurePreset": "release-foo1"
}
]
}
Now assume I select the preset release-foo1. This would render the following variables, among others:
binaryDir = "cmake-build-release-foo1"
FOO_VERSION = "1"
LD_LIBRARY_PATH = "/opt/foo/1.0/lib"
Is there a way to query these results for a given preset? For example, given release-foo1, I want to know the resulting binaryDir.
Of course I could parse the JSON myself, but that seems tedious, especially because of the cross references and substitutions which are being made by CMake.
You can use the -N flag to print the computed presets info without running the configure or generate steps. After modifying the presets file in your question to work, I created an empty CMakeLists.txt next to it and ran this test (*** stands in for my PATH):
$ cmake --preset=release-foo1 -N
Preset CMake variables:
CMAKE_BUILD_TYPE="Release"
FOO_VERSION="1"
Preset environment variables:
FOO_VERSION="1"
LD_LIBRARY_PATH="/opt/foo/1.0/lib"
PATH="/opt/foo/1/bin:***"
This is good enough for debugging your presets. If you need to go further, you could process the output using standard unix command line tools, like awk, grep, cut, etc.
I think this is the best you can do for now (CMake <=3.21) since there's nothing else in the command line documentation or the file API for IDEs that I could find.
Building on Alex Reinkings answer, you can get the binaryDir (and other parts) with the -N option if you tweek the preset a little bit.
You need to add a environment variable and reference it for binaryDir:.
See this tweeked example from CMake documentation:
{
"version": 2,
"cmakeMinimumRequired": {
"major": 3,
"minor": 20,
"patch": 0
},
"configurePresets": [
{
"name": "default",
"displayName": "Default Config",
"description": "Default build using Ninja generator",
"generator": "Ninja",
"binaryDir": "$env{BUILD_DIR}",
"cacheVariables": {
"FIRST_CACHE_VARIABLE": {
"type": "BOOL",
"value": "OFF"
},
"SECOND_CACHE_VARIABLE": "ON"
},
"environment": {
"BUILD_DIR": "${sourceDir}/build/default",
"MY_ENVIRONMENT_VARIABLE": "Test",
"PATH": "$env{HOME}/ninja/bin:$penv{PATH}"
},
"vendor": {
"example.com/ExampleIDE/1.0": {
"autoFormat": true
}
}
},
{
"name": "ninja-multi",
"inherits": "default",
"displayName": "Ninja Multi-Config",
"description": "Default build using Ninja Multi-Config generator",
"generator": "Ninja Multi-Config"
}
],
"buildPresets": [
{
"name": "default",
"configurePreset": "default"
}
],
"testPresets": [
{
"name": "default",
"configurePreset": "default",
"output": {"outputOnFailure": true},
"execution": {"noTestsAction": "error", "stopOnFailure": true}
}
],
"vendor": {
"example.com/ExampleIDE/1.0": {
"autoFormat": false
}
}
}
This will give you:
$ cmake --preset default -N
Preset CMake variables:
FIRST_CACHE_VARIABLE:BOOL="OFF"
SECOND_CACHE_VARIABLE="ON"
Preset environment variables:
=>BUILD_DIR="D:/tmp/presets/build/default"<=
MY_ENVIRONMENT_VARIABLE="Test"
PATH="..."
I'd like to prevent src/config.js to be bundled in scripts/app-bundle.js
I saw that previously the syntax was:
"bundles": {
"dist/app-build": {
"includes": [
"[**/*.js]",
"**/*.html!text",
"**/*.css!text",
"cloneya",
"dexie",
"jquery",
"jquery-ui",
"medium-editor-webpack",
"moment",
"polymer/mutationobservers",
"safe-json-stringify"
],
excludes: [
"config.js" // So our wildcard globbing doesn't include this config file
],
...
However the new syntax is different: aurelia.json:
"bundles": [
{
"name": "app-bundle.js",
"source": [
"[**/*.js]",
"**/*.{css,html}"
],
"excludes" : [
"**/config.js"
]
},
My temptative 'exclude' statement doesn't do the trick
Solution is actually given on the GitHub page: https://github.com/aurelia/cli
Optionally, you can define an exclude list by setting the source
property to be an object containing both an include and exclude array
of patterns. This is helpful when you're trying to define multiple
bundles from your source code.
{
"name": "app-bundle.js",
"source": {
"include": [
"[**/*.js]",
"**/*.{css,html}"
],
"exclude": [
"**/sub-module/**/*",
]
}
},
{
"name": "sub-module-bundle.js",
"source": [
"**/sub-module/**/*",
]
}
Make sure you have version > 0.19.0