linterhub / usage-parser Goto Github PK

It's unclear why script is failing during execution. Please, add handling of errors.

Project have many linting errors. Please, fix it.

Please, create next implementation of templatizer function:

• template shemas import
• writing options "args:" or w/o prefix

Please add support for .pullapprove.yml and make @linterhub/maintainers as reviewers for this repo.

If parser running many times in cycle, result json file consist arguments of all previous iterations. Please, fix it.

Please, create templates:

• args.json - schema of result
• option.json - default schema of option

Please, add validation of configuration, which should have:

• validation of config correctness
• empty values filling with defaults

Sometimes, in options section it is no information what can be interpreted that option is not flag. But, in examples there is so.

Example:

  Options:
    --plugin          Include third-party plugins [Can be set multiple times]

  Examples:
    $ xo --plugin=html --extension=html

Please, add parsing of "args" prefix by examples.

Please, add functionality for identifying of prefix "args" for non-flag arguments.

Please, add a template for an issue and pull request, also add labels for repository

Recommendation:

• pull request template
• issue templates

Please, find solution of parsing argument types (string, number, etc.) in templatizer function

Default value is not assigning to options now. Please, fix it.

Now, after parsing an argument, default value in it description isn't deleted. Please, make so, that default value will be deleted from description.

Please, create an implementation for splitSection function. This function should slice a section into array of options. Every element of array is string (option with description).

Please, create parsing of all arguments, which should be with linterhub prefix in templatizer function.

Please, add travis configuration.

Config must be include following command:

• npm run test
• linting

Now, in parser using only one pattern of default value of argument:
/\[default: (.*?)]/

So, default value will be parsed only if it looks like [default: 70].

Please, add other patterns.

I find next errors:

• sections Usage and Examples parsing only one time (only first)
• section Examples isn't parsing when it CAPITALIZED

Please, fix it.

Please, do refactoring of tests, because parts of code in test/config/test.js, test/sections/test.js, test/result/test.js are the same.

Also, please, do refactoring of tests filenames (folder files #52).

Next solutions isn't good:

• function handle of handle.js file

const context = require('./template/context.js');
context.options = [];

• function templatizer of templatizer.js file

const argumentsTemplate = JSON.parse(fs.readFileSync('./src/template/args.json'));
argumentsTemplate.definitions.arguments.properties = {};

It it necessary, because javascript filling empty templates, and for each next iteration of parser, js using this filled templates. Please, find best solution, how to fix it

Please, create implementation for parseSection function, which should get a section with some title from usage. For example, a section with title "Options:".

Sections are not parsing, if:

• Section data is after an empty line. Example:

Options:

    -h, --help                                                  output usage information
    -V, --version                                               output the version number

Examples:

    htmlhint

Please, fix it.

Now, users are not able to configure prefixes of arguments (linterhub:, args:). Please, make it configurable.

Please, add tests for two or more sections of one type (two "options:", three "examples", etc.).

Please, add new property "parseWhole" [boolean] to configuration.

If it true and no sections with names and postfixes was found, the whole file will be parsed. Each part, which starts at new line and begin with "-" will be parsed as option, and part between option and next option will be parsed as option description.

If description of option have two spaces between words or parts, will be parsed only first part before two spaces. Please, fix it.

Also, when description multiline, each line concatenates with previous without spaces.

Also, when description multiline, argument name becomes "null".

If option have only short name, it id in result json will be "null". Please, fix it.

Please, add parsing of delimiter between arguments and values (for non-flag args).

Please, add packages tests (one valid and one failing) with gulp. Tests should be:

• get helps of packages from given array of package names (with help of npx)
• send this helps to parser() and parse it
• with the help of gulp, identify, if parsing is clear or it is failing and print errors

Please, add tests for configValidation function:

• empty input json
• valid input json
• without fields (for every field)
• with other fields

Please, add inital docs:

• package.json
• gitignore
• README.md - with structure
• e.t.c

Please, add linting to project for js files.

Please, add tests of sections (with help of chai) what are using in parser.

Tests:

• if section is absent (for every section)
• if section consist invalid data (for every section)
• if there are other sections, what are not in use in parser

Function splitSection of handle.js looks unclear, it's hard to undestand it logic. Please, do refactoring.

In result json file, linterhub:path argument description is null. Please, fix it.

Now, all configuration of parsing is in handle.js file. It is better to move configuration to a separate file.

We have next config:

• object argumentTemplate, that is a template for an array of objects, that return by handle()
• object context, that has configuration of sections and its data.

Please, create base implementation of parser function:

• add function to parse linter's help (w/o implementation)
• add function to templatise result schema (w/o implementation, only template import)

Please, create implementation for handle function. This function should:

1. Get all suitable sections from parseSection function
2. Send it to splitSection function, which will split sections into options
3. Get a two-dimensional array of all options from parseOption function

Undefined option adds to result, if no linterhub:path required (in usage "file" or "path" is not found). Please, fix it

Please, create implementation for parseOption function, which should slice an option with description into array and return it. This array will be write into json later.

Structure of array:

• option (short version)
• option (long version)
• boolean, if option has value (needed later to add prefix "args:" if true)
• default value of option
• description of option

We have new algorithm of parsing (#38). Please, add tests for it.

Now, delimiter from config is not in use. Please, add it using.

Convert the information about the commands, options and flags of console app to JSON Object. This information is usually available when executing help command (possible synonyms: -h, --help, ?, etc.).

Example from docopt project:

Naval Fate.

Usage:
  naval_fate ship new <name>...
  naval_fate ship <name> move <x> <y> [--speed=<kn>]
  naval_fate ship shoot <x> <y>
  naval_fate mine (set|remove) <x> <y> [--moored|--drifting]
  naval_fate -h | --help
  naval_fate --version

Options:
  -h --help     Show this screen.
  --version     Show version.
  --speed=<kn>  Speed in knots [default: 10].
  --moored      Moored (anchored) mine.
  --drifting    Drifting mine.

So, this information could be parsed and converted to JSON similar to args definition. It should describe available arguments/commands/flags, name, type (if possible), default value and so on. There is no need to create exactly the same structure, args mentioned just as an example.

Implementation should include:

• The JSON schema of result (if it will be not the same as args above);
• Console app that able to parse a string with usage information and return JSON object according to schema;
• In case of any errors or if parsing failed by any reason - output a message with detailed explanation;
• Pre-processing logic that may clean up and restructure input string to make it parsable;

Some examples for args schema are here.

When constructing JSON Object give attention to next recommendation:

• Use full names of arguments. E.g. --help, not -h

Default value of option option is undefined when it is no defaults for it. Please, fix it, because it makes testing of parsing harder.

User should be able to send config and help (which will be parsed) as:

1. path to file
2. stdin

Please, add this features to index.js.

Please, add a linting to project.

Next comands:

• lint - command to check all json and js files on project
• before-commit - runs before each of commit. Include: lint

Recommendation:

• eslint - for linting
• husky - for run before-commit command

Please, add tests of result of parser. Tests have input string & output json and should compare result of parsing input string with output json, what should be.

I've found next errors:

• when description is fully on next line with option, it is not parsing
• exception when postfix or names [config] is string (not an array)
• parseWhole always changing to default config value

Now, it is parsing only one section name. But, some linters have "Options" instead of "Options:", or custom names: "Basic configuration:" (example from eslint).

Please, change algorithm of parsing of required sections (now it is only section "Options:") to the same way:

• Trying to parse primary names "Options:" and "Options:\n", if it not found, next:
• Trying to parse secondary names "Options" (other patterns may be added later), if it not found, next:

Please, fix 2 problems:

• different sections not parsing, only first
• section name isn't passed to the findSection function