watson-ruby Gem Version Build Status

an inline issue manager

watson (mirror) is a tool for creating and tracking bug reports, issues, and internal notes in code.
It is avaliable in two flavors, watson-ruby and watson-perl

See watson in action here (mirror)

See the RDoc documentation here (mirror)

Installation

watson-ruby has been tested with Ruby v2.1 and RubyGems v2.0.3 (on Arch Linux)
watson-ruby requires the json gem

From Repo

watson-ruby is avaliable as a RubyGems (link). You can either download it directory from gem using

gem install watson-ruby

Or you can clone this repo and install with Rake

git clone https://github.com/nhmood/watson-ruby.git .
cd watson-ruby
bundle install
bundle exec rake

Recent Updates

  • watson now supports Asana
  • watson now supports GitLab
  • vim-unite-watson.vim integrates watson right into vim!
  • watson now supports multiple export modes
    • JSON output allows you to integrate watson parse data into your own apps
    • Silent output allows you to use watson without a printout; tie it to git hooks for streamlined issue posting

Usage

For a quick idea of how to use watson, check out the app demo! (mirror)
See below for a description of what all the command line arguments do.

Supported Languages

If you see something missing from the list please either file an issue or submit a pull request (comment parsing happens in lib/watson/paser.rb)

  • C / C++
  • Objective C
  • C#
  • Java
  • Javascript
  • PHP
  • Go
  • Scala
  • Erlang
  • Fortran
  • SQL / PL
  • Lua
  • HTML
  • SASS SCSS
  • Haskell
  • Bash / Zsh
  • Ruby
  • Perl
  • Python
  • Coffeescript
  • Clojure
  • VimL
  • Markdown
  • HTML
  • Emacslisp
  • LaTex
  • Handlebars

Command line arguments

Usage: watson [OPTION]...
Running watson with no arguments will parse with settings in RC file
If no RC file exists, default RC file will be created

   -c, --context-depth   lines of context to provide with posted issue
   -d, --dirs            list of directories to search in
   --debug               enable debug prints from specified class
                         all debug prints enabled if no arguments passed
   -f, --files           list of files to search in
   --format              set output format for watson
                         [print, json, unite, silent]
   -h, --help            print help
   -i, --ignore          list of files, directories, or types to ignore
   -p, --parse-depth     depth to recursively parse directories
   -r, --remote          list / create tokens for Bitbucket/GitHub
   -s, --show            filter results (files listed) based on issue status
                         [all, clean, dirty]
   -t, --tags            list of tags to search for
   -u, --update          update remote repos with current issues
   -v, --version         print watson version and info

Any number of files, tags, dirs, and ignores can be listed after flag
Ignored files should be space separated
To use *.filetype identifier, encapsulate in "" to avoid shell substitutions

Report bugs to: [email protected]
watson home page: <http://goosecode.com/watson>
[goosecode] labs | 2012-2013

-c, --context-lines [LINES]

This parameter specifies how many lines of context watson should include when posting issues to remote repos.
When this parameter is set from the command line, the .watsonrc config file is written with the value; the command line option effectively sets the default value for this feature in the current directory.
The default value is set to 15 (and can be found in the lib/watson/command.rb file).

-d, --dirs [DIRS]

This parameter specifies which directories watson should parse through.
It should be followed by a space separated list of directories that should be parsed.
If watson is run without this parameter, the current directory is parsed.

--debug [CLASS]

This parameter enables debug prints for the specified class.
It should be followed by a space separated list of classes that should print debug messages.
The list of classes are found in lib/watson.
If passed without arguments, debug prints in ALL classes of watson will be enabled.

-f, --files [FILES]

This parameter specifies which files watson should parse through.
It should be followed by a space separated list of files that should be parsed.

--format [PRINT, JSON, UNITE, SILENT]

This parameter specifies how watson should output the issues that it finds.
If passed with print, the regular printing will occur, either to Unix less or STDOUT (depending on system).
If passed with json, the output will be in the form of JSON, and will be stored in .watsonresults.

  • This particular option is useful if attempting to intergrate watson into other tools / platforms.

If passed with unite, the output will be compatible with the unite.vim plugin
(see the Unite section below or visit alpaca-tc/vim-unite-watson

If passed with silent, watson will have no output.

  • This particular option is useful if remote posting to GitHub, Bitbucket, or GitLab is desired without the visual component of watson.
  • For example, you could set up a git commit hook to post issues to GitHub/Bitbucket/GitLab, but avoid the giant print out every time.

-h, --help

This parameter displays the list of avaliable options for watson.

-i, --ignore [IGNORES]

This parameter specifies which files and directories watson should ignore when parsing.
It should be followed by a space separated list of files and/or directories to be ignored.
This parameter should be used as an opposite to -d/-f, when there are more files that should be parsed in a directory than should be ignored.

-p, --parse-depth [PARSE_DEPTH]

This parameter specifies how deep watson should recursively parse directories.
The 'depth' is defined as how many levels deep from the top-most specified directory to parse.
If individual directories are passed with the -d (--dirs) flag, each will be parsed PARSE_DEPTH layers, regardless of their depth from the current directory.
If watson is run without this parameter, the parsing depth is unlimited and will search through all subdirectories found.

-r, --remote [GITHUB, BITBUCKET, GITLAB, ASANA]

This parameter is used to both list currently established remotes as well as setup new ones.
If passed without any options, the currently established remotes will be listed.
If passed with a github, bitbucket, gitlab, asana argument, watson will proceed to ask some questions to set up the corresponding remote.

-s, --show [ALL, CLEAN, DIRTY]

This parameter is used to specify which types of files and entries are shown when watson is run.
If passed with the clean argument, only files that watson did NOT find issues in will be displayed.
If passed with the dirty argument, only files that watson DID find issues in will be displayed.
If passed with the all argument, watson will display all files, regardless of their issue status.
The default behavior of watson is the all option.

-t, --tags [TAGS]

This parameter is used to specify which tags watson should look for when parsing.
The tag currently supports any regular character and number combination, no special (!@#$...) characters.

-u, --update

This parameter is used to update remote repos with new issues.
watson does not post new issues by default therefore this parameter is required to push up to GitHub/Bitbucket/GitLab.
watson does pull issue status by default, therefore you will always be notified of resolved issues on GitHub/Bitbucket.

-v, --version

This parameter displays the current version of watson you are running.

.watsonrc

watson supports an RC file that allows for reusing commong settings without repeating command line arguments every time.

The .watsonrc is placed in every directory that watson is run from as opposed to a unified file (in ~/.watsonrc for example). The thought process behind this is that each project may have a different set of folders to ignore, directories to search through, and tags to look for.
For example, a C/C++ project might want to look in src/ and ignore obj/ whereas a Ruby project might want to look in lib/ and ignore assets/.

A base .watsonrc is created in the users $HOME directory and used as the template for creating all subsequent config files.
Any changes made in the $HOME/.watsonrc will carry on towards new configs, but will not update previously created configs.
The $HOME/.watsonrc is directly copied, not merged, as to avoid confusion as to the source of config parameters.
The $HOME/.watsonrc contains all remote GitHub API tokens created. The individual project .watsonrc will contain only the username associated with the API token, and will reference $HOME/.watsonrc to obtain the API token on runtime.

The .watsonrc file is fairly straightforward...
[dirs] - This is a newline separated list of directories to look in while parsing.

[tags] - This is a newline separated list of tags to look for while parsing.

[ignore] - This is a newline separated list of files / folders to ignore while parsing.
This supports wildcard type selecting by providing .filename (no * required)

[context_depth] - This value determines how many lines of context should be grabbed for each issue when posting to a remote.

[type] - This field allows for custom filetype/comment associations without having to edit the gem itself.
The format of filetype entries should be

".type"   => ["comment1", "comment2"]
".nhmood" => ["@@", "***"]
".cc"     => ["//"]

[show_type] - This field allows for setting the default show type, similar to the -s/--show command line arg
If set to clean, only files that watson did NOT find issues in will be displayed.
If set to dirty, only files that watson DID find issues in will be displayed.
If set to all, watson will display all files, regardless of their issue status.
The default behavior of watson is the all option.

[tag_format] - Customize the tag format for your project. Defaults to [TAG] - COMMENT. "TAG" and "COMMENT" are required keywords.

[(github/bitbucket/gitlab/asana)_api] - If a remote is established, the API key for the corresponding remote is stored here.
Currently, OAuth has yet to be implemented for Bitbucket so the Bitbucket username is stored here.

[(github/bitbucket/gitlab)_repo] - The repo name / path is stored here.

[(github/gitlab)_endpoint] - The endpoint in case of GitHub Enterprise of GitLab configuration.

[asana_project] - The Asana project in case of Asana integration is stored here.

[asana_workspace] - The Asana workspace in case of Asana integration is stored here.

The remote related .watsonrc options shouldn't need to be edited manually, as they are automatically populated when the -r, --remote setup is called.

unite.vim

vim-unite-watson.vim by alpaca-tc

alpaca-tc was kind enough to create a unite.vim plugin to integrate with watson!
unite.vim allows you to create user interfaces within vim and lets you search and display any kind of information.
The unite-vim-watson.vim plugin allows you to:

  • Generate a list of issues directly inside vim
  • Search and select files displayed in the watson output
  • Jump directly to the line where the issue is located (!!)

If you use vim I definitely recommend using vim-unite-watson.vim, it works wonders for productivity!

See it in action here (GIF demo)

Special Thanks

Special thanks to @samirahmed for his super Ruby help and encouraging the Ruby port!
Special thanks to @eugenekolo [email] for his super Perl help!
Special thanks to @crowell for testing out watson-ruby!

FAQ

  • Will inline issues get deleted if I close them on GitHub/Bitbucket/GitLab?
    No, watson won't touch your source code, it will only inform you that issues have been closed remotely.

  • Why Ruby?
    I wanted to learn Ruby and this seemed like a pretty decent project.

  • Why is the Ruby version different from the Perl version?
    The Ruby version was developed after the Perl version was made. Because of this, it was a lot easier to add on features that were thought of while/after making the Perl version as the plumbing was still being setup.
    With a combination of wanting to finish watson-ruby as well as laziness, some of the improvements that were added to watson-ruby have yet to be pulled back into watson-perl.
    If you are interested in helping out or maintaining watson-perl let me know!