The below is generated from a README at https://github.com/cosmicexplorer/helm-rg. MELPA: https://melpa.org/#/helm-rg !`helm-rg' example usage (./emacs-helm-rg.png) Search massive codebases extremely fast, using `ripgrep' (https://github.com/BurntSushi/ripgrep) and `helm' (https://github.com/emacs-helm/helm). Inspired by `helm-ag' (https://github.com/syohex/emacs-helm-ag) and `f3' (https://github.com/cosmicexplorer/f3). Also check out rg.el (https://github.com/dajva/rg.el), which I haven't used much but seems pretty cool. Usage: *See the `ripgrep' whirlwind tour (https://github.com/BurntSushi/ripgrep#whirlwind-tour) for further information on invoking `ripgrep'.* - Invoke the interactive function `helm-rg' to start a search with `ripgrep' in the current directory. - `helm' is used to browse the results and update the output as you type. - Each line has the file path, the line number, and the column number of the start of the match, and each part is highlighted differently. - Use `TAB' to invoke the helm persistent action, which previews the result and highlights the matched text in the preview. - Use `RET' to visit the file containing the result, move point to the start of the match, and recenter. - The result's buffer is displayed with `helm-rg-display-buffer-normal-method' (which defaults to `switch-to-buffer'). - Use a prefix argument (`C-u RET') to open the buffer with `helm-rg-display-buffer-alternate-method' (which defaults to `pop-to-buffer'). - The text entered into the minibuffer is interpreted into a PCRE (https://pcre.org) regexp to pass to `ripgrep'. - `helm-rg''s pattern syntax is basically PCRE, but single spaces basically act as a more powerful conjunction operator. - For example, the pattern `a b' in the minibuffer is transformed into `a.*b|b.*a'. - The single space can be used to find lines with any permutation of the regexps on either side of the space. - Two spaces in a row will search for a literal single space. - `ripgrep''s `--smart-case' option is used so that case-sensitive search is only on if any of the characters in the pattern are capitalized. - For example, `ab' (conceptually) searches `[Aa][bB]', but `Ab' in the minibuffer will only search for the pattern `Ab' with `ripgrep', because it has at least one uppercase letter. - Use `M-d' to select a new directory to search from. - Use `M-g' to input a glob pattern to filter files by, e.g. `*.py'. - The glob pattern defaults to the value of `helm-rg-default-glob-string', which is an empty string (matches every file) unless you customize it. - Pressing `M-g' again shows the same minibuffer prompt for the glob pattern, with the string that was previously input. - Use `' and `' to go up and down by files in the results. - `' and `' simply go up and down by match result, and there may be many matches for your pattern in a single file, even multiple on a single line (which `ripgrep' reports as multiple separate results). - The `' and `' keys will move up or down until it lands on a result from a different file than it started on. - When moving by file, `helm-rg' will cycle around the results list, but it will print a harmless error message instead of looping infinitely if all results are from the same file. - Use the interactive autoloaded function `helm-rg-display-help' to see the ripgrep command's usage info. TODO: *items checked completed here are ready to be added to the docs above* - [x] make a keybinding to drop into an "edit mode" and edit file content inline in results like `helm-ag' (https://github.com/syohex/emacs-helm-ag) - *currently called "bounce mode"* in the alpha stage - [x] needs to dedup results from the same line - [x] should also merge the colorations - [x] this might be easier without using the `--vimgrep' flag (!!!) - [x] can insert markers on either side of each line to find the text added or removed - [x] can change the filename by editing the file line - [x] needs to reset all the file data for each entry if the file name is being changed!!! - [x] can expand the windows of text beyond single lines at a time - using `helm-rg--expand-match-context' and/or `helm-rg--spread-match-context' - [x] and pop into another buffer for a quick view if you want - can use `helm-rg--visit-current-file-for-bounce' - [ ] can expand up and down from file header lines to add lines from the top or bottom of the file! - [ ] can use newlines in inserted text - not for file names -- newlines are still removed there - would need to use text properties to move by match results then, for everything that uses `helm-rg--apply-matches-with-file-for-bounce' basically - [x] visiting the file should go to the appropriate line of the file! - [x] color all results in the file in the async action! - [x] don't recolor when switching to a different result in the same file! - [x] don't color matches whenever file path matches `helm-rg-shallow-highlight-files-regexp' - [ ] use `ripgrep' file types instead of flattening globbing out into `helm-rg-default-glob-string' - user defines file types in a `defcustom', and can interactively toggle the accepted file types - user can also set the default set of file types - as a dir-local variable!! - [ ] add testing - [ ] should be testing all of our interactive functions - in all configurations (for all permutations of `defcustom' values) - [ ] also everything that's called by helm - does helm have any frameworks to make integration testing easier? - [ ] publish `update-commentary.el' and the associated machinery - as an npm package, MELPA package, pandoc writer, *???* - [ ] make a keybinding for running `helm-rg' on dired marked files - then you could do an `f3' search, bounce to dired, then immediately `helm-rg' on just the file paths from the `f3' search, *which would be sick* License: GPL 3.0+ (./LICENSE) End Commentary