ellocate

ellocate copied to clipboard

[[https://melpa.org/#/ellocate][file:https://melpa.org/packages/ellocate-badge.svg]]

Ellocate is a re-implementation of locate written in Emacs-Lisp. It supports all completion backends (ivy, helm, ido, etc)

• Comparison #+caption: Comparison with other file searchers | | Ellocate | Locate (and elisp wrappers like counsel-locate) | Find (and elisp wrappers like counsel-find) | |-------------------------------------------------------------+------------+-----------------------------------------------------------+----------------------------------------------| | Speed | Fast | Fast | Slow | | Easy to configure multiple search paths | Yes | No | Yes (always searches down current directory) | | Switch search databases used depending on current directory | Yes | With a lot of effort | N/A | | Works the same on every platform that has the dependencies | Yes | No (Some systems use mlocate while others use GNU locate) | Yes | | Dependencies | Find | mlocate or GNU locate | Find | | Tramp support | I think so | No | Yes |

#+html:

• Usage The function =ellocate= scans recursively down the directory and creates a database file to store the results in as specified by =ellocate-scan-dirs=. By default =ellocate= only searches all directories below the current directory. To search all files in the current database either run =ellocate-all= or run =ellocate= with a universal argument =ellocate-scan-dirs= is configured like this: #+BEGIN_SRC emacs-lisp (setq ellocate-scan-dirs '(("~/" "~/ellocate-home-db") ("/mnt/USB" "~/ellocate-usb-drive-db") ;; I want this directory to be re-scanned on first search after every emacs restart by not creating a database file for it ("/mnt/USB2" nil) ;; Never make your scan paths overlap like this: ;; ("/mnt/USB2/newFolder" nil) ;; This is pretty useful if you have many USB devices, but remember to not overlap like this would if it wasn't commented (because the subdirectory /mnt/USB is also scanned as defined above) ;; ("/mnt/" nil) )) #+END_SRC Where the first element in all of the lists inside =ellocate-scan-dirs= is the base directory to scan and the second is in which file to store the search results so that =ellocate= can read read it if Emacs is restarted to give =ellocate= faster initial startup times. If the database is nil, =ellocate= will have to re-scan the directory if run every time Emacs is restarted instead of just reading a database containing the scanned files.

=ellocate-clear= clears the cache corresponding to your current directory and =ellocate-clear-all= clears all defined caches. These both forces =ellocate= to refresh its search entries. Do this if your file-system has changed and you want that change reflected in =ellocate=.

• Q&A ** Why not just use locate with one huge database containing everything? Let's say you have a 2 TB HDD mounted in =/dev/hdd= that you occasionally want to search through, but right now you want to find something new in =/home/user= but to do that you have to generate a new locate search database which means it has to search through everything inside =/dev/hdd= and that will probably take like 30 minutes if it's full of files. With ellocate on the other hand you can selectively delete and reconstruct the search database corresponding to =/home/user=. You also have a lot more options like you can tell ellocate to not save the database you build from =/dev/hdd= to disk, so that on every restart after running ellocate you automatically re-scan it. I think you should also be able to create search databases over ssh connections, which is impossible with locate unless you mount the remote computer as a local drive.