thedataleek / black-arrow Goto Github PK

Currently for very large results the numbers aren't great. Adding some formatting to help readability would be nice.

---------------
Files Searched: 11100
Files Matched: 2
Lines Searched: 11290394
Duration: 16.575687885284424

Turn on case insensitivity by default and only on if flag or uppercase in search term

Current behavior when sometimes using ctrl-c to interrupt -> lots of errors.

Expected behavior -> Nice happy cleanup

Why should someone use this? Some examples of what it can do would help convince people that it's useful and cool. Screenshots would be nice to preserve the colors in the output

Need to be able to configure the run based on some config file (either global or local)

Requested, would be nice to have a replace feature built in.

I've found a condition where the final .join() call in main hangs. It appears to be related to the number of matches and not to the number of files. I've hacked a bit and found that all of the processes that are started think they have exited (at least by printing on the last line of the function).

Here's the setup:
linuxMint 64bit
python 3.7.4
I'm using this repo (just happened to be sitting on my filesys): https://github.com/PySimpleGUI/PySimpleGUI

Searching for "asdfadsf" exits correctly. Searching for "the" hangs indefinitely.

I'm suspicious that a queue is holding a process open, but haven't gotten that far yet.

I'll dig some more, but I thought I'd go through the process and raise an issue. :) Thanks for sharing this project! Super cool!

Would be interesting to compare against the silver searcher, grep/find, etc.

Grep has this feature built in

https://stackoverflow.com/questions/9081/grep-a-file-but-show-several-surrounding-lines

Would be nice to replicate as well.

This might be impossible. Currently comparing to ag, blackarrow is about 3 times slower than ag. This is pretty good for a pure-python approach vs whatever compiled language ag is written in. but can we do better?

I need to create a writeup that explains the multi-processing architecture (I have a few diagrams floating around somewhere...)

I also need to add more comments explaining step-by-step how/what/why/etc.

Add fine tuning over how many levels deep to search

As referenced in #14 in a conversation with @mohitabz, the current implementation of the depth'ing functionality will not behave correctly when absolute paths are provided.