scripter17 / jrep Goto Github PK

So --min df=50 waits until 50 files match before printing the directory name

Or something like that, I'm not sure of how to best implement this

Currently there are 18 of these and 6 more in-the-works for a total of 24

I'm not sure if merging them into a --limit-style option is a good idea. Actually scratch that I'm sure it isn't

I'm not entirely sure how to format the JSON

Maybe something like this?

{
	"directories":[
		{
			"name":"a",
			"directories":[],
			"files":[
				{
					"name":"aaa.py",
					"matches":[
						{
							"match":"abcdefg",
							"span":[0,6]
						}
					]
				}
			]
		},
		{
			"name":"b",
			"directories":[]
		}
	],
	"files":[]
}

Obviously with all the runData info

I doubt this'll work with --order ending with anything other than print-dir print-name print-matches or --depth-first. It could probably be done with considerable effort but the output would be exactly the same

It sucks we only have STDOUT and STDERR as output options

Attempting to do str.encode on a string with Unicode surrogate pairs (???) causes problems so for now I'm doing string.encode(errors="ignore")

Todo:

• Learn what the h*ck a Unicode surrogate pair is
• Give the user options to handle it better
• Find a way to make str.encode not throw errors

--count/--limit would take 1 or more args and the arguments would be file-match, fm, file, `total, etc.

I don't care if I have to make my own help formatter class I've done more for less

--limit tdt=1:5,2:3 would trigger after ./a, ./b, ./c ./d ./e OR ./a/a ./a/b ./a/c

Maybe have 1:5+2:3 or something be AND?

So instead of printing every match for regex 1 then every match for regex 2, etc. the matches could be interweaved to make further processing via JREP easier

With --name-sub and --dir-name-sub, it's possible to have multiple of the same filename in a directory

Maybe change --no-duplicates to a --limit-style option with a default option of tmt?

Currently it only cleans backslashes, carriage returns, and newlines. It ignores \033, the most common character in ANSI escape codes

--order replace sub match-test, --order match-test replace sub, etc

Maybe --file-regex xyz *2?

Not sure how to implement that