[Feature Request] Keep old results and read path list from file (for incremental scans)

[RaymondBakerIBM]

To facilitate incremental scans, we need to keep old results in the baseline instead of the default behavior of deleting all files which weren't scanned in the current scan.

Sometimes there will be too many paths to pass into the cli (larger than ARG_MAX) so there needs to be another way to provide paths. stdin is already used for the baseline file so pass it in as a path for the program to read.

Associated PR
#153

[Maciej-Sitarz-IBM]

I also encountered a problem that required passing a list of several files to detect-secrets as CLI args.
Solution that I implemented in my projects was to use xargs:

sort -u <(git ls-files) |
  tr '\n' '\0' | 
  split --separator='\0' -n r/1/4 | 
  xargs -r0x detect-secrets -v scan --update .secrets.baseline ...

Explanation of each part of the above oneliner:

• sort -u <(git ls-files) - returns all files from the repo
• tr '\n' '\0' - replaces new line characters with NUL (so xargs can handle files with spaces in filenames)
• split --separator='\0' -n r/1/4 - splits the list into chunks and processes the first one (allows parallel computing)
• xargs -r0x detect-secrets -v scan --update .secrets.baseline ... - xargs passes the line from stdin as CLI args to detect-secrets

The remaining issue is that when xargs receives more lines than ARG_MAX it'll call detect-secrets multiple times.
Each such call to xargs call to detect-secrets causes non scanned files to be removed from .secrets.baseline.
This seems to be fixed by the mentioned PR:

• Allow for incremental scans #153