Feature suggestion : allow files filtering when used with -recursive

[PaulRivier]

Hi,

thank you very much for cshatag.

I would like to use it for my significant files, while avoiding the overhead (CPU, IO, time, space) of checking thousands of non-significant files for which a silent data corruption would be mostly without consequences. An analogy can be made with what backup software usually provide, or what git and .gitignore provide, where a set of patterns allows to include or exclude files. On this topic, restic backup is also written in Go, and have some nice filtering/inclusion/exclusion features.

I could do some hacks with 'find' and 'xargs' to have a 'cshatag' exec call on each eligible file, but that would slow down execution.

Would you consider adding some support for more fine-grain inclusion than whole directory content ? Maybe the most direct way to implement this would be to accept a file as input, either parameter or stdin, with 1 line per file-path to process, and then just loop over it ?

What do you think ?

Regards,

[midnight-wonderer]

I could do some hacks with 'find' and 'xargs' to have a 'cshatag' exec call on each eligible file, but that would slow down execution.

I did this and noticed no significant slowdown at all.
Are you sure there is nothing wrong with your find or xargs arguments?
Despite the line length limitation, I never observed a hit (multiple cshatag invocations), and just a single cshatag invocation takes hours. (System Monitor crashed if I mouse over the cshatag process to reveal the command. The command arguments can be very long.)

Hardly imagine a scenario where filenames via STDIN will be a noticeable improvement.