Help not really helpful; several feature requests

[DerPoet]

--help does not tell which values are available for -c

And I would really appreciate if

• there were an option --csv
• I could type abbreviated arguments, e.g. --units b or -c +ino
• I had an option --no-bar-graphs to leave out the bar graphs
• there were options for restricting/filtering the filesystems shown, e.g. --type ext4,xfs or --no-type tmpfs or --min-use 90% or --disk lvm,ssd or maybe something like --filter 'type=ext4,xfs and disk!=ssd and use>90'
• there were an option --no-header (especially with --csv)

[DerPoet]

and an option for sorting by any column and an --asciioption for output in plain ascii (for dumber terminals in the n-th login level)

[Canop]

So many questions. Let's handle them one per one

1) --help does not tell which values are available for -c.

That's true. I'm not sure of how I'll handle this without replacing the library I use for arg parsing. In the meantime, I just improved the error message:

I'm pondering adding a --list-cols which would output a table of all column names and their meaning.

2) The --csv option : No objection. Please make a separate dedicated issue.

3) I'll add other abbreviated column values. Suggestions welcome

4) --no-bar-graphs : It would be some new columns with just the percentages use_percent & inodes_percent. I note to add those

5) Are those filterings really useful ? If it's for post-treatment you can use jq. What's the reason here ?

6) the --no-header option doesn't look terribly useful

7) sorting: see https://github.com/Canop/lfs/issues/37

8) --ascii : No objection. Please make a separate dedicated issue.

[Canop]

(in the next release)

[DerPoet]

Hello Canop, thank you for your quick response and the work you do! Even more thanks for implementing so many of the suggestions coming from outside. As you asked, I've opened separate issues for --csv and --ascii.

Regarding the filtering: Sure, it's not really necessary, jq/grep/awk/... can do the same thing. It would just be nice to do the filtering within the tool. And it would, for example, avoid mistakes by having a grep matching also an entry in a non-intended column. (The older I become, the more I'm in for safe code, after having seen the weirdest ways that things go wrong. Maybe that's only my personal quirk.)

Best regards, Christian

[Canop]

I'm not at all in favor of grepping the output of lfs, I agree it's too dangerous. The alternative I would use is json + jq.

I'll certainly do the filtering right in lfs if I feel like it would be used by several users. It mostly depends on whether many people start to use lfs and would need this feature. It could be a separate issue and we'd see if people chime in to say they're interested.

[DerPoet]

Okay, I've added a comment to #41 . Thank you.

[lemmy04]

I would love to have a proper manpage.

[Canop]

@lemmy04 Can you please crate a separate issue for that, so that we don't forget?

[Canop]

Most of what's here has been moved to other issues. Closing this one.