Merge bash_completions changes from upstream

[dioni21]

Motivation and Context

The current bash_completion contrib code in openzfs is very old, and
some changes have been added since.

Description

The original repo is at https://github.com/Aneurin/zfs-bash

I just made a diff and brought it all in a single chunk

How Has This Been Tested?

I've been using the original @Aneurin code since my first deploy of ZoL.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance enhancement (non-breaking change which improves efficiency)
• Code cleanup (non-breaking change which makes code smaller or more readable)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation (a change to man pages or other documentation)

Checklist:

• My code follows the ZFS on Linux code style requirements.
• I have updated the documentation accordingly.
• I have read the contributing document.
• I have added tests to cover my changes.
• I have run the ZFS Test Suite with this change applied.
• All commit messages are properly formatted and contain Signed-off-by.

[dioni21]

@ggzengel would you please take a look? The last commit in this file was yours, and it motivated me to do the update.

[dioni21]

BTW: I was not sure if it was safe to keep the shopt -u hostcomplete command in here. Reviewers, what do you think?

[Aneurin]

BTW: I was not sure if it was safe to keep the shopt -u hostcomplete command in here. Reviewers, what do you think?

Just for some context to that, the rationale for disabling hostname completion is a combination of two things:

• Most people using this will also be using bash_completion, which disables hostcomplete.
• For everyone else, completion of snapshot names won't work unless hostcomplete is disabled, which is a near-existential problem for ZFS command-line completion.

It's been several years but IIRC there was no other way to opt out of hostcomplete just for a single command. Also I think the old version in practice simply didn't work at all without bash_completion, because the commit that includes this change is titled "Remove dependency on bash_completion" (Aneurin/zfs-bash@1bd8d4f).

[behlendorf]

It's good to see these updated and the changes look good to me but I haven't tested them myself. I'm all for pulling them in as long as @Aneurin doesn't have any objection.

Not to be greedy, but we have added several new subcommands over the last few years (zpool checkpoint|initialize|resilver|trim|wait) and it would be nice to update the script in a separate PR to cover them.

[Aneurin]

I certainly have no objection.

I've never used any of the new commands (actually, never heard of some of them) so I've not had an impetus to make any updates for a few years, but I'll review the current manpages with a view to adding the missing commands and options. That's unlikely to happen before the weekend though.

[codecov]

Codecov Report

Merging #10456 into master will decrease coverage by 0.78%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master   #10456      +/-   ##
==========================================
- Coverage   80.24%   79.46%   -0.79%     
==========================================
  Files         293      393     +100     
  Lines       83882   123859   +39977     
==========================================
+ Hits        67313    98427   +31114     
- Misses      16569    25432    +8863     

Flag	Coverage Δ
#kernel	80.04% <ø> (-0.21%)	⬇️
#user	65.55% <ø> (?)

Impacted Files	Coverage Δ
module/os/linux/spl/spl-zlib.c	55.35% <0.00%> (-28.58%)	⬇️
module/zfs/zcp_set.c	87.09% <0.00%> (-12.91%)	⬇️
module/zfs/zfs_ratelimit.c	90.47% <0.00%> (-9.53%)	⬇️
module/zfs/dmu_redact.c	82.15% <0.00%> (-6.24%)	⬇️
module/zfs/zfs_byteswap.c	79.45% <0.00%> (-5.85%)	⬇️
module/zcommon/zfs_deleg.c	92.45% <0.00%> (-5.55%)	⬇️
module/zcommon/zfeature_common.c	92.85% <0.00%> (-4.82%)	⬇️
module/zfs/dsl_deleg.c	90.07% <0.00%> (-4.48%)	⬇️
module/zfs/zio_inject.c	75.80% <0.00%> (-4.41%)	⬇️
module/zfs/fm.c	36.77% <0.00%> (-4.10%)	⬇️
... and 278 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 883a40f...f041c9c. Read the comment docs.

[dioni21]

I am super curious, and did check what @Aneurin said:

• Most people using this will also be using bash_completion, which disables hostcomplete.

Indeed, it is set in my system, Fedora 31, at /usr/share/bash-completion/bash_completion

Maybe we could agree on a safe path:

• Do not unset hostcomplete in here, let bash_completion be responsible for that
• Add some readme file explaining that user should unset it if NOT using full bash_completion pack

Sorry for my inexperience, but should I do this as another commit to my branch or as a force push?

Also, what if a subsequent PR could make this installable? Probably best as a separate sub-package. I can do the RPM specs change, but I do not feel comfortable messing with other package systems.

[behlendorf]

Sorry for my inexperience, but should I do this as another commit to my branch or as a force push?

I think it makes sense to merge this as is. I'll go ahead and get it merged right away.

Also, what if a subsequent PR could make this installable? Probably best as a separate sub-package. I can do the RPM specs change, but I do not feel comfortable messing with other package systems.

This we should definitely do as a subsequent PR. Making the change to the RPM spec file should be sufficient since we don't maintain other native packaging. Would we really need a sub-package, is there any harm in installing them as part of the zfs package?

[dioni21]

Would we really need a sub-package, is there any harm in installing them as part of the zfs package?

I was afraid to add another dependency (bash-completion), but I just looked into how other packages do, and they add its own completion file without requiring the main framwork. Ok, I'll see into adding it to zfs main package ASAP.