#128 quota(1) reports "none" for file systems with quota limits

[Petr Písař]

If a file system has set quota limits but the user does not occupy any space or inodes, quota tool reports "none":

# /tmp/linuxquota-git/quota -u test 
Disk quotas for user test (uid 500): none

and that looks like no quota limits are set, but that's not true as can verified in the verbose mode:

# /tmp/linuxquota-git/quota -u test -v
Disk quotas for user test (uid 500): 
     Filesystem  blocks   quota   limit   grace   files   quota   limit   grace
     /dev/loop0       0      10      20               0       0       0        
     /dev/loop1       3       0       0               1       0       0        

and thats confusing. (At least for me and some other users I talked to.)

Attached patch tracks whether there are indeed no quota limits and changes the message to be clear why no details were printed in the non-verbose mode.

Another question, not addressed with the patch, is why actually the non-verbose mode exist. I would assume reporting the details for every file system with quota limits set regardless of used space very helpfull. Is there any reason for that? I could not find the reason as the git history starts after introducing the -v option.

[Jan Kara]

This behavior (including -v option) has been there since I've started taking care of quota tools (i.e. before 1999) and I guess it was originally present in the BSD tools quota-tools have been derived from. I guess the rationale for this behavior is to skip reporting devices user isn't using to make output shorter but I admit the behavior was a bit confusing to me as well and I almost always use -v when testing quota changes. I just never dared to change the behavior due to backward compatibility concerns.

Looking at it now, I don't think there are too many scripts parsing quota(1) output without -v so I guess your change is fine. Maybe we could keep the original "none" output for the "no quota limits set" case and just add a new "no limited resources used"? Because in my experience that's what is causing most of the confusion...

[Petr Písař]

No problem. See the attached patch that replaces the original one.

[Jan Kara]

Thanks. Patch applied!