Re: version 2.2 --remove-older-than behavior changed

[Eric Zolf]

Hi,

On 16/02/2023 20:41, qx6uwumzvv@liamekaens.com wrote:
> IMHO, "No increment is older than ...." is normal behavior and should 
> not cause a warning, especially since nonzero status codes are 
> considered abnormal termination by most shells (even the Python docs 
> <https://docs.python.org/3/library/sys.html#sys.exit> mention this), and 
> can cause backup scripts to terminate.

If there is a warning message, and there always was a warning message, 
then it's not a "normal" behavior, at least not in my books.

For example, if you're short on disk space, want to remove some 
increments, and choose the wrong selector, then you'd be happy to notice 
that nothing was removed (and your next backup might fail).

So, yes, this is abnormal (or do you call the command just to have it 
remove nothing?) but not a fatal error.

Check the man page of a few commands and you'll see that it's not 
uncommon (ls, grep, rsync...).

KR, Eric

> On 2023-02-16 10:25, ewl+rdiffbackup-at-lavar.de |rdiff-backup-users| 
> wrote:
> > Change log and man page indeed:
> >
> > CHG: return codes have changed and are now more detailed, see man-page 
> > for details
> >
> > CHG: start to introduce more specific return codes (1 for error, 2 for 
> > warning, 4 for single file error, etc), which can be combined (e.g. 3 
> > for error and warning), this will take time to introduce everywhere
> >
> >
> > On 16 February 2023 13:06:27 GMT+01:00, tbsky<tbsky@annsky.com>  wrote:
> > > ewl+rdiffbackup@lavar.de  於 2023-02-16 15:04:
> > > > Did you check the documentation? Tl;Dr: it is documented!
> > > I am sorry. I only "man rdiff-backup", but I didn't see anything 
> > > about the changed behavior.
> > > what document should I read?
> > > thanks a lot for help.