Correct "-n" behaviour when going past 100% of "-s" size

[Andrew Wood]

Date: Wed, 31 Oct 2018 14:04:58 +0100
From: Marcel
Subject: pv -n -s - unexpected/wrong calculation

Hi Andrew,

I hope you don't mind if I approach you directly. I am not sure if it
is a bug or by design, but I found that pv does not behave as
expected.

Lets assume I would like to tar 200.000 bytes. If I set size=100000
and run pv like so:

tar cf - . | pv -n -s $size

I will get an output that looks for example like this simplified example:

10

40

90

100

90

40

10

And if i set size smaller than 100.000 pv would even go to negative numbers.

I actually had expected that values would simply go above 100(%) or
that pv would stop at 100 (maybe even raise a warning, if told so).

Maybe you could enlighten me in this matter?

Cheers,

Marcel

---

Date: Wed, 21 Nov 2018 19:59:14 +0000
From: Andrew Wood
To: Marcel
Subject: Re: pv -n -s - unexpected/wrong calculation

Thanks for your email - sorry for the delayed response.

I've not had a chance to have a look through the code, but what happens when
you don't use -n - does it show the same behaviour? Also which version of
PV are you using?

---

Date: Sat, 24 Nov 2018 22:45:51 +0100
From: Marcel
To: Andrew Wood
Subject: Re: pv -n -s - unexpected/wrong calculation

Hi Andrew,

thanks for your reply. Here is what I tried:

tar --xz -cf - . | pv --interval 0.5 -s 10000 | split -b 1G - /tmp/testme_

The size is way to small and as expected I got negative numbers with -n.

Without -n, I get the following:

^C47MiB 0:00:02 [3.38MiB/s] [=======================================================] 88801% ETA 0:00:00

So, the percentage also seems to be affected and ETA stays 0:00:00 all the time.

Hope that helps. If I can be of any further help, please do not hesitate and get back to me!

pv --version = 1.6.6

Cheers,
Marcel

---

Date: Sun, 28 Mar 2021 20:06:40 +0100
From: Andrew Wood
To: Marcel
Subject: Re: pv -n -s - unexpected/wrong calculation

My apologies for the exceptionally long delay here!

I've had another look and I can reproduce the problem with "-n", in that I
get negative numbers and general nonsense, but normal display mode works
fine for me (i.e. I get the right % values). I wouldn't expect ETA to do
anything sensible anyway.

I will add a note to the TODO to investigate the behaviour of "-n" when the
input byte count exceeds the value of "-s" since it is clearly wrong.

[Andrew Wood]

This is corrected in the next release after 1.6.20, and was due to some extra conditions in numeric mode that had been mistakenly copied from the visual progress display.