pmrep csv output depends on py version

[Nathan Scott <nathans@xxxxxxxxxx>]

Hi Marko,

I'm seeing test qa/1069 produce different output depending on the
python version (2 vs 3).  It looks like we're not using precision
(self.precision, that is) in this case, should we be?  Maybe some
judicious use of "round(value, self.precision)" might fix this?

$ ./check -q -l 1069
1069 2s ... - output mismatch (see 1069.out.bad)
36c36
< sample.milliseconds,2000-05-01 20:36:47,s.milliseconds,util,380433679.314
---
> sample.milliseconds,2000-05-01 20:36:47,s.milliseconds,util,380433679.31412864
38c38
< sample.milliseconds,2000-05-01 20:36:49,s.milliseconds,util,380435679.381
---
> sample.milliseconds,2000-05-01 20:36:49,s.milliseconds,util,380435679.38064176
40c40
< sample.milliseconds,2000-05-01 20:36:51,s.milliseconds,util,380437679.382
---
> sample.milliseconds,2000-05-01 20:36:51,s.milliseconds,util,380437679.38192904
42c42
< sample.milliseconds,2000-05-01 20:36:53,s.milliseconds,util,380439679.413
---
> sample.milliseconds,2000-05-01 20:36:53,s.milliseconds,util,380439679.41344017
47c47
< sample.milliseconds|s.milliseconds|util|380433679.314
---
> sample.milliseconds|s.milliseconds|util|380433679.31412864
49c49
< sample.milliseconds|s.milliseconds|util|380435679.381
---
> sample.milliseconds|s.milliseconds|util|380435679.38064176
51c51
< sample.milliseconds|s.milliseconds|util|380437679.382
---
> sample.milliseconds|s.milliseconds|util|380437679.38192904
53c53
< sample.milliseconds|s.milliseconds|util|380439679.413
---
> sample.milliseconds|s.milliseconds|util|380439679.41344017


(python3 produces the greater-precision variant, FWIW)

cheers.

--
Nathan