6
91%
master: 91%

Ran 09 Dec 2019 06:56PM UTC

Files 16

Run time 1s

Badge

Embed ▾

Committed 09 Dec 2019 06:45PM UTC coverage: 85.972% (-0.08%) from 86.052%

Job # PKGS="python=3.6 numpy=1.14"

Build Type

push

travis-ci

Committed by

web-flow

Commit Message

Merge #923

923: Preserve caches on context switch v2 r=hgrecco a=crusaderky

Fix bug in #909 where, every time a context is reactivated, it creates a brand new entry in UnitRegistry._caches - thus not only nullifying the whole point of the PR, but also creating a nasty memory leak for users that invoke ``.to`` with an inline context many times.

Not sure what went wrong with the first PR - if I rerun the included benchmark, it doesn't reproduce the same results anymore and clearly shows that nothing is cached.

Same benchmark with this PR:
```python
import pint
ureg = pint.UnitRegistry()
src = ureg('1 m')
dst = ureg.Unit('km')
%timeit src.to(dst)
%timeit -n 1 -r 1 src.to(dst, 'chemistry')
%timeit src.to(dst, 'chemistry')
ureg.enable_contexts('chemistry')
%timeit src.to(dst)
```
29.9 µs ± 1.4 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)
57.3 ms ± 0 ns per loop (mean ± std. dev. of 1 run, 1 loop each)
76.9 µs ± 1.69 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)
43.7 µs ± 955 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each)

Co-authored-by: Guido Imperiale <crusaderky@gmail.com>

Run Details

2911 of 3386 relevant lines covered (85.97%)

0.86 hits per line

hgrecco / pint / 1331 / 6
91%
master: 91%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 1331.6 (PKGS="python=3.6 numpy=1.14")

hgrecco / pint / 1331 / 6 91% master: 91%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 1331.6 (PKGS="python=3.6 numpy=1.14")

hgrecco / pint / 1331 / 6
91%
master: 91%

README BADGES
x