Closed luizirber closed 4 years ago
Merging #35 into master will not change coverage. The diff coverage is
100.00%
.
@@ Coverage Diff @@
## master #35 +/- ##
=======================================
Coverage 98.89% 98.89%
=======================================
Files 2 2
Lines 181 181
=======================================
Hits 179 179
Misses 2 2
Impacted Files | Coverage Δ | |
---|---|---|
src/lib.rs | 100.00% <ø> (ø) |
|
src/compression.rs | 97.93% <100.00%> (ø) |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 4b9ec1c...4d61c07. Read the comment docs.
So, what are the drawbacks of adding
Send
? Other than requiring bumping niffler to3.0.0
? =P
I also tried using this branch in sourmash and decoct, and both worked fine (I had to fix one place in sourmash where I was asking for a R: io::Read
instead of R: io::Read + Send
, and hence why the bump to 3.0.0
). Does anyone know about something that implements Read
but is not Send
?
Does anyone know about something that implements Read but is not Send?
I don't think there's any but don't quote me on that... I'm sure there's someone out there with a custom Read implementation that is not Send
So, what are the drawbacks of adding Send?
I didn't see any, for me we can merge this pull request
Other than requiring bumping niffler to 3.0.0?
Yes but I want to integrate get_reader_seek
and get_writer_seek
and bgzip (we can do this easily) before release 3.0.
needletail had to rollback niffler adoption because PyO3 requires a Send bound. This PR adds
Send
bounds to all places where aBox<dyn Read>
orBox<dyn Write>
is used, and... it all passes the tests?So, what are the drawbacks of adding
Send
? Other than requiring bumping niffler to3.0.0
? =P(similar discussion: https://fasterthanli.me/articles/getting-in-and-out-of-trouble-with-rust-futures)