Multiple PRs #2662
Multiple PRs #2662
Conversation
adithyaov
force-pushed
the
rm-imm-cloning
branch
from
6a78aba to
61d72d2
Compare
adithyaov
force-pushed
the
rm-imm-cloning
branch
2 times, most recently
from
b8b6bfa to
1875e5d
Compare
adithyaov
changed the title
adithyaov
force-pushed
the
rm-imm-cloning
branch
3 times, most recently
from
371720b to
89bcc28
Compare
adithyaov
force-pushed
the
rm-imm-cloning
branch
2 times, most recently
from
d15e389 to
1f6ac70
Compare
adithyaov
force-pushed
the
rm-imm-cloning
branch
from
ae6a9e1 to
abdb716
Compare
| in Skip (MECSBuffer remElem st1) | ||
| Yield arr st1 -> | ||
| let len = Array.length arr | ||
| in if (len .&. 1) == 1 |
| in Yield w16 | ||
| (MECSYieldAndBuffer arr1 lstElem st1) | ||
| Skip s -> Skip (MECSBuffer remElem s) | ||
| Stop -> Stop -- Here the last Word8 is lost |
There was a problem hiding this comment.
In case of strict decoding, this should be an error.
There was a problem hiding this comment.
This has nothing to do with utf16 decoding.
There was a problem hiding this comment.
This is only used in the benchmark for now anyway, so I don't want to spend more time on this. We can change this if this use-case ever arises.
There was a problem hiding this comment.
If it has nothing to do with it then we should not use it in decoding :-) Because decoding should fail if it finds the input is truncated. If this drops silently, decoder won't even know that some part of the input was dropped, it will just happily use whatever it got.
adithyaov
force-pushed
the
rm-imm-cloning
branch
2 times, most recently
from
f5e9741 to
1e7224a
Compare
| , decodeUtf16' | ||
| , decodeUtf16 |
There was a problem hiding this comment.
We can keep these unexposed. Don't know if we'll ever use this without taking byte-order into consideration.
- Add identity tests - Add benchmarks - Add Word16 to Word8 helpers
adithyaov
force-pushed
the
rm-imm-cloning
branch
from
1e7224a to
d8d7b29
Compare
| copyStreamUtf16 :: Handle -> Handle -> IO () | ||
| copyStreamUtf16 inh outh = | ||
| Stream.fold (Handle.writeChunks outh) | ||
| $ fmap Array.castUnsafe $ Array.chunksOf (arrayPayloadSize (16 * 1024)) |
There was a problem hiding this comment.
We can use defaultChunkSize instead.
There was a problem hiding this comment.
Keeping the number of elements constant for comparison.
No description provided.