Basic refactoring of HashMap #32058
Basic refactoring of HashMap #32058
Conversation
|
travis is failing |
|
Sadly, I am currently contractually barred from reviewing this code. r? @cgaebel |
|
Ah shoot. r? @apasel422 |
|
uhhhh r? @bluss ??? |
|
bors where r u |
|
no homu, no highfive, aaaah |
|
I'll take a look when travis passess. |
pczarn
force-pushed
the
hashmap-initial-refactoring
branch
from
c9c4e93 to
d67cf45
Compare
|
Sorry, forgot to push changes |
|
|
|
fixed |
| } | ||
| } | ||
|
|
||
| impl<K, V, M> DerefMut for FullBucket<K, V, M> where M: DerefMut<Target=RawTable<K, V>> { |
There was a problem hiding this comment.
I don't think that this impl is safe - giving arbitrary access to the underlying RawTable means that a user could resize or replace the RawTable out from under the FullBucket, making it invalid.
There was a problem hiding this comment.
An alternative would be to put an equivalent but unsafe method on Put to fetch the table. By making the method unsafe, any violation of memory safety would still have to use unsafe code.
|
Added fixes for @gereeter's review. |
|
The type signatures are getting a little crazy for a simple hashtable, but that's neither the singular doing of this patch, nor necessarily uncalled for. Looks good to me. |
|
⌛ Testing commit 64adca7 with merge 1abcf47... |
|
💔 Test failed - auto-win-msvc-64-opt-rustbuild |
|
@bors: retry On Tuesday, March 22, 2016, bors notifications@github.com wrote:
|
No description provided.