doc: add a notice that unchecked resolving is faster for BufferBackend
#92
Conversation
|
Codecov ReportAll modified and coverable lines are covered by tests ✅
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. Additional details and impacted files@@ Coverage Diff @@
## master #92 +/- ##
==========================================
- Coverage 61.10% 60.98% -0.12%
==========================================
Files 12 12
Lines 509 510 +1
==========================================
Hits 311 311
- Misses 198 199 +1 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
It says outdated deps, not sure if that PR has to do anything with it |
Robbepop
left a comment
Robbepop
left a comment
There was a problem hiding this comment.
Thanks for the PR! Indeed this is probably a useful information to users.
| //! | ||
| //! 2. `BufferBackend` is slow with checked resolving because its internal representation | ||
| //! is extremely sensible to the correctness of the symbols, thus a lot of checks | ||
| //! are performed. If you will only use symbols probided by the same instance of |
There was a problem hiding this comment.
| //! are performed. If you will only use symbols probided by the same instance of | |
| //! are performed. If you will only use symbols provided by the same instance of |
|
@Robbepop I didn't fix the typo, it is in the repo now |
Ah sorry I forgot about that. Feel free to open another PR to fix it if you want. |
3 participants
Judging only by docs it looked like
BufferBackendis extremely inferior. After checking it's implementation, running benchmarks and comparing implementation to other backends it seems like absolutely superior backend, assuming theSymbolis correct.This PR adds a notice that
resolve_unchedkedis quite fast forBufferBackend, exceeding other backends (23-28 % faster on my machine).