Closed kzaher closed 6 years ago
Merging #67 into master will increase coverage by
0.44%
. The diff coverage is93.1%
.
@@ Coverage Diff @@
## master #67 +/- ##
==========================================
+ Coverage 79.59% 80.03% +0.44%
==========================================
Files 10 10
Lines 1132 1137 +5
==========================================
+ Hits 901 910 +9
+ Misses 231 227 -4
Impacted Files | Coverage Δ | |
---|---|---|
Sources/PostgreSQL/Result.swift | 67.85% <100%> (+2.63%) |
:arrow_up: |
Sources/PostgreSQL/Bind/Bind.swift | 78.28% <92.85%> (+1.04%) |
:arrow_up: |
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 abba023...60f6b8c. Read the comment docs.
Hi.
Thank you for the fix, this is indeed a serious issue and a neat patch.
While technically this is indeed a change in the public API in theory, users of the library aren't supposed to use the Bind class directly. Most notably, the postgresql-driver module is completely unaffected by this change, so I'm in favor of just tagging and releasing this as 2.1.2 officially too.
This is one way how one could solve #66.
It breaks public interface, but I would argue that public interface had a design flaw in the first place because it was not clear who owns the passed buffer and who is responsible for cleaning up memory.
That was the actual cause for the bug in the first place.
I would suggest changing that interface to be more explicit, but if backwards compatibility is wanted, then there are other solutions.