Open jpiautinity opened 6 years ago
Merging #110 into master will increase coverage by
0.01%
. The diff coverage is100%
.
@@ Coverage Diff @@
## master #110 +/- ##
==========================================
+ Coverage 21.5% 21.51% +0.01%
==========================================
Files 71 71
Lines 19914 19917 +3
Branches 5728 5729 +1
==========================================
+ Hits 4283 4286 +3
- Misses 14922 14924 +2
+ Partials 709 707 -2
Impacted Files | Coverage Δ | |
---|---|---|
lib/SQL/Translator/Parser/MySQL.pm | 71.72% <100%> (-0.82%) |
:arrow_down: |
lib/SQL/Translator/Producer/MySQL.pm | 91.08% <0%> (+0.55%) |
: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 86c68a0...d19e2c6. Read the comment docs.
Summary: This merge request removes the forced 2 precision and 8 digits size limit for floating point number data types in the MySQL field normalization and instead force removes size for these data types.
Some unit tests are updated to reflect this change.
Reason for this change: While sane size limits are useful for fixed point data types like decimal such limits make little sense for floating point data types. Setting the size for floating point data types does not result in less storage being used but will just try to badly truncate the input and mask the output (or raise an error in strict mode on bad input).
For use cases where size restricted data is required or desired it is highly recommended to use fixed point data types as they don't have impression errors in their given number space as opposed to IEEE 754 floating point numbers (As an example: 0.1 and 0.2 are not directly representable as IEEE 754 number no matter how good the precision is). IEEE 754 floating point numbers are simply not designed for such use cases but to closer approximate natural numbers.
Being able to set size limits to floating point numbers can also encourage the use of bad practices like the classic case of money represented as floating point numbers.
Lastly, the MySQL documentation recommends not using precision and number of digits for maximum portability. https://dev.mysql.com/doc/refman/en/floating-point-types.html