Thank you, @jg-rp, for identifying this bug and helping to implement the fix.
Summary
This PR addresses a bug in the sum filter when handling arrays containing floats and thus yield a result that is a float.
Previously, the sum filter would return a BigDecimal, which would render in scientific notation, when summing floats. This behaviour was inconsistent with other math filters.
The change in this PR ensures that the sum filter returns a float instead of a BigDecimal when summing floats, aligning its behaviour with other math filters.
Thank you, @jg-rp, for identifying this bug and helping to implement the fix.
Summary
This PR addresses a bug in the
sumfilter when handling arrays containing floats and thus yield a result that is a float.Previously, the
sumfilter would return aBigDecimal, which would render in scientific notation, when summing floats. This behaviour was inconsistent with other math filters.The change in this PR ensures that the
sumfilter returns a float instead of aBigDecimalwhen summing floats, aligning its behaviour with other math filters.For example:
https://github.com/Shopify/liquid/blob/0b9318222bcc09681e52fd5b8e70262274e673bf/lib/liquid/standardfilters.rb#L908-L911
Changes
sumfilter inlib/liquid/standardfilters.rbto convert BigDecimal results to floats when summing floats.test/integration/standard_filter_test.rbto verify the correct behaviour of thesumfilter when handling floats.