Repository navigation
Parse float32 fields with float32 precision in MapTo - #449
Merged
Merged
Conversation
Parsing as a float64 and then narrowing to a float32 rounds twice, so the shortest float32 string for some values, such as 7.038531e-26, read back one ULP off.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Describe the pull request
Follow-up to #448. That PR made
ReflectFromwrite float32 fields as the shortest float32 string, butMapTostill parsed them withstrconv.ParseFloat(str, 64)and then narrowed to float32. Rounding twice (text → float64 → float32) isn't always the same as rounding once. Running all 2³² float32 bit patterns through format → parse found one value, and its negative, that came back wrong: bits0x15ae43fdis written as7.038531e-26and read back one ULP off as7.0385313e-26. Before #448 this value round-tripped, because all 17 digits were written.parseValuenow parses withstrconv.ParseFloat(str, t.Bits()), so float32 fields round once, to the nearest float32. With this change, every finite float32 round-trips exactly.TestStructRoundTrip_Float32gets a case for7.038531e-26. It fails without the fix and passes with it.go test -race ./...,go vet ./...andgofmt -l .pass.One side effect: an out-of-range float32 such as
1e39now fails with strconv's "value out of range" error instead of "value … overflows float32". No test checks that message.Link to the issue: #447
Checklist
AI usage
Made with Amp (Claude). It found the double rounding by checking every float32 bit pattern while reviewing #448, then wrote the fix and the test and ran the checks above.