seq: fix default float precision (%f); fix 0 scientific printing
#7384
Conversation
|
GNU testsuite comparison: |
|
GNU testsuite comparison: |
|
Retrigger CI with a force push, but I think the error count difference came from |
There was a problem hiding this comment.
PR Overview
This PR addresses inconsistencies in float precision handling and scientific notation formatting, and enables additional tests for verifying these behaviors.
- Update default float precision to 6 across all variants.
- Fix the capitalization of 0 in scientific notation and remove redundant code.
- Enable and add tests for default formatting for %f, %g, and %E formats.
Reviewed Changes
| File | Description |
|---|---|
| src/uucore/src/lib/features/format/num_format.rs | Adjusts default precision and fixes scientific formatting for both zero and non-zero. |
| tests/by-util/test_seq.rs | Enables and adds tests to verify default precision and formatting behavior. |
Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.
|
GNU testsuite comparison: |
|
GNU testsuite comparison: |
| @@ -293,13 +293,7 @@ impl Formatter for Float { | |||
|
|
|||
| let precision = match precision { | |||
| Some(CanAsterisk::Fixed(x)) => x, | |||
| None => { | |||
| if matches!(variant, FloatVariant::Shortest) { | |||
There was a problem hiding this comment.
Hmm good catch. I was wondering how this came to be and it seems to be this: 4c5326f
Looks like that commit only looked at %g and check the other formats.
There was a problem hiding this comment.
|
Just did another rebase. Anything blocking this? (I have other changes on top of this that are starting to pile up...) |
|
GNU testsuite comparison: |
0.0E+00 was not capitalized properly when using `%E` format. Fixes uutils#7382. Test: cargo test --package uucore --all-features float Test: cargo run printf "%E\n" 0 => 0.000000E+00
The default precision is 6, no matter the format. This applies to all float formats, not just "%g" (aka FloatVariant::Shortest). Fixes uutils#7361.
Add tests for some of the default float formats (%f, %g, %E), mostly to check that the default precision is correctly set to 6 digits.
Those tests appear to have been fixed, enable them.
|
GNU testsuite comparison: |
|
Thanks ! |
The fact that printf and seq path are totally different is, TBH, a bit confusing...
printf uses precision 6 because of this subtle line in
spec.rs:Spec::write(Self::Floatcase):2 small additional changes (I can split to 2 other PR if needed, they felt small enough to just tack on this one):
%Ewas wrong, fix that as well.seq: Enable test_auto_precision and test_undefined
Those tests appear to have been fixed, enable them.
seq: Add tests for default float formats
Add tests for some of the default float formats (%f, %g, %E), mostly
to check that the default precision is correctly set to 6 digits.
uucore: format: Fix default Float precision in try_from_spec
The default precision is 6, no matter the format. This applies
to all float formats, not just "%g" (aka FloatVariant::Shortest).
Fixes #7361.
uucore: format: Fix capitalization of 0 in scientific formating
0.0E+00 was not capitalized properly when using
%Eformat.Fixes #7382.
Test: cargo test --package uucore --all-features float
Test: cargo run printf "%E\n" 0 => 0.000000E+00