Fix floating point formatting, #799 #801
Conversation
|
Took me a while to realise that render_int would actually take a floating point number and internally do the right thing with the sign |
stdlib/std.jsonnet
Outdated
| @@ -535,7 +535,7 @@ limitations under the License. | |||
| local whole = std.floor(n_); | |||
| local dot_size = if prec == 0 && !ensure_pt then 0 else 1; | |||
| local zp = zero_pad - prec - dot_size; | |||
| local str = render_int(std.sign(n__) * whole, zp, 0, blank, sign, 10, ''); | |||
| local str = render_int(n__, zp, 0, blank, sign, 10, ''); | |||
There was a problem hiding this comment.
Could you modify the comment for render_int explaining what it does with non-whole numbers? This behavior is counter-intuitive, so it's better to have it spelled out explicitly.
There was a problem hiding this comment.
Have attempted to do it less hackily now
| @@ -189,6 +189,7 @@ std.assertEqual(std.format('%-12.4f', [910.3]), '910.3000 ') && | |||
| std.assertEqual(std.format('%#.0f', [910.3]), '910.') && | |||
| std.assertEqual(std.format('%#.0f', [910]), '910.') && | |||
| std.assertEqual(std.format('%.3f', [1000000001]), '1000000001.000') && | |||
| std.assertEqual(std.format('%f', [-0.1]), '-0.100000') && | |||
There was a problem hiding this comment.
I would also add `std.format('%d', [-0.1])'. The result will be '-0', I think. Not sure if that's ok.
There was a problem hiding this comment.
I think it's not OK because it's different to what Python does.
stdlib/std.jsonnet
Outdated
| // min_chars must be a whole number >= 0 | ||
| // It is the field width, i.e. std.length() of the result should be >= min_chars | ||
| // min_digits must be a whole number >= 0. It's the number of zeroes to pad with. | ||
| // blank must be a boolean, if true a postive number has a ' ' in place of the '-'. |
There was a problem hiding this comment.
The detailed explanation is good, but I think these lines are confusing.
- "blank must be a boolean, if true a postive number has a ' ' in place of the '-'." <- It's weird to say that a positive number has something in place of '-'. A positive number has no '-'. Perhaps it would be better to say that it adds an additional ' ' in front of the number, so that it is aligned with negative numbers with the same number of digits.
- plus must be a boolean, if true a '+' is used in place of the '-'. <- Same, but even more confusing. Naively, it sounds like we would format
-5as"+5"whenplusis true.
| @@ -53,6 +53,27 @@ std.assertEqual(std.format('thing-%-5.3d', [10.3]), 'thing-010 ') && | |||
| std.assertEqual(std.format('thing-%#-5.3d', [10.3]), 'thing-010 ') && | |||
| std.assertEqual(std.format('thing-%#-5.3i', [10.3]), 'thing-010 ') && | |||
| std.assertEqual(std.format('thing-%#-5.3u', [10.3]), 'thing-010 ') && | |||
| std.assertEqual(std.format('thing-%5.3d', [-10.3]), 'thing- -010') && | |||
There was a problem hiding this comment.
Nice to have all these cases covered!
No description provided.