strformat float printing doesn't print ".0" portion [regression]

[kaushalmodi]

Earlier strformat printed a float 1.0 as "1.0". Now it prints just "1".

It was nice to have that ".0" portion printed that gave a clarity that the value being printed was a float.

Example

import strformat

let
  foo: float = 1.0
  bar: float = 1.5
echo fmt"{foo}" # Prints "1" -- bad
echo fmt"{bar}" # Prints "1.5" -- good

Current Output

1
1.5

Expected Output

1.0
1.5

Possible Solution

Additional Information

$ nim -v
Nim Compiler Version 0.19.9 [Linux: amd64]
Compiled at 2019-04-22
Copyright (c) 2006-2019 by Andreas Rumpf

git hash: 4f9366975441be889a8cd4fbfb4e41f6830dc542
active boot switches: -d:release

/cc @Araq @krux02

---

I discovered as I was rerunning all the snippets in my Nim notes with the latest devel. git diff highlighted all the missing ".0"'s:

[Araq]

This was never explicitly in the program though as far as I know. Did your libc change?

[kaushalmodi]

Hmm, I definitely haven't updated my libc (glibc that ships with the OS, right?). My OS version is still RHEL 6.8.

Are you not able to reproduce the same ".0" truncation?

[kaushalmodi]

@Araq The regression of this behavior from stable to devel was confirmed on IRC.

Ref: https://irclogs.nim-lang.org/23-04-2019.html#07:01:10

[kaushalmodi]

If it helps trace down this problem, in my git diff, I had also noticed the resolution reduction when printing floats in general too. Printing 1.23456789 printed something like 1.23456 instead. I can share the exact examples when I get to my PC.

[kaushalmodi]

@Araq Thank you for that quick fix! As I was past this regression fixes and evaluating other blocks in my notes, I ended up with #11092.