Wrong @sprintf output with # flag

[mwlidar]

julia> @sprintf("%#.3g", 100.0 - eps(100.0))
"1000"

julia> @sprintf("%#.3g", 100.0)
"100."

This is with Julia 1.6.3

[inkydragon]

Test code

versioninfo()
using Printf
Printf.format(Printf.Format("%#.3g"), 100.0 - eps(100.0))
# should == "100."

• ✔️ 1.6.0 (2021-03-24) f9720dc
• ✔️ 1.6.1 (2021-04-23) 6aaedec
• ✔️ 1.6.2-pre.18 (2021-05-11 15:06 UTC) 6056838
• ❌ 1.6.2-pre.19 (2021-05-11 15:06 UTC) 4ab8c79
  make Printf's general format C11 compliant #40689 ping @bicycle1885
• ❌ 1.6.2 (2021-07-14) 1b93d53
• ❌ 1.6.7 (2022-07-19) 3b76b25
• ❌ 1.7.3 (2022-05-06) 742b9ab
• ❌ 1.8.0-rc4 (2022-08-08) 7853436
• ❌ 1.9.0-DEV.1125 (2022-08-13) e1fa6a5

[KristofferC]

Why was this added to the 1.8 milestone?

[vtjnash]

Showing the wrong answer seems quite a bad regression

[KristofferC]

Yes, but not vs 1.7. If all supported Julia versions are released with this bug, we surely can release 1.8 with it as well.

[gitboy16]

Maybe worth adding the backport 1.8 flag on this one?