Bug 22940 – Underscore disallowed after decimal separator in floating point literals

https://dlang.org/spec/lex.html#floatliteral says: Floating literals can have embedded ‘_’ characters to improve readability, and which are ignored. This code (zzz2.d) void main () { real r = 1._0L; } when compiled with dmd 2.099 gives: $dmd zzz2 zzz2.d(3): Error: no property `_0L` for type `int` expected: Software should behave as the manual says or the other way round.

The paragraph is misleading but that behavior matches the grammar in the spec, hence marking it as an enhancement.

Changing the component to the spec because this issue most likely isn't fixable on the compiler side. `_0` is a valid identifier and hence would introduce an ambiguity for `1._0L` (is it a double literal or UFCS call `_0(1)`?)

@ntrel created dlang/dlang.org pull request #3642 "Lex float" fixing this issue: - Fix Issue 22940 - Underscore disallowed after decimal separator in floating point literals https://github.com/dlang/dlang.org/pull/3642

dlang/dlang.org pull request #3642 " [spec/lex] Improve floating point docs" was merged into master: - b1de8c3a4b46a908badbb4cb9914e27ff41dd04c by Nick Treleaven: Fix Issue 22940 - Underscore disallowed after decimal separator in floating point literals https://github.com/dlang/dlang.org/pull/3642