fix: show default values for single boolean flags with default=False

[timonrieger]

Previously, single boolean flags (without secondary --no-* options) with
default=False would not display their default value in help text, even
when show_default=True was set. This made it inconsistent with dual-flag
options and other parameter types.

Removed the special case in _get_default_string() that hid defaults for
single boolean flags with falsy default values. Now all boolean flags
consistently show their default values in help text.

    dry_run: bool = typer.Option(
        False,
        "--dry-run",
        help="Simulate uploads without actually uploading",
    ),

Before: --dry-run Simulate uploads without actually uploading
After: --dry-run Simulate uploads without actually uploading [default: False]

Fixes inconsistency where show_default=True had no effect for single
boolean flags with default=False.

[timonrieger]

alternatively we could do

     elif (
         isinstance(obj, TyperOption)
         and obj.is_bool_flag
         and not obj.secondary_opts
         and not default_value
+        and not (obj.show_default or ctx.show_default)
     ):
         default_string = ""

[derOtto]

added some tests as pr to this pr: timonrieger#1

[derOtto]

alternatively we could do

     elif (
         isinstance(obj, TyperOption)
         and obj.is_bool_flag
         and not obj.secondary_opts
         and not default_value
+        and not (obj.show_default or ctx.show_default)
     ):
         default_string = ""

Colleagues of mine and I discussed the two options you suggested and we like this option here better.
We think the method _get_default_string should return the default string accordingly to the show_default parms.

[timonrieger]

Colleagues of mine and I discussed the two options you suggested and we like this option here better. We think the method _get_default_string should return the default string accordingly to the show_default parms.

sounds good. your tests cover the alternative implementation, right? I’ll wait for the maintainers response for the way to go, before merging it 😄

[derOtto]

Colleagues of mine and I discussed the two options you suggested and we like this option here better. We think the method _get_default_string should return the default string accordingly to the show_default parms.

sounds good. your tests cover the alternative implementation, right? I’ll wait for the maintainers response for the way to go, before merging it 😄

the tests should cover both implementations, as the desired functionality is tested, but the branch is forked from your first / current implementation