nk_do_property() uses nk_strtod() instead of NK_STRTOD()

[sleeptightAnsiC]

You can provide your own definition of NK_STRTOD(), but nk_do_property() will ignore this and still use internal nk_strtod() instead of user-provided macro:

Nuklear/src/nuklear.h

Lines 3818 to 3821 in fcd64f8

	#ifndef NK_STRTOD
	#define NK_STRTOD nk_strtod
	NK_API double nk_strtod(const char *str, char **endptr);
	#endif

Nuklear/src/nuklear_property.c

Lines 283 to 287 in fcd64f8

	case NK_PROPERTY_DOUBLE:
	nk_string_float_limit(buffer, NK_MAX_FLOAT_PRECISION);
	variant->value.d = nk_strtod(buffer, 0);
	variant->value.d = NK_CLAMP(variant->min_value.d, variant->value.d, variant->max_value.d);
	break;

Should be an easy fix, unless there are more issues like this one, or if there is some hidden dependency issue that I missed.

[RobLoach]

Thahnks for putting together these follow up issues.

[PavelSharp]

This indeed resolves the inconsistency that currently exists in Nuklear.
In the code, the NK_PROPERTY_FLOAT branch calls nk_strtof:

Nuklear/src/nuklear_property.c

Lines 278 to 282 in fcd64f8

	case NK_PROPERTY_FLOAT:
	nk_string_float_limit(buffer, NK_MAX_FLOAT_PRECISION);
	variant->value.f = nk_strtof(buffer, 0);
	variant->value.f = NK_CLAMP(variant->min_value.f, variant->value.f, variant->max_value.f);
	break;

nk_strtof implementation is:

Nuklear/src/nuklear_util.c

Lines 224 to 232 in fcd64f8

	NK_API float
	nk_strtof(const char *str, char **endptr)
	{
	float float_value;
	double double_value;
	double_value = NK_STRTOD(str, endptr);
	float_value = (float)double_value;
	return float_value;
	}

Therefore, in the NK_PROPERTY_DOUBLE branch we should definitely call NK_STRTOD. This will ensure consistent behavior between the float and double property handling.

[sleeptightAnsiC]

Note: be careful when fixing this or #256: nk_strtod() is missing a #ifndef NK_STRTOD guard.

[sleeptightAnsiC]

Note: make sure to also deal with this at the same time: #700