InputDouble disallows inf and nan input

[PeterJohnson]

Version/Branch of Dear ImGui:

Version: 64b1e44
Branch: docking

My Issue/Question:

InputDouble() does not allow inf or nan be input, despite these being recognized by sscanf() (see https://en.cppreference.com/w/cpp/string/byte/strtof) and output by printf(). This means an infinity/NaN value can be displayed, but not set. This is caused by ImGuiInputTextFlags_CharsScientific not allowing these characters: https://github.com/ocornut/imgui/blob/d72e1563d4274e0f1a21973cf2d98563d579d377/imgui_widgets.cpp#L3985C9-L3987C27

A workaround is to use InputScalar() and not setting ImGuiInputTextFlags_CharsScientific.

Standalone, minimal, complete and verifiable example:

ImGui::Begin("Example Bug");
double v = std::numeric_limits<double>::infinity();
InputDouble("test", &v);
ImGui::End();

[ocornut]

Lately I have been wondering if the char filtering is a good idea at all. I wonder if we should just tend to disable filtering and perform the scan>store>display sequence naturally when Enter is pressed.

[ocornut]

I removed the filter and "nan"/"inf" etc. parse fine now, so this is solved :)