@ does not quite work as intended when in UTF8 mode

[lightbinder]

@ behaves idiosyncratically in UTF8 mode as follows:

If you have a line like

I uwu youwu.@ And I uwu youwu toowu.\

... you will indeed trigger the clickwait, but the rest of the line will be completely cut off.

On the other hand if you have a line like

I uwu youwu.`@`And I uwu youwu toowu.\

... the second half of the line will indeed display, but one line down.

On the final (gripping) hand, this

I uwu youwu.@/
I uwu youwu toowu.\

... works exactly as intended. Praise be to the onsen and onani gods.

[lightbinder]

It turns out that this behavior exists in all of legacy mode.

14 April 2023 Update
Now investigating if the "ignore the rest of the line" behavior is intentional and in line with mainline NScripter. If so, I'm still going to have to make modifications, as the program currently allows you to put a ` after the @ and still have it work, just on a new line. This may well be not orthogonal to the way the program is supposed to work.

[lightbinder]

For additional detail and a workaround see here.

[lightbinder]

In ONScripter_text.cpp:

        if (textgosub_label){
            saveon_flag = false;

            textgosub_clickstr_state = CLICK_WAIT;
            if (script_h.getStringBuffer()[string_buffer_offset] == 0x0)
                textgosub_clickstr_state |= CLICK_EOL;
            gosubReal(textgosub_label, script_h.getWait(), true);

            event_mode = IDLE_EVENT_MODE;
            waitEvent(0);

            return false;
        }

This block triggers when we see an @. The issue is gosubReal(textgosub_label, script_h.getWait(), true); -- or more specifically the script_h.getWait() part of the equation. It is a small inline function in ScriptHandler.h:

inline char *getWait(){ return wait_script?wait_script:next_script; };

The entire script object is iterated on until we hit the next line to execute. This is why all text after @ is ignored, unless there is a special command token such as ` or / next.

[lightbinder]

While this is a annoying problem, because of the existence of the functional workaround this bugfix is deprioritized at present.