fix #40: key length comparison in config.c

[miuirussia]

We already have checking emptiness of string in parsers, so empty value is valid. Fixes #40

[bayes1701]

Probably it's better to skip lines with empty values in awg-quick:

@@ -49,8 +49,10 @@ parse_options() {
        shopt -s nocasematch
        while read -r line || [[ -n $line ]]; do
                stripped="${line%%\#*}"
+               [[ -z "${stripped//[[:space:]]/}" ]] && continue # skip empty lines
                key="${stripped%%=*}"; key="${key##*([[:space:]])}"; key="${key%%*([[:space:]])}"
                value="${stripped#*=}"; value="${value##*([[:space:]])}"; value="${value%%*([[:space:]])}"
+               [[ $stripped == *"="* && -z "$value" ]] && continue # skip lines with empty values
                [[ $key == "["* ]] && interface_section=0
                [[ $key == "[Interface]" ]] && interface_section=1
                if [[ $interface_section -eq 1 ]]; then

The reason is we are setting flags for found keys:

amneziawg-tools/src/config.c

Lines 580 to 584 in 5d6179a

	} else if (key_match("I1")) {
	ret = parse_awg_string(&ctx->device->i1, "I1", value);
	if (ret)
	ctx->device->flags |= WGDEVICE_HAS_I1;
	} else if (key_match("I2")) {

... and we can set a key unintentionally to an empty string (like PrivateKey=)

[miuirussia]

no, not better, as some scripts use directly complied binary, not scripts.

and empty PrivateKey will throw exception Key is not the correct length or format:

[bayes1701]

I think setting ctx->device->flags for empty values may have corner cases

[miuirussia]

If you think, so test it and show me cases with error