re-align COBOL_TYPE_UNSIGNED_BINARY code to use switch

[GitMensch]

I'd have sent in a PR, but I'm not sure about the not-yet-checked-in changes with the dynamic memory changing the related code, so here's the suggestion to change those lines:

gixsql/runtime/libgixsql/SqlVar.cpp

Lines 281 to 316 in debeeea

	case COBOL_TYPE_UNSIGNED_BINARY:
	if (this->length == 1) { // 1 byte
	uint8_t n8 = *((uint8_t*)addr);
	snprintf((char*)realdata, length, "%d", n8);
	}
	else {
	if (this->length == 2) { // 1 byte
	uint8_t n8 = *((uint8_t*)addr);
	snprintf((char*)realdata, length, "%d", n8);
	}
	else {
	if (this->length == 3 || this->length == 4) { // 2 bytes
	uint16_t n16 = *((uint16_t*)addr);
	n16 = COB_BSWAP_16(n16);
	snprintf((char*)realdata, length, "%d", n16);
	}
	else {
	if (this->length >= 5 || this->length <= 9) { // 4 bytes
	uint32_t n32 = *((uint32_t*)addr);
	n32 = COB_BSWAP_32(n32);
	snprintf((char*)realdata, length, "%d", n32);
	}
	else {
	if (this->length >= 10 || this->length <= 18) { // 8 bytes
	uint64_t n64 = *((uint64_t*)addr);
	n64 = COB_BSWAP_64(n64);
	snprintf((char*)realdata, length, "%d", n64);
	}
	else {
	// Should never happen - TODO: log fatal and abort
	}
	}
	}
	}
	}

(and similar) to use switch (this->length) instead of nested if/else.

I may send in a PR for that when wanted, too.

[mridoni]

This is undergoing a massive overhaul due to #124, so a PR would be probably wasted, but I am scheduling this for v1.0.20