Missing break in switch-case?

[Schildkroet]

Hello,
is it intended, that there is no break for each case here?

nanoprintf/nanoprintf.h

Lines 402 to 407 in 64333b3

	case 'i':
	case 'd': tmp_conv = NPF_FMT_SPEC_CONV_SIGNED_INT;
	case 'o': if (tmp_conv == -1) { tmp_conv = NPF_FMT_SPEC_CONV_OCTAL; }
	case 'u': if (tmp_conv == -1) { tmp_conv = NPF_FMT_SPEC_CONV_UNSIGNED_INT; }
	case 'X': if (tmp_conv == -1) { out_spec->case_adjust = 0; }
	case 'x': if (tmp_conv == -1) { tmp_conv = NPF_FMT_SPEC_CONV_HEX_INT; }

[charlesnicholson]

Yeah, it's intended fall-through; each case is guarded by tmp_conv during the fall-through. It's also awful and ridiculous and stupid, but arm-gcc knows exactly what to do with it.

The underlying issue was that nanoprintf was non-conformant: the C standard requires that printf + friends ignore the 0 flag when precision is specified, so you need to do some special-casing iff an integer type and 0 and precision were specified. That could have been a separate if statement with a bunch of or sub-clauses after the switch case, but it added ~100 bytes to the compiled code on cortex-m0 architectures. So, I did a disgusting Duff's-Device-ish thing here, and it ended up being a freebie in terms of size.

#244
https://github.com/charlesnicholson/nanoprintf/pull/245/files