https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91916
Bug ID: 91916
Summary: Maybe a dead code in socket.d
Product: gcc
Version: 10.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: d
Assignee: ibuclaw at gdcproject dot org
Reporter: marxin at gcc dot gnu.org
CC: ibuclaw at gcc dot gnu.org
Target Milestone: ---
Looking at the ./libphobos/src/std/socket.d file, when building:
/dev/shm/mliska/gcc_89622158a277a4f1200d2652d66f8c1b9c1e4012/objdir/./gcc/gdc
-B/dev/shm/mliska/gcc_89622158a277a4f1200d2652d66f8c1b9c1e4012/objdir/./gcc/
-B/usr/local/x86_64-pc-linux-gnu/bin/ -B/usr/local/x86_64-pc-linux-gnu/lib/
-isystem /usr/local/x86_64-pc-linux-gnu/include -isystem
/usr/local/x86_64-pc-linux-gnu/sys-include -fPIC -O2 -g -nostdinc -I
../../../../libphobos/src -I ../../../../libphobos/libdruntime -I
../libdruntime -I . -c ../../../../libphobos/src/std/socket.d -fversion=Shared
-o std/.libs/socket.o
I see there's a function formatValue which contains if-elseif chain that has a
dead code:
Statement can throw: formatValue (w_9(D), D.19387, f_5(D));
formatValue (struct Appender!(immutable(string)) & w, SocketOption val, const
struct FormatSpec!char & f)
{
...
<bb 3> :
if (val_6(D) == 1)
goto <bb 4>; [INV]
else
goto <bb 5>; [INV]
<bb 4> :
D.19387.length = 5;
D.19387.ptr = "DEBUG";
formatValue (w_9(D), D.19387, f_5(D));
// predicted unlikely by early return (on trees) predictor.
goto <bb 53>; [INV]
<bb 5> :
if (val_6(D) == 6)
goto <bb 6>; [INV]
else
goto <bb 7>; [INV]
...
<bb 35> :
if (val_6(D) == 1)
goto <bb 36>; [INV]
else
goto <bb 37>; [INV]
<bb 36> :
D.19435.length = 11;
D.19435.ptr = "TCP_NODELAY";
formatValue (w_9(D), D.19435, f_5(D));
// predicted unlikely by early return (on trees) predictor.
goto <bb 53>; [INV]
So as seen:
/// Specifies a socket option:
enum SocketOption: int
{
DEBUG = SO_DEBUG, /// Record debugging
information
...
TCP_NODELAY = .TCP_NODELAY, /// Disable the Nagle algorithm
for send coalescing
these two have same integer representation == 1. Maybe it's a bug?
