| Message ID | 1334499233-6344-21-git-send-email-sw@weilnetz.de |
|---|---|
| State | Superseded |
| Headers | show |
On Sun, Apr 15, 2012 at 14:13, Stefan Weil <sw@weilnetz.de> wrote: > The MinGW-w64 compiler complains about a non-void function > which does not return a value. > > This is not true, but we can help the compiler to > see that by rewriting the code. > > Signed-off-by: Stefan Weil <sw@weilnetz.de> > --- > block/raw-win32.c | 6 ++---- > 1 files changed, 2 insertions(+), 4 deletions(-) > > diff --git a/block/raw-win32.c b/block/raw-win32.c > index e4b0b75..a420f73 100644 > --- a/block/raw-win32.c > +++ b/block/raw-win32.c > @@ -335,12 +335,10 @@ static int find_device_type(BlockDriverState *bs, const char *filename) > return FTYPE_HARDDISK; > case DRIVE_CDROM: > return FTYPE_CD; > - default: > - return FTYPE_FILE; If these DRIVE_ constants are enums (or later changed to such), a compiler may complain about missing default case. I'd prefer then to keep 'default:', but add 'break;' in place of the return. Would this work in your case too? > } > - } else { > - return FTYPE_FILE; > } > + > + return FTYPE_FILE; This and else change is fine. > } > > static int hdev_probe_device(const char *filename) > -- > 1.7.0.4 >
Am 15.04.2012 19:13, schrieb Blue Swirl: > On Sun, Apr 15, 2012 at 14:13, Stefan Weil<sw@weilnetz.de> wrote: >> The MinGW-w64 compiler complains about a non-void function >> which does not return a value. >> >> This is not true, but we can help the compiler to >> see that by rewriting the code. >> >> Signed-off-by: Stefan Weil<sw@weilnetz.de> >> --- >> block/raw-win32.c | 6 ++---- >> 1 files changed, 2 insertions(+), 4 deletions(-) >> >> diff --git a/block/raw-win32.c b/block/raw-win32.c >> index e4b0b75..a420f73 100644 >> --- a/block/raw-win32.c >> +++ b/block/raw-win32.c >> @@ -335,12 +335,10 @@ static int find_device_type(BlockDriverState *bs, const char *filename) >> return FTYPE_HARDDISK; >> case DRIVE_CDROM: >> return FTYPE_CD; >> - default: >> - return FTYPE_FILE; > If these DRIVE_ constants are enums (or later changed to such), a > compiler may complain about missing default case. I'd prefer then to > keep 'default:', but add 'break;' in place of the return. Would this > work in your case too? > >> } >> - } else { >> - return FTYPE_FILE; >> } >> + >> + return FTYPE_FILE; > This and else change is fine. > This patch is not strictly needed for w64 because it only fixes a compiler warning, so I'll remove it from this patch series. Introducing an enum (also for block/raw-posix.c) and re-writing the switch statement can be done by separate patches. Regards, Stefan W.
diff --git a/block/raw-win32.c b/block/raw-win32.c index e4b0b75..a420f73 100644 --- a/block/raw-win32.c +++ b/block/raw-win32.c @@ -335,12 +335,10 @@ static int find_device_type(BlockDriverState *bs, const char *filename) return FTYPE_HARDDISK; case DRIVE_CDROM: return FTYPE_CD; - default: - return FTYPE_FILE; } - } else { - return FTYPE_FILE; } + + return FTYPE_FILE; } static int hdev_probe_device(const char *filename)
The MinGW-w64 compiler complains about a non-void function which does not return a value. This is not true, but we can help the compiler to see that by rewriting the code. Signed-off-by: Stefan Weil <sw@weilnetz.de> --- block/raw-win32.c | 6 ++---- 1 files changed, 2 insertions(+), 4 deletions(-)