Peter Maydell <peter.mayd...@linaro.org> writes: > On 24 June 2016 at 09:17, Paolo Bonzini <pbonz...@redhat.com> wrote: >> >> >> On 24/06/2016 10:15, Peter Maydell wrote: >>>>>> >>> @@ -31,6 +31,7 @@ typedef struct FWCfgState FWCfgState; >>>>>> >>> typedef struct HCIInfo HCIInfo; >>>>>> >>> typedef struct I2CBus I2CBus; >>>>>> >>> typedef struct I2SCodec I2SCodec; >>>>>> >>> +typedef struct IRQState *qemu_irq; >>>>>> >>> typedef struct ISABus ISABus; >>>>>> >>> typedef struct ISADevice ISADevice; >>>>>> >>> typedef struct IsaDma IsaDma; >>>>> >> >>>>> >> Everything else in typedefs.h is a "typedef struct Thing Thing", >>>>> >> but qemu_irq is different... >>>> > >>>> > We want to keep our readers on their toes! >>> It would mean you now have to decide whether the file is orderd >>> by the types being defined or by the underlying implementation >>> type (previously both orders were the same)... >> >> Indeed, and renaming the struct is trivial because it's used in a >> handful of places only. > > It would still be different by being a pointer-to-Foo, not a Foo.
Hiding pointer-ness behind a typedef is a bad idea more often than not. What do you want me to do, if anything?