On 28.01.2016 11:04, Andreas Schwab wrote: > John Paul Adrian Glaubitz <glaub...@physik.fu-berlin.de> writes: > >> [suggestion to use "void" as dummy return type] >> >> Wait. Do you think this would actually allow ghc to determine the >> return type later? If I remember correctly, ghc currently initially >> declares the function prototype with return type void*, doesn't it? > Replace a lie with a different lie. Spot the pattern? > > Andreas. I am sorry, I fail to spot any pattern. As I understood you, you are opposed to changing gcc because a program that lies to gcc fails to get the result this program expects. But I don't see any pattern in "replacing a lie with a different lie". Please be more specific in what your message should tell the recipients.
In case you refer to the ppc64el issue of ghc, and try to imply that changing lies at it fits has precedence in ghc history (and thus ghc either needs to stop lying or find a new lie that just happens to work), I strongly disagree. In that case, ghc developers not "replace a lie with a different lie", but replaced a lie "(empty parameter list)" with the truth "(unknown parameter list)". Regards, Michael Karcher
