Обсуждение: Preprocessor condition fix
Here is a one-line patch to fix a wrong preprocessor condition in pg_regress, found because the VS 2015 compiler warns on the cast in the 32-bit branch where apparently earlier versions did not. According to git grep, this is the only place where WIN64 is used without the leading underscore. -- Christian
Christian Ullrich <chris@chrullrich.net> writes: > Here is a one-line patch to fix a wrong preprocessor condition in > pg_regress, found because the VS 2015 compiler warns on the cast in the > 32-bit branch where apparently earlier versions did not. Pushed, thanks. > According to git grep, this is the only place where WIN64 is used > without the leading underscore. Hm, my grep found another one ... regards, tom lane
* Tom Lane wrote: > Christian Ullrich <chris@chrullrich.net> writes: >> According to git grep, this is the only place where WIN64 is used >> without the leading underscore. > > Hm, my grep found another one ... Oh, sorry. I saw that one, but thought it was intentional because _WIN64 is defined automatically anyway. -- Christian
Christian Ullrich <chris@chrullrich.net> writes: > * Tom Lane wrote: >> Hm, my grep found another one ... > Oh, sorry. I saw that one, but thought it was intentional because _WIN64 > is defined automatically anyway. Oh? Then we should not need that one (the /D switch in win32.mak) at all. Should we just remove it? regards, tom lane
* From: Tom Lane [mailto:tgl@sss.pgh.pa.us] > Christian Ullrich <chris@chrullrich.net> writes: > > * Tom Lane wrote: > >> Hm, my grep found another one ... > > > Oh, sorry. I saw that one, but thought it was intentional because _WIN64 > > is defined automatically anyway. > > Oh? Then we should not need that one (the /D switch in win32.mak) at all. > Should we just remove it? We have both confirmed several times that nothing depends on it. I think it can go. -- Christian
Christian Ullrich <chris@chrullrich.net> writes: > * Tom Lane wrote: >> Oh? Then we should not need that one (the /D switch in win32.mak) at all. >> Should we just remove it? > We have both confirmed several times that nothing depends on it. I think it can go. Done. regards, tom lane