On Friday, July 13, 2012 12:58:21 PM UTC+2, Dima Pasechnik wrote: > > But WINSOCKAPI should not be defined on Cygwin, in the first place. > I guess it's a problem with ECL which defines it unconditionally, rather than only on (genuine) Windows .
> In my sys/types the fd_set relevant ifdef does not care about > _WINSOCKAPI_. Indeed, it is > # if !(defined (_POSIX_SOURCE) || defined (_WINSOCK_H) || defined > (__USE_W32_SOCKETS)) > Oh, I see, I have: # if !(defined (_POSIX_SOURCE) || defined (_WINSOCK_H) || defined (_WINSOCKAPI_) || defined (__USE_W32_SOCKETS)) I had freshly installed Cygwin 1.7.15 last week , but also installed a CVS version at some point because their gcc is too broken. So that may be the source of the discrepancey in types.h It could also be that I'm using the gcc4 package version 4.5.3 rather than 4.3.4 (if by any chance the types.h file belongs to that package). > > No wonder it has no effect. Do you have anything different from this? > If yes, when did you last update your Cygwin? I run 1.7.15. > Sometimes people install headers from MINGW (say), and this wreaks > havoc.. > No mingw related stuff installed here. -- -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
