Yes I know there are traces of GCC, but it never really materialized. The guy (Tom Bates) who did all these ports that you can seen on http://ituglib.connect-community.org left the company end of 2007 (made redundant, actually), so all the stuff you can see there from 2007 or older is his stuff (I took care to make it available to the public again and re-packaged it). I now sort of took over, so all the stuff that is newer, is mine, but of course based on his work. He seems to have had a gcc port, but never published it, maybe because it never worked properly, or because he got forbidden, as we do have our own compiler. I asked him several times for gcc, never got any reply (but got replies to from him on other things), so I assume there was something going on behind the scene. If you look into that floss-0.9* package, at the ToDo file, you'll see that "remove GNU-C stuff or get gcc ported" is on the agenda. I'm pretty sure it'd be the former rather than the later... On the other hand: a compiler is obliged to ignore unknown pragmas, the C99 Standard (n1256.pdf), 6.10.6.1 says: "Any such pragma that is not recognized by the implementation is ignored"). And the (god old) times where gcc started nethack when finding a #pragma have long passed ;-)
Bye, Jojo -----Original Message----- From: Bruno Haible [mailto:[email protected]] Sent: Sunday, October 03, 2010 2:56 AM To: Schmitz, Joachim Cc: Paul Eggert; Paolo Bonzini; 'bug-gnulib' Subject: Re: Fwd: sed porting trouble Btw, Joachim: I can see in the libfloss sources that there is some GCC port for Tandem NSR. Does it support #pragma columns? If not, then the idiom needs to look like this: #if defined __TANDEM && !defined __GNUC__ @PRAGMA_COLUMNS@ #endif Bruno
