Date: Thu, 23 Jun 2011 20:09:40 +0200 From: Koop Mast <kwm@rainbow-runner.nl> To: Rainer Hurling <rhurlin@gwdg.de> Cc: ports@freebsd.org Subject: Re: math/ggobi: compile error with clang Message-ID: <4E038164.2070907@rainbow-runner.nl> In-Reply-To: <4E037A4A.4000904@gwdg.de> References: <4E036F01.2020708@gwdg.de> <4E03725E.7070802@rainbow-runner.nl> <4E037A4A.4000904@gwdg.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On 6/23/2011 19:39 , Rainer Hurling wrote: > On 23.06.2011 19:05 (UTC+1), Koop Mast wrote: >> On 6/23/2011 18:51 , Rainer Hurling wrote: >>> I just discovered, that math/ggobi does not compile with clang, see: >>> >>> http://pointyhat.freebsd.org/errorlogs/amd64-errorlogs/e.9-exp.20110616185105/ggobi-2.1.9.log >>> >>> >>> >>> It seems, the main error is in 'line control' statement of >>> src/ggobi-data.c, l.45 >>> >>> #line 0 "data.gob" >>> static void ggobi_data_class_init (GGobiDataClass * c) G_GNUC_UNUSED; >>> >>> The official cpp online docs on >>> http://gcc.gnu.org/onlinedocs/cpp/Line-Control.html defines >>> >>> #line linenum >>> linenum is a non-negative decimal integer constant. It specifies the >>> line number which should be reported for the following line of input. >>> Subsequent lines are counted from linenum. >>> >>> It seems gcc accepts the 0 (zero) as valid, whilst clang produces an >>> error '#line directive requires a positive integer argument'. >>> >>> What is the meaning of line 0 in this context? At line 1 (the first >>> line in src/data.gob) there is the initiation '%h{' for the whole file >>> contents. Does 0 mean, take the whole file? >>> >>> Would it be correct to change value 0 into 1 (#line 1)? At least, gcc >>> is able to build that and it seems to work fine. >>> >>> My system is compiled with gcc, so it would be nice, if someone would >>> give it a try with clang. >>> >>> Many thanks in advance, >>> Rainer Hurling >> >> I had the same problem with devel/gob2 ages ago. This following is the >> solution I settled on. I don't know if this is correct or not but it >> seems to work fine. >> http://www.freebsd.org/cgi/cvsweb.cgi/~checkout~/ports/devel/gob2/files/patch-src_out.c?rev=1.1;content-type=text%2Fplain >> > > Koop, many thanks for the fast reaction. This is exactly what I did > (=changing 0 in 1), gcc accepted that and until now it seems to work. > > Does anybody knows if changing 0 into 1 is ok in this context? > > I am not able to test, if this works with clang. If someone is willing > to try, there is a diff attached. > > Rainer > >> -Koop No port revision bump is needed here. -Koop
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4E038164.2070907>