Date: Fri, 29 Mar 2002 11:44:06 +0100 (CET) From: Alexander Leidinger <Alexander@Leidinger.net> To: winter@jurai.net Cc: jmallett@FreeBSD.org, green@FreeBSD.org, sw@anthologeek.net, freebsd-current@FreeBSD.org Subject: Re: cvs commit: ports/lang Makefile ports/lang/icc Makefile distinfo pkg-comment pkg-descr pkg-plist ports/lang/icc/files patch-include Message-ID: <200203291044.g2TAi6h8000665@Magelan.Leidinger.net> In-Reply-To: <20020328181725.I45452-100000@sasami.jurai.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On 28 Mär, Matthew N. Dodd wrote: >> I've tried to give it a start, so I also allowed __ICC in pcpu.h, now it >> fails with: > > I got most of src/bin src/sbin src/usr.bin src/usr.sbin and lib to compile > with it. Does libm work? At my first try (January) ICC complained about some files. But perhaps BDE's commits to the libm solved these... > Libc had some issues with malloc and mmap() and wouldn't function when > installed, but other things worked fine. > > My patches to src/share/mk/ are here: > > ftp://ftp.jurai.net/users/winter/icc.mk.diff > > This allows you to set 'USE_ICC' and 'ICFLAGS' and build stuff. Personally I would prefer ICCCFLAGS or ICC_CFLAGS. My review of your patch: bsd.lib.mk: - Shouldn't ICC_PICFLAG be encapsulated withhin '.if !defined(ICC_PICFLAC)' like PICFLAG is? bsd.prog.mk: - Does ${LOCALBASE} instead of /usr/local work here? sys.mk: - Have a look at the '#ICC' line. - The single suffix rules are wrong (line 127, 195). This has to get splitted into an icc compile line and a CC link line. Perhaps with a temporary file (mtemp(1)) which gets rm'ed after linking. - You've commented bsd.cpu.mk out. We should add USE_ICC stuff there too. > Setting 'CFLAGS' to nil and 'NO_WARNS' is also a good idea. > > icc.cfg: I reordered the lines a little bit: > -Ulinux > -U__linux__ > -U__linux > -D__FreeBSD__=5 > -D__ELF__=1 hould I add these to the port? Seems to be a good idea to me... > -D__ICC__=1 ICC already defines '__ICC', doe we really need this? > -D__attribute__(x)= > -D__GNUC__=2 Ok as a short term solution, but IMHO this should get solved in the source. > -nolib_inline I have to look this up, but if this means some of the distributable icc libs don't get linked into the programs, then I don't think this is a good idea. This makes the complete userland dependend on the icc port. > -X > -I/usr/include Why? Did I missed a linux secific include file? > I also added a few lines to my > /usr/local/intel/compiler50/ia32/bin/iccvars.csh > setenv ICFLAGS '-O3 -tpp6 -ip' > setenv USE_ICC > setenv CFLAGS > setenv CWARNFLAGS > setenv NO_WARNS yes I think they should reside in /etc/make.conf... Bye, Alexander. -- Give a man a fish and you feed him for a day; teach him to use the Net and he won't bother you for weeks. http://www.Leidinger.net Alexander @ Leidinger.net GPG fingerprint = C518 BC70 E67F 143F BE91 3365 79E2 9C60 B006 3FE7 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200203291044.g2TAi6h8000665>