Date: Mon, 14 Jun 2010 20:56:00 +1000 (EST) From: Bruce Evans <brde@optusnet.com.au> To: =?utf-8?Q?Dag-Erling_Sm=C3=B8rgrav?= <des@des.no> Cc: arch@FreeBSD.org Subject: Re: Change to config(8) for OFED Message-ID: <20100614201541.B37598@delplex.bde.org> In-Reply-To: <86zkz05bra.fsf@ds4.des.no> References: <alpine.BSF.2.00.1006111611380.1435@desktop> <20100612.101458.10150326125744273.imp@bsdimp.com> <20100612.111805.94843338670897167.imp@bsdimp.com> <86zkz05bra.fsf@ds4.des.no>
next in thread | previous in thread | raw e-mail | index | archive | help
This message is in MIME format. The first part should be readable text,
while the remaining parts are likely unreadable without MIME-aware tools.
--0-542411416-1276512960=:37598
Content-Type: TEXT/PLAIN; charset=X-UNKNOWN; format=flowed
Content-Transfer-Encoding: QUOTED-PRINTABLE
On Sat, 12 Jun 2010, [utf-8] Dag-Erling Sm=C3=B8rgrav wrote:
> "M. Warner Losh" <imp@bsdimp.com> writes:
>> but NORMAL_C is
>>
>> NORMAL_C=3D ${CC} -c ${CFLAGS} ${WERROR} ${PROF} ${.IMPSRC}
>>
>> and .IMPSRC is null. That seems like a bug to me, but I'm not sure if
>> the bug is that .IMPSRC is computed wrong, or if it really should be
>> ${.ALLSRC}
>
> .IMPSRC is the *implicit* source and is only valid in *implicit* rules.
However, NORMAL_C and related macros have always depended on it being
defined to something reasonable so that they can abuse it in explicit
rules.
> Consider the following:
>
> .SUFFIXES: .foo .bar
> .foo.bar: foo2bar.sh
> /bin/sh foo2bar.sh ${.IMPSRC} ${.TARGET}
>
> When building hello.bar from hello.foo, .IMPSRC is "hello.foo" but .ALLSR=
C
> is "hello.foo foo2bar.sh".
>
> GNU make has a way of defining more complex implicit rules (pattern
> rules). If we had something similar, we could do:
>
> foo_%.o: %.c
> ${NORMAL_C}
BSD make can handle simple cases like this using pattern filters and
substitution. E.g.:
%%%
ABNORMAL_C=3D ${CC} -c ${CFLAGS} ${WERROR} ${PROF} ${.ALLSRC:M*.c:S/^foo_//=
}
# Untested more careful patterns: ${.ALLSRC:M^foo*.c$:S/^foo_//}
bar.o: foo_bar.c stuff.nonc
=09${ABNORMAL_C}
%%%
This has no alternative of using .IMPSRC instead of .ALLSRC since its
=2EIMPSRC is empty with the above dependencies. The rule for getting a
nonempty .IMPSRC in an explicit dependency with BSD makef seems to be:
some of the sources must match the target according to a suffix
rule; then the first matching one (according to the order of the
sources in the dependency list for the target) becomes .IMPSRC,
so .IMPSRC always contains either 0 or 1 name.
I've don't remember having seen the exact rule documented.
Bruce
--0-542411416-1276512960=:37598--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100614201541.B37598>