On Apr 17, 2001, "Robert Collins" <[EMAIL PROTECTED]> wrote:
> I think so. I'm using gcc 2.95.3 - that does dependencies as a side
> effect doesn't it ?
Yep. Check .deps/cf_gen.Po
>> Nope. I'm saying it would make sure BUILT_SOURCES are built before
>> any other source files are compiled.
>
> -Original Message-
> From: Alexandre Oliva [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, April 17, 2001 3:16 PM
> To: Robert Collins
> Cc: Lars J. Aas; [EMAIL PROTECTED]
> Subject: Re: target dependency bug
>
>
> On Apr 17, 2001, "Robert Collins"
> <[EMAIL PROTECTED]> wrote:
>
> > cf_g
On Apr 17, 2001, "Robert Collins" <[EMAIL PROTECTED]> wrote:
> cf_gen_defines.h is being built. But cf_gen.o isn't being updated when
> cf_gen_defines.h is updated.
Did you --enable-dependency-tracking?
> And AFAICT, automake has enough info to make cf_gen_defines.h a
> dependency of cf_gen.o
> -Original Message-
> From: Alexandre Oliva [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, April 17, 2001 2:20 PM
> To: Robert Collins
> Cc: Lars J. Aas; [EMAIL PROTECTED]
> Subject: Re: target dependency bug
>
>
> On Apr 17, 2001, "Robert Collins"
> <[EMAIL PROTECTED]> wrote:
>
> > If I
On Apr 17, 2001, "Robert Collins" <[EMAIL PROTECTED]> wrote:
> If I understand you correctly, that should result in a Makefile.in with
> cf_gen_defines.h as a dependancy of cf_gen.$(OBJEXT): ?
Nope. It's some other target that would depend on cf_gen_defines.h.
BUILT_SOURCES are global, not spec
This was the original makefile.am fragment:
cf_gen_SOURCES = cf_gen.c defines.h
nodist_cf_gen_SOURCES = cf_gen_defines.h
BUILT_SOURCES = cf_gen_defines.h
If I understand you correctly, that should result in a Makefile.in with
cf_gen_defines.h as a dependancy of cf_gen.$(OBJEXT): ?
I'm testin
On Apr 17, 2001, "Robert Collins" <[EMAIL PROTECTED]> wrote:
> Thats the bug I reported: a built_source that doesn't get encoded as a
> dependency by automake.
It's definitely not encoded as an explicit dependency of files that
depend on it (how would automake figure it out?), but BUILT_SOURCES
> -Original Message-
> From: Alexandre Oliva [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, April 17, 2001 1:53 PM
> To: Robert Collins
> Cc: Lars J. Aas; [EMAIL PROTECTED]
> Subject: Re: target dependency bug
>
>
> On Apr 16, 2001, "Robert Collins"
> <[EMAIL PROTECTED]> wrote:
>
> > I'm
On Apr 16, 2001, "Robert Collins" <[EMAIL PROTECTED]> wrote:
> I'm happy with --enable-dependency-tracking... but will all the end
> users?
If they want dependence tracking, yes. Note that it will be enabled
by default if it can be obtained as a side effect of compilation.
Also note that, if th
- Original Message -
From: "Lars J. Aas" <[EMAIL PROTECTED]>
To: "Robert Collins" <[EMAIL PROTECTED]>
Cc: <[EMAIL PROTECTED]>
Sent: Tuesday, April 17, 2001 2:40 AM
Subject: Re: target dependency bug
> On Tue, Apr 17, 2001 at 12:16:01AM +1000, Robert Collins wrote:
> : given the following
On Tue, Apr 17, 2001 at 12:16:01AM +1000, Robert Collins wrote:
: given the following
:
: cf_gen_SOURCES = cf_gen.c defines.h
: nodist_cf_gen_SOURCES = cf_gen_defines.h
: BUILT_SOURCES = cf_gen_defines.h
:
: I think cf_gen_defines.h should be one of the dependencies of
: cf_gen.$(OBJEXT). It's n
given the following
cf_gen_SOURCES = cf_gen.c defines.h
nodist_cf_gen_SOURCES = cf_gen_defines.h
BUILT_SOURCES = cf_gen_defines.h
I think cf_gen_defines.h should be one of the dependencies of
cf_gen.$(OBJEXT). It's not though in my generated Makefile here.
Suggestions?
Rob
12 matches
Mail list logo