rra at stanford.edu
Wed Apr 15 20:35:12 UTC 2009
Julien ÉLIE <julien at trigofacile.com> writes:
> When lib/perl.c is modified, perl.o is regenerated but libinn.a is *not*,
> which leads to problems afterwards.
It shouldn't; lib/perl.o isn't included in libinn.a. Although hm, I bet
there aren't explicit dependencies and rebuild rules for it in the
Makefiles that use it. We probably need something like the rebuild rule
for libinn.a. (Or switch the entire build system to non-recursive
> I see that perl.c is not in $(SOURCES) but instead, perl.c/o is treated
> differently and added in a few Makefile rules (like "depend").
> Is there a problem not to put it in $(SOURCES)?
It's not included in libinn.a because libinn.a would then have a
dependency on the Perl libraries, which are only needed for innd and nnrpd
and not all the other stuff in INN that links with libinn.
> I also see that we have:
> # Sources for additional functions only built to replace missing system ones.
> EXTRA_SOURCES = alloca.c asprintf.c fseeko.c ftello.c getaddrinfo.c getnameinfo.c \
> But I do not see when they are compiled. If for instance getaddrinfo is
> missing, I do not understand what will trigger the compilation of
Autoconf via AC_REPLACE_FUNC / LIBOBJS.
Russ Allbery (rra at stanford.edu) <http://www.eyrie.org/~eagle/>
Please send questions to the list rather than mailing me directly.
<http://www.eyrie.org/~eagle/faqs/questions.html> explains why.
More information about the inn-workers