Re: cvs commit: httpd-2.0/server Makefile.in

["William A. Rowe, Jr." <wr...@rowe-clan.net>]

Can I ask a stupid question about the fragment below?

I was under the impression that libapr.so and libaprutil.so were exported
from themselves, not from libhttpd.so.  Is that assumption correct?

If so, why regen exports.c of libhttpd.so upon changes to libapr/aprutil
headers?

Bill

At 09:41 AM 3/11/2003, trawick@apache.org wrote:
>  +EXPORT_DIRS = $(top_srcdir)/include $(top_srcdir)/os/$(OS_DIR) $(APR_INCLUDEDIR) $(APU_INCLUDEDIR) $(top_srcdir)/modules/http
>   
>  +# If export_files is a dependency here, but we remove it during this stage,
>  +# when exports.c is generated, make will not detect that export_files is no
>  +# longer here and deadlock.  So, export_files can't be a dependency of
>  +# delete-exports.
>   delete-exports:
>        @if test -f exports.c; then \
>  -                 headers="`find $(EXPORT_FILES) -newer exports.c`" ; \
>  -                 if test -n "$$headers"; then \
>  -                     echo Found newer headers. Will rebuild exports.c. ; \
>  -                     echo rm -f exports.c ; \
>  -                     rm -f exports.c ; \
>  -                 fi \
>  +         if test -f export_files; then \
>  +             files=`cat export_files`; \
>  +             headers="`find $$files -newer exports.c`"; \
>  +             if test -n "$$headers"; then \
>  +                echo Found newer headers. Will rebuild exports.c.; \
>  +                echo rm -f exports.c export_files; \
>  +                rm -f exports.c export_files; \
>  +             fi; \
>  +         else \
>  +             rm -f exports.c; \
>  +         fi; \
>        fi