Re: [xml] Release of libxml2-2.6.16

[Daniel Veillard <veillard redhat com>]

On Fri, Nov 12, 2004 at 08:15:34AM +0100, Gerrit P. Haase wrote:
> Daniel Veillard wrote:
> > I disagree. The fact that you can't regenerate the C file should not 
> > prevent
> > compiling it because it should be present in $(srcdir) anyway.
> >  paphio:~/XML -> tar tvzf libxml2-2.6.16.tar.gz | grep testapi.c
> >  -rw-rw-r-- 50138/69    1317100 2004-11-09 17:12:04 
> >  libxml2-2.6.16/testapi.c
> >  paphio:~/XML ->
> > Now you need to understand why your make didn't found it in $(srcdir)
> > that's where the problem is !
>
> Yes, I know I have the file here too, I don't know why gentest.py is 
> called.  It is probably a libtool / automake feature;)

  There is two things:
    1/ make may come to the rule that regeneration is needed, for example
       if it find the timestamp of gentest.py older than testapi.c one,
       that's normal, but will fail in a variety of situation.
    2/ whether regeneration was possible or not it should build using
       $(srcdir)/testapi.c

 Maybe the 

  testapi.c: gentest.py doc/libxml2-api.xml
        ...

lines in Makefile.am should really be rewritten

  $(srcdir)/testapi.c: $(srcdir)/gentest.py $(srcdir)/doc/libxml2-api.xml
        -@(if [ "$(PYTHON)" != "" ] ; then cd $(srcdir) $(PYTHON) gentest.py ; fi )

this may be worth trying on your side.

Daniel

-- 
Daniel Veillard      | Red Hat Desktop team http://redhat.com/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/