Re: MELT tutorial on the wiki

[Tom Tromey <tromey@redhat.com>]

>>>>> "Basile" == Basile STARYNKEVITCH <basile@starynkevitch.net> writes:

Basile> My current belief is that MELT is easier built (and used) as a
Basile> GCC-trunk (or future GCC-4.5) plugin melt.so.

Yeah, I looked at this, but I really wanted something where I did not
have to do much work... no applying patches or hacking scripts.  I have
very little time for this side project, so most barriers look too big at
this point.

Tom> In both cases I just configured with --disable-bootstrap --enable-melt,
Tom> plus the appropriate --enable-langauges option.  This is on x86 F11.

Basile> You probably missed some configure arguments. The MELT's
Basile> gcc/configure.ac ends with
Basile> ## Basile adds a notice if the MELT branch is configured without
Basile> ## --enable-melt
Basile> if test "$enabled_melt" != "yes" ; then
Basile>    AC_MSG_NOTICE(
Basile> [GCC MELT branch is configured WITHOUT enabling melt.
Basile>  Are you sure to want that?])
Basile> fi

I looked into this a little.  It looks like the PPL checks don't work
properly in the case where PPL is a system library.  I guess I need
--with-ppl=/usr ... I will try that later.

BTW: IMO, on a branch like this, it would be friendlier for users to:

* Check in the correct version of configure
* Enable MELT by default, don't require an option
* Die in configure if a prerequisite is missing
* Have the PPL checks detect the situation where PPL is a system library

Otherwise whatever error messages there might be are just lost in the
noise.

Tom