Hello All,
See http://gcc.gnu.org/wiki/MiddleEndLispTranslator for MELT
The MELT branch bootstrapped, in the sense that the Lisp compiler is
able to compile itself to C code. It is not the bootstrap in the usual
GCC sense (a GCC being able to compile itself - currently MELT GCC
behaves like the trunk in this respect). MELT branch is closely
following the trunk: I am doing svnmerge merge more than once a week
(without any issues so far)
I've got some questions to the list:
first, MELT is generating itself, so the generated C code
gcc/warm-basilys-0.c file is committed to SVN. In that respect, it is
like the configure script (a generated file which is in the SVN repository).
second, the generated warm-basilys-0.c is quite big (more than 250KLOC
or 13Mbytes), and a small change (e.g. one line) in
melt/warm-basilys.bysl triggers many changes (e.g. a thousand lines
changed) in the generated file warm-basilys-{0,1,2,3].c
I still feel that I have to commit it frequently, to make tracable and
reproducible all my changes.
Some questions:
1. Should I avoid committing warm-basilys-0.c frequently to lower the
Subversion server disk consumption? I could do that, but then my changes
would be less reproductible (in the sense that applying the diff between
two commits to a source tree would not be enough to make it
recompilable). My perception is that disk space is on svn://gcc.gnu.org/
cheap (but then, I am not paying it!).
2. I cannot figure out if tbhe GCC bugzilla can be used for work on a
branch, not only on the trunk (or older releases).
3. Are shared libraries *.la obtained by libtool usable inside
gcc/Makefile.in?
Regards.
--
Basile STARYNKEVITCH http://starynkevitch.net/Basile/
email: basile<at>starynkevitch<dot>net mobile: +33 6 8501 2359
8, rue de la Faiencerie, 92340 Bourg La Reine, France
*** opinions {are only mines, sont seulement les miennes} ***
