Re: Will tree-ssa be GCC 3.5?

[kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)]

    > If you want something done, you need to approach specific people who
    > are doing the work and solicit proposals from those people as to
    > funding and schedules.

    Those specific people will want to know that the rest of the GCC
    contributors will support them.  

No, you *cannot* depend on unfunded volunteer work for a funded development
project.  For example, if you're offering commercial support for a GCC-based
front end and find a middle-end bug, you can't tell your customer that you
can only fix it if these volunteers choose to do so for you.  Instead, you
either have to make a *contract* with these volunteers or have in-house
expertese to fix the problems.  That's not to say that you can't *ask* for
help from the GCC development community: you just can't *depend* on getting it.

    In the case of tree-ssa for example, this means that there has to be
    consensus that this project makes sense, can make a real improvement
    and therefore has to be included in release x.y.  

No, that's wrong.  As Robert says, any funded project that depends on tree-ssa
for its success (and, unless I'm misunderstanding something, gfortran is
*not* one of those) must also provide the funding to ensure that tree-ssa
is completed to the extent needed for that project and on the schedule
required for that project.