Re: Will tree-ssa be GCC 3.5?

[Robert Dewar <dewar at gnat dot com>]

Scott Robert Ladd wrote:

> Actually, intent was all I was interested in. If there is no intent to 
> make something work, no amount of money is going to motivate quality work.

That's also an odd point of view. What we as a company work on has much 
more to do with our customers interests, than with our abstract 
intentions. If you have a customer who is interested in seeing feature Y 
implemented, and feature Y is reasonable, and worthwhile, and the 
customer has sufficient funds, then the project can be started up, 
successfully, quite independently of whether anyone had any intent prior 
to the funded project (for example, in the GNAT arena, the 
Initialize_Scalars feature was one that was devised in direct response 
to needs of a large customer, and we worked with that customer to invent 
the necessary feature and figure out how to implement it. We hadn't even 
thought of this idea (along with the related -gnatV switch) until we had 
these discussions.

In the current situation, if users are willing to fund significant work 
on the Fortran 95 compiler, and if this compiler requires additional 
work to be done to get tree-ssa to work well, then extra resources (for 
example more people) will be put into this tree-ssa development. In fact 
it is absolutely obvious that those working on tree-ssa right now intend 
it to be successful, effective, and to be part of the mainline. Extra 
resources can make these intentions/wishes come to fruition.

> But the question is now largely moot, and I have learned what I needed 
> to know. Thank you for your opinions.

Remember that moot means arguable, and undecided :-)