Re: does gfortran backport any g95 code?

[Steve Kargl <sgk at troutmask dot apl dot washington dot edu>]

On Mon, Jul 25, 2005 at 10:06:50PM -0400, Jack Howarth wrote:
>    You guys might want to keep some general level of
> communication open with Andy. In asking about whether
> he ever looked at the current gfortran source code,
> Andy had the following comment...

You're beating a dead horse.  For over a year after the
split, Andy purposely obfuscated his code to make it much
harder to follow his development.  He gratutiously moved
functions within a file, renamed variables, added/deleted
whitespace, changed tabs to spaces, made formatting changes
to comments (ie splitting and combining lines without need),
and moving functions between files without need.

>    I don't watch their source much, so I can't tell how much they borrow
> from me.  Right after the split, I totally rewrote array handling and
> scalarization, which I thought were deeply flawed.  Many of the data
> structures and even conventions have changed in the last two years, so I
> would surmise that it is getting more difficult to merge things all the
> time.

Yep.  The gfortran guys originally tried to work with Andy.
The split was caused by Andy and he went out of his way
to make sure it was hard "to borrow" from him.

> Which brings up the question, do the gfortran developers consider their
> array handling and scalarization deeply flawed and have you attempted
> your own rewrite?

It could probably use a review and some work, but there are only
so many gfortran code contributors with free time to hack.

-- 
Steve