[Bug fortran/86248] [7/8/9/10 Regression] LEN_TRIM in specification expression causes link failure

[paul.richard.thomas at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86248

--- Comment #6 from paul.richard.thomas at gmail dot com <paul.richard.thomas at gmail dot com> ---
Hi Bill,

If you look at pr44265, I took over the patch from Ian Sandoe and
fixed one or two of the wrinkles associated with it. I do not seem to
have given it as much thought as I should. Had I done so, I would have
gone down the route suggested by Tobias Burnus and apparently used by
g95 and NAG, namely to mark the length of the return value as unknown
and initialize it on scope entry.

I don't have time tonight but I'll put it on my list of TODOs. It
looks like fairly low hanging fruit to me so I'll take a look see in
the next few days. Messing with name mangling cannot be the right way
to go that is for sure.

Thanks for prodding me

Paul

On Mon, 21 Oct 2019 at 19:39, longb at cray dot com
<gcc-bugzilla@gcc.gnu.org> wrote:
>
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86248
>
> --- Comment #5 from Bill Long <longb at cray dot com> ---
> Any progress on this?
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.