On Tue, 15 Oct 2024 at 19:16, Richard (Rikki) Andrew Cattermole via Digitalmars-d <digitalmars-d@puremagic.com> wrote:
On 15/10/2024 10:07 PM, Manu wrote:
> ...to make this worse; Rainer has effectively checked-out too. We've
> lost another one of our finest.

He implemented the negative annotation exportation switches for dmd at
end of last year.

(dllimport and visibility)

Druntime/phobos are not currently offered as a shared library though.

That was a MAJOR contribution to dmd.

Oh I know it... I mean, I have always said Rainer is one of the greatest, most competent, most high impact, and most quiet and humble contributors we've ever had... but I'm not sure what your point is. He seems to have checked out... that's fine, he's welcome to do so, but it reinforces a serious problem.
Key-person risk is a HUGE practical problem for D :/

> I now recognise a really major conundrum; I've recently returned to D to
> start a company with a greenfields project. VisualD failing is
> essentially terminal. I'm not sure what to do. I don't have time
> available to try and pick up the project and work it myself, but the
> current state is really pushing at the border of forcing me to
> completely rewrite all my code in C++ on account of ecosystem reliability.

Over the past year, VisualD has broken a couple of times.

The fix appears to update & recompile it.

To be clear, it's broken in almost every single semantic-related task that it performs.
The way you describe here is because MS have changed their VS release model so make major revision increments every couple of months rather than every couple of years... and so yes, it needs to be regularly recompiled, but that's not the cause of any functionality regression.

Personally I only use VisualD for debugging, and Intellij for development.
 
It's still the most competent debugger on earth.