This is really sounding very much like my experience except that mine only
took several months to evolve into something that I could diagnose instead
of several years like yours.  It will all make perfect sense once you see
the inside of the pressure regulator.

I will try to describe the innards of the regulator.
Imagine a vertical piece of 3/16" ID metal pipe -- this is the overflow back
to the gas tank out the bottom and the top is covered by a round flat disc
that blocks the exit.  This disc is attached to a rubber diaphragm that will
lift the disc off the end of the overflow pipe when the correct pressure is
reached.  The outside diameter of the rubber diaphragm is sealed between the
two chambers of the housing.  In the upper chamber of the housing is a
compression coil spring that preloads the disc onto the end of the overflow
pipe.  The lower chamber of the housing is connected to the fuel rail and is
full of fuel at the regulated pressure.  The upper chamber has a port that
is plumbed to manifold vacuum so as to subtract the manifold vacuum from the
spring pressure.  As fuel is pumped into the system the pressure rises until
it is enough to overcome the spring preload and raise the disc up off the
overflow pipe thereby sending excess fuel back to the tank.

The peculiar behavior of this malfunction occurs because the way the
regulator disc is attached to the rubber diaphragm allows it to become
disconnected when the fuel pressure is high and then to reattach itself when
the engine is turned off and the spring pushes the diaphragm back down onto
the disc.  The disc has a small sphere the size of a bb welded onto it's
upper surface.  This ball is held to the diaphragm by a thin sheet metal
inverted cone that is part of the diaphragm.  The cone has slits on all four
sides as a way to form it around the ball after the ball is welded to the
disc so the 'leaves' of the cone are springy.  After many years wear of the
ball and the leaves of the cone make it possible for the fuel pressure to
pull the ball out of the cone so that the regulator no longer works and the
pressure goes sky high.  After the engine is turned off, the pressure goes
away, the spring pushes the diaphragm back down, and the ball snaps back
into the cone -- perhaps at a little different clocking so it may stay
attached for some time.

In the beginning my car would only stall at idle once every week or so and
would always start right back up and be fine for another week.  At the end
it would run OK for 15 minutes in the morning and stall at every stop after
that.  I finally put a fuel pressure gage on it and ran it for 15 minutes so
I got to see the pressure shoot up to over 100 and my fuel hoses all try to
straighten out.

It can quite easy to test the pressure yourself depending on the pipe
layout. I'm not familiar with the E30 - has it got a separate cold start
injector? That can be a convenient place to remove the pipe and replace
with a gauge.  But best to de-pressurize the rail first.

You could always get an adjustable one off Ebay - some have a pressure
gauge fitted. Then keep it for your next car. Or try a breakers yard -
they're not something that normally totally fail. They're often not that
accurate, though.

Check Ebay for a secondhand one? I've got an old but working 36 psi one
from a Rover EFI somewhere.