Thank you for your kind recommendations.
In the end... I didn't run out of tools... but time.
I first looked at my output of "/proc/partitions", but found some oddities.
Did not dare to use -- copied into another machine (just in case).
Then I tried "gpart", but presence of extended partitions confused it.
It could not determine the original partition table layout.
Testdisk arrived at the same result.
Perhaps part of the problem was... that the data which overwrote
my partition table... was not random trash, but contained another parition table,
originating from a CompactFlash card.
In the end... I set the corrupted disk aside for analysis
(later extracted some source code from the Windows partition)
and started installing things onto a new disk (faster too).
Since everything I stored under Linux (read: most things)
were successfully evacuated, I obtained a working system reasonably fast
(and a new kernel too).
So things are... nearly back to normal.
And there's a CD-ROM on my bookshelf.
With double backups of my current partition table.
