new database code

[John E. Malmberg]

Andrew Tridgell <tridge at linuxcare.com> wrote:

> what sort of optimisations? It should be pretty fast on most systems
> already. It will be slower on those without mmap(), but probably still
> acceptably fast.

The OpenVMS operating system has Indexed file support.  A quick glance at
the the description that you reference shows that much of the functionality
is readily available in the native record manager.

The VMS CRTL has mmap() but the program in the tests directory failed to
share data.

It also has a in memory locking manager that can also be used, but there is
no direct analog to standard C functions.

Using the native locking manager instead of the .lck file or other database
file for keeping track of running processes would be a speed enhancement for
VMS, but I am not sure on how much, since it is basically run once code at
the daemon startup.

> > Is the "locking" library going away?
>
> all the old shmem locking code is gone. locking.c is has been
> extensively modified to use the new tdb code.
>
> > If not, will it be changed to not reference routines in server.c?  This
kept
> > me from converting it to a shared image.
>
> what routines are you referring to?

I will have to look that up in a little bit, as I find I did not keep as
detailed notes on this as I thought.  Basically it involves uid.c and
server.c and the IIRC: unbecome_user() and exit_server().

I have to run off just right now.

-John
wb8tyw at qsl.net