[PATCH] pnp.ids
Pierre Ossman
drzeus-list at drzeus.cx
Mon Feb 7 09:00:49 PST 2005
David Zeuthen wrote:
> Hi,
>
> No, I still don't think it make sense to distribute this file with
> hal; the only thing it does, in my view, is to increase complexity
> since we would need code to open/read/close yet another file. We can
> already do the same with device information files; we really don't
> need another format. Do you disagree?
It's a startup issue, so I think flexibility can be prioritised in these
cases.
As for distributing the file with hal, that was just meant to be a
temporary measure. When a proper page is set up for mainting the data in
the file it shouldn't be too much of a hassle getting it included in the
hwdata packages of the different distributions.
>
> Actually, I've been considering giving the pci.ids and usb.ids the
> same treatment; write a script that converts the files to a table (.h
> file) and use that for look-up in hald. While doing that we should
> change the look-up to use a binary search also or whatever to do
> better than O(n). If someone got a patch for this, I'd be happy to
> take it :-)
Is there really a performance issue with this? In any case this can
still be done by loading the files into memory in a more intelligent
way. Duplicating the data as header files seems like a maintenance
nightmare to me. I find the idea that hal needs to be recompiled every
time a new piece of hardware is released a bit strange.
Rgds
Pierre
_______________________________________________
hal mailing list
hal at lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/hal
More information about the Hal
mailing list