On Tue, May 30, 2000, Daniel Streetman <[EMAIL PROTECTED]> wrote:
> On Tue, 30 May 2000, Johannes Erdfelt wrote:
>
> > On Tue, May 30, 2000, Thomas Sailer <[EMAIL PROTECTED]> wrote:
> > > Daniel Streetman wrote:
> > >
> > > > I'm assuming the 'devices' file is intended for users, not programs.
> > >
> > > Wrong assumption 8-) The file is also intended to be machine
> > > parseable.
> > >
> > > That's how lsusb -t gets the topology, I suggest looking at it
> > > (usb.in.tum.de, CVS module usbutils).
> >
> > Ugh, before anyone starts, please don't bring the binary vs machine
> > parsable ASCII war onto linux-usb. Leave it on linux-kernel please.
>
> Yes, ugh. My concern was not a binary vs ASCII issue, I just didn't want
> to read the entire 'devices' file on every connect/disconnect.
Just as an aside (which you may or may not know), you can poll or select
on the 'devices' file to be notified when something changes. It doesn't
tell you what changed tho.
> > This is incorrect. Please look at my last patch. I don't know which -ac
> > release it went into (or if even did go in), but it was sent on by Randy
> > atleast.
>
> Ok. I like that (the patch) better than manually asking the devices for
> their config descriptors.
It's easier, but the original reason for this was that some devices
don't like sending back the descriptors again after it's already done
it. This is a violation of the USB spec, but we can workaround it
easily.
JE
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
