Re: [hpoj-devel] Success with G85

[Allen B. <ba...@lo...>]

"PASCHAL,DAVID (HP-Roseville,ex1)" wrote:
> 
> Allen Barnett wrote:
> > I installed 0.8 of the OfficeJet drivers per the very detailed
> > instructions in the README files and have successfully
> > scanned from, and
> > printed to, a G85 connected through its USB port.
> 
> Regarding the "very detailed instructions", do you think they are too
> detailed in the README/INSTALL/*-HOWTO files?  I'm thinking they are and
> that I should provide simpler (less detailed) setup instructions there and
> move most of the details into separate manual pages for each daemon or
> application.  Do you have any thoughts one way or the other?


I had to perform each step as outlined in your instructions for both
scanning and printing, so I do not consider them too detailed! I had to
download and install SANE (which is now separated into two pieces, by
the way, sane-backends-1.0.4 and sane-frontends-1.0.4; I still haven't
been able to connect to ftp.mostang.com; I ended up getting the backend
source from TUCOWS). So, I'm not sure how you could simplify them, at
least until distributions catch up with the latest versions of SANE, the
2.4 kernel and so forth. Of course, a GUI installer for the totally
hapless might be nice :-)



> How was your experience with updating printer.c in the Linux kernel source
> (if you weren't already running a recent-enough kernel)?  I'm trying to
> decide if I need to provide a sufficiently up-to-date version with the hpoj
> package so people don't have to mess with downloading and updating their
> kernel (or just extracting that file).

I had actually upgraded to 2.4.2 and installed the hotplug scripts a
couple of days ago for an unrelated reason, so I just had to add the
printer.o module to the kernel config. When I plugged in the USB cable,
hotplug went ahead and loaded the driver. Remarkable.

For comparison, the XFree86 DRI OpenGL driver has a kernel module as
well as user-space components. Both components are provided in the CVS
repository at dri.sourceforge.net. However, some version of the kernel
module also shows up in the kernel source. These occasionally get out of
sync, causing mysterious failures (till someone knocks their head after
checking their X log). If the printer.o module is relatively stable or
at least relatively independent of the user-space driver, then it might
be more convenient to provide it with the HPOJ source, otherwise, I'd
leave it out.

Allen