call for testing: print/hplip update, pass 2

[Juergen Lock]

On Sat, Jul 11, 2009 at 12:07:05AM +0200, Juergen Lock wrote:
> On Thu, Jul 09, 2009 at 10:44:46PM +0400, Dmitry Marakasov wrote:
> > * Juergen Lock (nox at jelal.kn-bremen.de) wrote:
> > 
> > > I got a new printer today (the old one's scanner died... ): and now hplip
> > > has usb issues with that one:
> > > 
> > >  hp-systray reports on startup: (and also on refresh)
> > 
> > Thanks for testing, Jurgen! I've merged your changes into the port
> > and polished it a bit, so now it builds on all FreeBSD versions and
> > without plist problems.
> > 
> > Unfortunately I'm unable to test this and I have very shallow vision
> > of functionality and regressions of this update.
> > 
> > First of all, here's latest version of the port:
> > http://people.freebsd.org/~amdmi3/hplip.shar
> > Please use it as a base for further changes (and also please check
> > that I haven't missed any of your enchantments).
> > 
> This got lost:
> 
> Index: Makefile
> @@ -84,7 +84,12 @@
>  		${WRKSRC}/ui4/devmgr5.py
>  	@${REINPLACE_CMD} -e 's|/etc/dbus-1/system.d|${LOCALBASE}/etc/dbus-1/system.d|g ; \
>  			s|/usr/share/dbus-1/system-services|${LOCALBASE}/share/dbus-1/system-services|g' \
> -		${WRKSRC}/Makefile.am ${WRKSRC}/Makefile.in
> +		${WRKSRC}/Makefile.am ${WRKSRC}/Makefile.in \
> +		${WRKSRC}/check.py ${WRKSRC}/hpssd.py \
> +		${WRKSRC}/base/g.py ${WRKSRC}/fax/backend/hpfax.py \
> +		${WRKSRC}/installer/distros.dat \
> +		${WRKSRC}/installer/core_install.py \
> +		${WRKSRC}/prnt/hpijs/ljzjs.cpp
>  	@${REINPLACE_CMD} 's|/etc/xdg|${PREFIX}&|g' \
>  		${WRKSRC}/configure ${WRKSRC}/configure.in
>  	@${REINPLACE_CMD} 's|-lpthread|${PTHREAD_LIBS}|g' \
> 
> > As I understood, the update has regressions compared to the hplip
> > version which is currently in ports, thus it's not possible to just
> > update it. Can you (both Jurgen and Chris) give me a brief summary of
> > how the update is better and worse than the current version? If it has
> > enough working/new features, it would be nice to commit this change as
> > hplip-devel. And, in either case, how hard is it (if possible) to
> > fix the regressions?
> 
> Well the usb issue with my new printer also exists with the old version
> thats currently in ports as I now found out, and the only other issue
> that I remember (printing ascii or ps to the fax queue) could just as
> well be a problem/incompatiblity with the new cups version...  And
> additionally, the old version's hp-setup seems to be incompatible with
> today's cups, at least it was unable to add the network print queues
> for my new printer when I tried it now.  (unlike the new version.  And
> also the gui certainly has improved.)
> 
>  Btw that testing just unconvered another bug: the package needs to add
> hpaio to ${LOCALBASE)/etc/sane.d/dll.conf too...
> 
>  I guess I should test the fax queue on Linux, and possibly usb on head...
> (fixing usb would at least need someone that still knows the old usb
> stack I'd say.  _If_ it works on head. :)

Alright, I just looked at this again and got usb back to working here
(7-stable/amd64) with the following simple patch:

Index: io/hpmud/musb.c
@@ -122,7 +122,7 @@ static int get_string_descriptor(usb_dev
    {
       ret = usb_control_msg(dev, USB_ENDPOINT_IN, USB_REQ_GET_DESCRIPTOR, (USB_DT_STRING << 8) + index, 
                0x409, tbuf, sizeof(tbuf), LIBUSB_CONTROL_REQ_TIMEOUT);
-      if (ret==0)
+      if (ret==0 || ret == -EIO)
       {
 	 /* This retry is necessary for lj1000 and lj1005. des 12/12/07
          Also HP Photosmart 42xx seems to suffer transient errors with serial id */

 My guess is other usb stacks return 0 here where ours returns -EIO...
(or at least or old one does, still haven't tested this on 8.0 because
head has ata issues on this box, see
	http://docs.freebsd.org/cgi/mid.cgi?20090628101656.GA38983
and
	http://lists.freebsd.org/pipermail/freebsd-current/2009-July/009013.html
.)

 HTH,
	Juergen