[sane-devel] SANE 1.0.12 on Mac OSX 10.2.8

Henning Meier-Geinitz henning@meier-geinitz.de
Sat, 15 Nov 2003 18:36:24 +0100


Hi,

On Tue, Nov 11, 2003 at 07:49:40PM -0600, jakobie wrote:
> I am trying to get SANE working on my mac, it can see the scanner but 
> that is about it. I ran debug for both USB and the scanner. I'm working 
> with Mac OSX 10.2.8, SANE 1.0.12, UMAX Astra 2200 via USB.

The umax backend uses fork() which doesn't work correctly with USB
scanners on MacOS X. So until the backend is concerted to use
the sanei_thread stuff, it won't be able to scan.

> ./scanimage
> [sanei_debug] Setting debug level of umax to 255.
> [umax] sane_init
> [umax] This is sane-umax version 1.0 build 40
[...]
> scanimage: no SANE devices found
> [umax] sane_exit
> [jakobie:/usr/local/bin] jakobie% clear

Huh? The detection only works after the second run of scanimage?

> [jakobie:/usr/local/bin] jakobie% setenv SANE_DEBUG_UMAX=255; 
> ./scanimage
> [sanei_debug] Setting debug level of umax to 255.
> [umax] sane_init
> [umax] This is sane-umax version 1.0 build 40
[...]
> [umax] sanei_usb_attach_matching_devices(usb 0x1606 0x0230)
> [umax] attach_scanner: libusb:001:007, connection_type 2
> [umax] attach_scanner: opening usb device libusb:001:007

found...

> [umax] start_scan
> [umax] starting scan
> [umax] Sending SCSI cmd 0x1b cdb len 6, param len 1, result len 0
> [umax]   SCSI command successfully executed
> [umax] do_calibration
> [umax] wait_scanner
> [umax] Sending SCSI cmd 0x00 cdb len 6, param len 0, result len 0
> [umax] error in sanei_pv8630_bulkread (got 02)
> [umax] scanner reports Error during device I/O, waiting ...
> [umax] Sending SCSI cmd 0x00 cdb len 6, param len 0, result len 0
> [umax] error in sanei_pv8630_bulkread (got a5)
> [umax] Sending SCSI cmd 0x00 cdb len 6, param len 0, result len 0
>  [umax] error in sanei_pv8630_bulkread (got a5)
> [umax] Sending SCSI cmd 0x00 cdb len 6, param len 0, result len 0

If I understand the code correctly, it should loop up to 100 times
here until the scanner is ready.

For me that doesn't look like the fork() problem. But maybe the author
of the umax backend has an idea about that: Oliver Rauch
<Oliver.Rauch@Rauch-domain.de>.

Bye,
  Henning