[sane-devel] sane-backends-WIN32
Dave Burns
dave.burns@comcast.net
Mon, 21 Jun 2004 09:12:50 -0400
I hope I'm not completely out of context since I only get the digest and
have not been following this thread. But the following caught my eye:
"With xsane the stdout and sterr are not available because it is compiled
as GUI."
As a WIN32 programmer for years, I know this doesn't have to be true. Use
the following code to open a Win32 console window and rewire stdout/stderr
to its output. Wrap the whole thing in a conditional that's true if you want
debug output. Hope this helps. If I'm off topic, apologies.
db
#include <windows.h>
#include <io.h> /* For _open_osfhandle( ) */
#include <fcntl.h> /* For _O_TEXT */
{
int hCrt_StdOut;
int hCrt_StdErr;
FILE *hf_stdout;
FILE *hf_stderr;
/* First create the console. This only opens the window. This will fail if
we are already running from a console window (which means we don't need
to connect any output streams). */
if (AllocConsole()) {
HANDLE stdoutBuffer;
CONSOLE_SCREEN_BUFFER_INFO bufferInfo;
/* Now connect the Win32 STD_OUTPUT pipe to the c-runtime stdout stream
*/
stdoutBuffer = GetStdHandle(STD_OUTPUT_HANDLE);
hCrt_StdOut = _open_osfhandle((long) stdoutBuffer, _O_TEXT);
hf_stdout = _fdopen(hCrt_StdOut, "w");
*stdout = *hf_stdout;
/* setvbuf(stdout, NULL, _IONBF, 0);*/
/* Now connect the Win32 STD_ERROR pipe to the c-runtime stderr stream
*/
hCrt_StdErr = _open_osfhandle((long) GetStdHandle(STD_ERROR_HANDLE),
_O_TEXT);
hf_stderr = _fdopen(hCrt_StdErr, "w");
*stderr = *hf_stderr;
setvbuf(stderr, NULL, _IONBF, 0);
/* now we increase the buffer size to 9999 lines. We only do this
if we created a console so that we don't mess with a user's
existing console. */
GetConsoleScreenBufferInfo(stdoutBuffer, &bufferInfo);
if (bufferInfo.dwSize.Y < 9999)
bufferInfo.dwSize.Y = 9999;
SetConsoleScreenBufferSize(stdoutBuffer, bufferInfo.dwSize);
}
}
-----Original Message-----
From: sane-devel-admin@lists.alioth.debian.org
[mailto:sane-devel-admin@lists.alioth.debian.org] On Behalf Of
sane-devel-request@lists.alioth.debian.org
From: Oliver Rauch <Oliver.Rauch@Rauch-Domain.DE>
To: sane-devel <sane-devel@lists.alioth.debian.org>
Organization:
Date: 19 Jun 2004 13:44:12 +0200
Subject: [sane-devel] sane-backends-WIN32
Hello.
Here a little experience report with sane-backends on WIN32.
I used xsane-win32-0.94-sane-1.0.14-complete-no-gimp.zip for my tests with a
UMAX Powerlook III SCSI-scanner on Windows 2000.
I can create debugging messages with
set SANE_DEBUG_UMAX=12
or
set SANE_DEBUG_SANEI_SCSI=128
and scanimage.
With xsane the stdout and sterr are not available because it is compiled as
GUI.
The status is:
1) xsane recognices the scanner without any changes.
2) The xsane interface comes up with correct options
=> INQUIRY is done without problems
3) Starting a scan or a preview makes two problems:
a) SCSI Buffer size seems to be limited to 32KB,
that is too small for this scanner in Color mode.
Increasing the size by umax.conf options does not work.
b) the START_SCAN command returns with I/O Error.
The SET_WINDOW command and SEND_GAMMA_DATA in grayscale mode does not make
any problems.
It is encouraging that the INQUIRY, SET_WINDOW and SEND_GAMMA_DATA commands
work without problems. I do not have an idea why the START_SCAN command
fails.
Please also report your experiences with sane-backends on WIN32.
I think that we only have to do some little sourcecode changes to make it
work. If we succeed with this then SANE gets the chance to be THE SCANNER
DRIVER STANDARD and may be one day the manufacturers start writing their own
(windows) drivers as SANE driver.
Best regards
Oliver
--__--__--
Message: 2
Date: Sat, 19 Jun 2004 14:09:17 +0200
From: Kees Lemmens <C.W.J.Lemmens@ewi.tudelft.nl>
To: sane-devel@lists.alioth.debian.org
Subject: [sane-devel] Re: strange rattle in epson 610 with 2.6.7 kernel
Hi,
First problem : rattle epson 610 (solved)
-------------
Just installed sane 1.0.13 and this works fine under 2.6.7. However turning
back to sane-1.0.12 immediately gives the alarming rattle back ...
BTW: I used this script to test without having to reboot/reinstall :
####################################################################
#!/bin/sh
#SANE=/opt/sane-1.0.13 # Newer xsane release
SANE=/usr/ # Standard sane-1.0.12 from Slackware 9.1
export LD_LIBRARY_PATH=$SANE/lib/
export LD_PRELOAD=$SANE/lib/libsane.so
exec /usr/bin/xsane $*
####################################################################
Second problem : compile 1.0.14 under Slackware fails
--------------
Whatever I try sane-1.0.14 doesn't compile under Slackware 9.1 !!
configure fails to detect the presence of basic routines like "strdup",
"getenv" and others under Slackware 9.1, thereby making it impposible to
compile (at least on my system).
Any ideas ?
giant:/tmp/sane-backends-1.0.14 252 % make making all in include
make[1]: Entering directory /tmp/sane-backends-1.0.14/include'
make[1]: Nothing to be done for all'.
make[1]: Leaving directory /tmp/sane-backends-1.0.14/include'
making all in lib
make[1]: Entering directory /tmp/sane-backends-1.0.14/lib'
gcc -c -DHAVE_CONFIG_H -I. -I. -I../include/sane -I../include
-DPATH_SANE_CONFIG_DIR=/opt/sane-1.0.14//etc/sane.d
-DPATH_SANE_DATA_DIR=/opt/sane-1.0.14//share -DV_MAJOR=1 -DV_MINOR=0
-I/usr/local/include/gphoto2 -g -O2 -W -Wall getenv.c
getenv.c:13:4: #error "Missing getenv() on this platform. Please
implement."
getenv.c: In function getenv':
getenv.c:6: warning: unused parameter name'
make[1]: *** [getenv.o] Error 1
make[1]: Leaving directory /tmp/sane-backends-1.0.14/lib'
make: *** [all-recursive] Error 1
giant:/tmp/sane-backends-1.0.14 254 % less include/sane/config.h ...
/* Define to 1 if you have the getenv' function. */
/* #undef HAVE_GETENV */
...
/* Define to 1 if you have the strcasecmp' function. */
/* #undef HAVE_STRCASECMP */
/* Define to 1 if you have the strdup' function. */
/* #undef HAVE_STRDUP */
Bye,
Kees Lemmens
On Fri, Jun 18, 2004 at 09:40:03AM +0200, Kees Lemmens wrote:
> Hi,
>
> I have an Epson Perfection 610 which worked fine for 2 years with
> 2.2.x and 2.4.x.
>
> Today I compiled 2.6.7 and now it suddenly produces strange and
> alarming sounds after the scan is finished and it returns to initial
position.
>
> I thought the scanner failed, but after rebooting in 2.4.26 it works
> fine again, so I am afraid it must be something in the 2.6.x kernel
> combined with the sane 1.0.12 backend that I use (the one provided
> with Slackware 9.1).
>
> Any suggestions ? Should I just install the latest backend or is it
> something else ?
>
> --
> Bye,
> Kees Lemmens.
> ----------------------------------------------------------------------
> -
> NOTICE: I consider all HTML based email as SPAM : forward => trash
--__--__--
_______________________________________________
sane-devel mailing list
sane-devel@lists.alioth.debian.org
http://lists.alioth.debian.org/mailman/listinfo/sane-devel
End of sane-devel Digest