please include Patch for sane-backend - hp3500c - fixing lights issues and probably other uninitialisation problems

[epek]

from a correspondence with troy rollo:

I asked:
> After finishing a scan of the first page the slay stays in its current
> > (ending) position. It does not return to its parking position (top)with
> > simple scan (reproducible) . Scanning further pages is hence not
> > possible until the scanner gets disconnected and reconnected again.
> > (Simple Scan crashes then, of course).

He answered
This looks like the problem fixed by commit
8ec394d6489dac4226fd22f6cac02dd1a89c400c which was included in version
1.0.21.

http://fedorapeople.org/gitweb?p=nphilipp/public_git/sane-backends.git;a=commitdiff;h=8ec394d6489dac4226fd22f6cac02dd1a89c400c

lists the following changes

diff --git a/backend/hp3500.c b/backend/hp3500.c

index 56b48b0..6844a5e 100644 (file)

--- a/backend/hp3500.c

+++ b/backend/hp3500.c

@@ -3392,27 +3392,28 @@ reader_process (void *pv)

   sigset_t sigterm_set;

   struct SIGACTION act;

   struct hp3500_write_info winfo;

+ int status;

   if (sanei_thread_is_forked ())

     {

       close (scanner->pipe_r);

- }

- sigfillset (&ignore_set);

- sigdelset (&ignore_set, SIGTERM);

-#if defined (__APPLE__) && defined (__MACH__)

- sigdelset (&ignore_set, SIGUSR2);

+ sigfillset (&ignore_set);

+ sigdelset (&ignore_set, SIGTERM);

+#if defined (__APPLE__) && defined (__MACH__)

+ sigdelset (&ignore_set, SIGUSR2);

 #endif

- sigprocmask (SIG_SETMASK, &ignore_set, 0);

+ sigprocmask (SIG_SETMASK, &ignore_set, 0);

- sigemptyset (&sigterm_set);

- sigaddset (&sigterm_set, SIGTERM);

+ sigemptyset (&sigterm_set);

+ sigaddset (&sigterm_set, SIGTERM);

- memset (&act, 0, sizeof (act));

-#ifdef _POSIX_SOURCE

- act.sa_handler = sigtermHandler;

+ memset (&act, 0, sizeof (act));

+#ifdef _POSIX_SOURCE

+ act.sa_handler = sigtermHandler;

 #endif

- sigaction (SIGTERM, &act, 0);

+ sigaction (SIGTERM, &act, 0);

+ }

   /* Warm up the lamp again if our last scan ended more than 5 minutes ago. */

@@ -3447,6 +3448,8 @@ reader_process (void *pv)

        scanner->actres_pixels.bottom - scanner->actres_pixels.top,

        scanner->resolution, scanner->mode, (rts8801_callback) writefunc,

        &winfo) >= 0)

- exit (SANE_STATUS_GOOD);

- exit (SANE_STATUS_IO_ERROR);

+ status = SANE_STATUS_GOOD;

+ status = SANE_STATUS_IO_ERROR;

+ close (scanner->pipe_w);

+ return status;

 }

[Rolf Leggewie (r0lf)]

epek, sorry for the late response. Is this something that still affects precise or later?

[Rolf Leggewie (r0lf)]

Answering my own question here.

If this indeed this was fixed in 1.0.21 then anything after oneiric should have this fix included. Thus, closing as fixed. Hardy and Lucid might still be affected but I am not sure about backporting. Feel free to supply a debdiff if you need support in lucid. If possible, an update to precise might be the wiser choice.