[sane-devel] sane-backends 1.0.20 Released

Julien BLACHE jb at jblache.org
Mon May 4 14:03:17 UTC 2009


Chris Bagwell <chris at cnpbagwell.com> wrote:

Hi Chris,

> Are you referring to sane-backends?  I can't find a --enable-pthreads
> option.  Also, I tried for the first time the --enable-fork-process
> option but still had no compile issues.

Yes, that's sane-backends and the option is actually
--disable-fork-process (which does exactly what --enable-pthreads
would do if it existed ;))

It switches sanei_thread from using fork() to using pthreads for
reader processes.

> Oh wait, grepping for pthreads in backend directory, I see references
> to pthread functions in mustek, pixma, and snapscan but I'm only
> linking in for mustek.

In scapscan pthread_*() functions are used when sanei_thread uses
pthreads to do some locking. They are not used otherwise.

> Also, can you verify those two backends align with what your seeing fail?

The first one I see failing is agfafocus. Any sanei_thread user needs
pthreads when --disable-fork-process is used, so there's some automake
foo to be done here.

Easiest would be a new variable for pthreads libs when
--disable-fork-process is used, so you can just define this variable
when needed and leave it empty otherwise. And add it to all the
sanei_thread users.


There's also the issue of excessive dependencies on libsane.so; we
should make a decision about that. Though it's already much better
having to fix a couple of variables in a Makefile rather than
extensively patching a Makefile :) My amount of patching has
dramatically decreased with this release, thanks for that :)

JB.

-- 
Julien BLACHE                                   <http://www.jblache.org> 
<jb at jblache.org>                                  GPG KeyID 0xF5D65169



More information about the sane-devel mailing list