Hi.
This still doesn't seem to be in CVS, so building Sane out-of-the-box
seems problematic at the moment.
Could anyone with write access check in the obvious fixes to
m3096g-scsi.h & sp15c-scsi.h ?
Thanks,
John.
Randolph Bentson wrote:
>
> John Vickers wrote:
> > It seems like m3096g-scsi.h & sp15c-scsi.h define `lint_catcher' in
> > conflicting ways. m3096g.c & sp15c.c seem to be the only compilation
> > units which use this identifer.
>
> I created the lint_catcher to reduce compiler complaints about
> defined, but unused structures to a single complaint. The
> structures are defined for commands which aren't implemented.
>
> > I don't know if this is what was intended, but I made `lint_catcher'
> > static in both places.
>
> John is correct in pointing that defining them static will pull
> them out of the external symbol table. Good catch--my compiler
> didn't complain.
>
> > This made the build run to completion, anyway. Anyone want to check
> > this in ?
>
> I've made the change to my source, but I don't have access to the
> CVS tree either. I hope someone will check in the proposed changes
>
> Randolph Bentson
> embarassed author of sp15c and m3096g
-------- Original Message --------
Subject: CVS build fails: backend/m3096g-scsi.h & backend/sp15c-scsi.h
Date: Sun, 30 Jan 2000 20:55:47 +0000
From: John Vickers <jvickers@dial.pipex.com>
Reply-To: sane-devel@mostang.com
Organization: ---
To: sane-devel@mostang.com
Hi.
After patching config.h.in, the build fails with:
gcc -L/usr/lib -L/usr/X11R6/lib -rdynamic -o scanimage scanimage.o
../backend/.libs/libsane.a ../lib/liblib.a -ldl -lnsl -ljpeg -lm
../backend/.libs/libsane.a(sp15c.o): In function
`sane_sp15c_get_devices':
/home/pkg/source/sane/sane/backend/sp15c.c:237: multiple definition of
`lint_catcher'
../backend/.libs/libsane.a(m3096g.o):/home/pkg/source/sane/sane/backend/m3096g.c:236:
first defined here
/usr/bin/ld: Warning: size of symbol `lint_catcher' changed from 64 to
72 in sp15c.o
collect2: ld returned 1 exit status
make[1]: *** [scanimage] Error 1
It seems like m3096g-scsi.h & sp15c-scsi.h define `lint_catcher' in
conflicting ways.
m3096g.c & sp15c.c seem to be the only compilation units which use this
identifer.
I don't know if this is what was intended, but I made `lint_catcher'
static in both places.
This made the build run to completion, anyway. Anyone want to check
this in ?
What is an sp15c or an m3096, anyway ???
Regards,
John.
Index: backend/m3096g-scsi.h
===================================================================
RCS file: /cvsroot/external/sane/backend/Attic/m3096g-scsi.h,v
retrieving revision 1.1.2.2
diff -c -p -r1.1.2.2 m3096g-scsi.h
*** m3096g-scsi.h 2000/01/26 03:51:45 1.1.2.2
--- m3096g-scsi.h 2000/01/30 20:54:31
*************** static scsiblk window_descriptor_blockB
*** 523,529 ****
/*
==================================================================== */
! scsiblk *lint_catcher[] =
{&reserve_unitB,
&release_unitB,
&inquiryB,
--- 523,529 ----
/*
==================================================================== */
! static scsiblk *lint_catcher[] =
{&reserve_unitB,
&release_unitB,
&inquiryB,
Index: backend/sp15c-scsi.h
===================================================================
RCS file: /cvsroot/external/sane/backend/Attic/sp15c-scsi.h,v
retrieving revision 1.1.2.2
diff -c -p -r1.1.2.2 sp15c-scsi.h
*** sp15c-scsi.h 2000/01/26 03:51:48 1.1.2.2
--- sp15c-scsi.h 2000/01/30 20:55:03
*************** static scsiblk window_descriptor_blockB
*** 525,531 ****
/*
==================================================================== */
! scsiblk *lint_catcher[] =
{&reserve_unitB,
&release_unitB,
&inquiryB,
--- 525,531 ----
/*
==================================================================== */
! static scsiblk *lint_catcher[] =
{&reserve_unitB,
&release_unitB,
&inquiryB,
---- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
This archive was generated by hypermail 2b29 : Mon Feb 14 2000 - 01:00:28 PST