Re: Cannot compile sane-1.0.2 on SuSE 6.4

From: Henning Meier-Geinitz (hmg-ml@gmx.de)
Date: Thu Apr 27 2000 - 14:02:18 PDT

  • Next message: Oliver Schwartz: "Transparency support for snapscan backend"

    Hi,

    > I believe this patch should fix both problems.

    That doesn't seem to be sufficient (at least for me). When I compile with
    NDEBUG I get:

    [...]
    microtek.c: In function id_microtek':
    microtek.c:1966: invalid lvalue in assignment
    [...]

    This line is:

          DBG_LEVEL = 1;

    I think this will resolve to (0) = 1;

    Maybe we should do the same as in the case of NDEBUG not set:

    #define DBG_LEVEL PASTE(sanei_debug_,BACKEND_NAME)

    #if defined(BACKEND_NAME) && !defined(STUBS)
    int PASTE(sanei_debug_,BACKEND_NAME);
    #endif

    and set the variable to 0 instead of DBG_LEVEL? I have to confine that I
    don't understand this fully, e.g. the #if defined(BACKEND_NAME) but
    BACKEND_NAME is already used 2 lines before.

    Further more I get a lot of warnings about empty bodies in if and else
    statements like this one:

    microtek.c:344: warning: empty body in an if-statement

    The corresponding line:
        if (sense[2] & 0x10) DBG(10, "sense: ERR_ILLBPP\n");

    The problem is that DBG is defined to be empty (if HAVE_VARARG_MACROS is
    defined) so the statement resolves to:

        if (sense[2] & 0x10) ;

        
    If it matters, my system is:
    Linux 2.2.14, i386, glibc2.0.7, egcs-1.1.2 and gcc 2.7.2.3

    Bye, Henning

    --
    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 : Thu Apr 27 2000 - 14:02:48 PDT