Re: SANE CVS snapshot (2001-05-27) available

From: Oliver Rauch (oliver.rauch@rauch-domain.de)
Date: Sun May 27 2001 - 14:23:43 PDT

  • Next message: Phil Cummins: "Re: EPSON ES-8500 success"

    Henning Meier-Geinitz wrote:
    >
    > Hi,
    >
    > On Sun, May 27, 2001 at 09:48:32PM +0200, Petter Reinholdtsen wrote:
    > > [Henning Meier-Geinitz]
    > > > I see. I just found out that sanei_codec_ascii is somehow used in
    > > > the xscanimage preferences. I will have a look at this.
    > >
    > > Yes, it seem to be used to save and load the preferences file. Which
    > > is a good idea when one think about it. :-)
    > >
    > > I updated the sane-frontend/sanei/ subdirectory to match the content
    > > of sane-frontend/sanei/. Parhaps we should install libsanei as well
    > > from sane-backends?
    >
    > Won't we get trouble at least with shared libraries because currently
    > it's linked to the backend libraries and then to backends and frontends?
    >

    That is a strange thing. As far as I experienced in such cases is that
    you can not say which function is used, it may happen that one routine
    uses one function and an other routine uses the other function.
    But after all, when both functions are the same this should work,
    but that really is bad!

    > > Other alternatives include rewrite the frontends to not use sanei_wire
    > > functions, or to use different symbol names in the frontends. Both
    > > are bad alternatives in my eyes.
    >
    > It's not only sanei_wire but also the debug functions. I just did some
    > changes (locally) to combine all necessary code in debug.c,
    > store_values.c (load_values, save_values) and wire.c (codec_ascii,
    > sanei_wire). codec_bin doesn't seem to be needed. Renaming everything
    > from sanei_w_* to wire_* or something like that works. It's a bigger
    > change (removing sanei/ completely and most of include/sane). At least
    > it works :-) I will look at this in more detail tomorrow and post a
    > patch.
    >
    > I'm not sure, what's the better approach.

    I think this is the correct way to do this.

    If we include it into the backends we have to link it to each backend,
    not only to sane-dll.
    When we make a common lib "libwire" o sthg like that we have to define an
    API for it. It is not part of the sane api.

    Including it into the frontend increases the frontend binary by about 20KB
    (may be less if we look thrugh the code again).

    XSane also uses a reduced set of the sanei_wire... etc routines.
    This way the frotends and backends really work independent in this case
    and I think this is important!

    Bye
    Oliver

    -- 
    Homepage:	http://www.rauch-domain.de
    sane-umax:	http://www.rauch-domain.de/sane-umax
    xsane:		http://www.xsane.org
    E-Mail:		mailto:Oliver.Rauch@rauch-domain.de
    

    -- 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 : Sun May 27 2001 - 15:02:25 PDT