EOF detection

From: Marian Eichholz (marian.eichholz@freenet-ag.de)
Date: Mon May 07 2001 - 08:10:25 PDT

  • Next message: abel deuring: "Re: [dev] Memory access violations and leaks"

    Hi friends,

    hey, Oliver made a 0.76 XSane version. Did I miss anything?

    Nevertheless, I browsed the source and I'm pretty shure, that I found
    the reason for the strange scanning behaviour of our little sm3600
    backend.

    If the sane_read() gives SANE_STATUS_EOF, the whole action is discarded.
    In other words:

    - the last valid scan data must be reported with SANE_STATUS_GOOD
    - the SANE_STATUS_EOF must be reported without access to the buffer.

    Hey, I call it somewhat in-sane, because the last call to sane_read()
    will smash the stack, if there is anything written to the buffer address
    (yes, it points to not available data), but perhaps we should't argue on
    that topic :-)

    I think, that we should release the backend now for the first volunteer
    testers, or more precisely: We should integrate it into the backend
    package.

    Is anyone out there, who wants to do it? The issue *now* is the
    prerequisite libusb, and I have no time to fiddle with "configure" to
    set up the Makefile properly. The rest ist almost trivial.

    Mit freundlichen Gruessen / Yours sincerely

    Marian Eichholz

    - - - - - - - - - - -
    Marian Eichholz
    Postmaster
    freenet.de AG Vorsitzender des Aufsichtsrates: Gerhard Schmid
    Deelbögenkamp 4c Vorstand: Eckhard Spoerr (Vors.), Axel Krieger
    22297 Hamburg Amtsgericht Hamburg, HRB 74048

    --
    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 May 07 2001 - 08:01:05 PDT