Re: Bit field issues in Solaris 2.8

From: Tom Hughes (thh_at_cyberscience.com)
Date: Thu Aug 07 2003 - 00:28:31 BST

  • Next message: Matt Seitz: "ups Next and Step acting like Cont"
    In message <200308062104.h76L4IEt022183_at_tpd.usa.alcatel.com>
              Steve Harris <slharris_at_tpd.usa.alcatel.com> wrote:
    
    > UPS expands the structure and displays all of the bit fields.
    > 
    > Problems:
    > Editing the first bit field of a byte will change the entire byte
    > (edited field + next 7 bits).
    > 
    > Editing any other bit proves futile as it will not change.
    
    Have you tried this in 3.38beta1 or a CVS build? Only I did rework
    the bitfield editing stuff to fix a number of bugs and that was probably
    too recent to be in 3.37 I think.
    
    I'm pretty sure it works right the moment with gcc and C code on
    Linux so I'd be quite surprised if it is failing on Solaris, but it
    is possible I guess.
    
    > Not only editing but if left to run in UPS, the program cannot seem
    > to manipulate the bit fields without altering a byte at a time. Outside
    > of UPS or in another debugger such as ddd, the bit fields are handled
    > properly.
    
    This seems extraordinarily unlikely as the code is still run by
    the processor int the normal way so I don't see how having ups
    attached as a debugger could cause the program itself to start
    behaving differently like this.
    
    One hing that may be catching you is that older versions of ups
    may misreport the value of bitfields I think - that was all part
    of my fixes.
    
    Tom
    
    -- 
    Tom Hughes (thh_at_cyberscience.com)
    Software Engineer, Cyberscience Corporation
    http://www.cyberscience.com/
    


    This archive was generated by hypermail 2.1.4 : Fri Aug 08 2003 - 11:25:41 BST