Re: flexfax: More compile problems with Linux 2.2

["David A. De Graaf" <degraaf@rhsnet.com>]

Robert Colquhoun wrote:
> 
> At 09:59 PM 5/24/99 -0400, you wrote:
> >
> >On 17-May-99 Robert Colquhoun wrote:
> >> Hi,
> >> 
> >> I have added the configure fixes(so that configure looks in the right place
> >> for the function prototypes, and to automatically select the utmp interface
> >> for linux).  As a temporary measure i also added the '#define fds_bits'
> hack.
> >
> >Still doesn't work?
> 
> oh...
> 
> >> Sometime later this week(depending on how busy i am) i will try to post the
> >> above plus some updated documentation as hylafax-v4.0pl2-rjc7.  This
> >> release will also contain the y2k leap year fix.
> >
> >The above is using the new rjc7 tarball...
> 
> Try getting rid of the #ifdefs i put around the #define in Dispatcher.c++,
> and see if that works(see patch below).
> 
> - Robert
> 
> ===================================================================
> RCS file: /usr/local/cvsroot/hylafax/util/Dispatcher.c++,v
> retrieving revision 1.3
> diff -c -r1.3 Dispatcher.c++
> *** Dispatcher.c++      1999/05/20 06:33:09     1.3
> --- Dispatcher.c++      1999/05/25 07:08:01
> ***************
> *** 45,53 ****
>       XXX This is a bad hack to get it to work with glibc 2.1, should
>       be removed and the code dependent on it fixed at the first opportunity.
>   */
> - #if defined __GLIBC__ >= 2 && __GLIBC_MINOR__ >= 1
>   #define fds_bits __fds_bits
> - #endif
> 
>   class FdMask : public fd_set {
>   public:
> --- 45,51 ----  
> 

I, too, am trying to compile hylafax under RedHat 6.0 without success.
Your patch, with the conditions removed, did allow to Dispatcher.c++
to compile.  This has moved the problem downstream to SendFaxJob.c++.
The number of "warning" messages has mushroomed from 27 with RH 5.2
to 417 with RH 6.0, and the make hasn't even finished!

I am just bewildered by this complex C++ code, and can only hope that
you and other smart programmers can fix these problems, and soon.

It is an awful situation when an apparently worthwhile upgrade like
RH 6.0 renders an essential major package like hylafax unusable.
I surely hope you can find the time and motivation to wade through all
the problems and fix them.  The rest of us are waiting hopefully for
news.

-- 
	David A. De Graaf    DATIX, Inc.    Hilton Head Is., SC  
	degraaf@rhsnet.com       843-785-3136, -3156 (fax)