Re: Mingw32

[James Ingham <jingham@cygnus.com>]

Mumit,

 > James Ingham <jingham@cygnus.com> writes:
 > > 
 > > Gdb uses a whole bunch of Cygwin functionality to do its job on
 > > Windows, and there are currently no plans to rewrite these sections to
 > > call directly into the Windows API.  That would be a pretty big job,
 > > with no clear payoff.  So, no, a Mingw32 version of Insight is not
 > > likely to happen.
 > > 
 > 
 > Actually, I've had a mingw port for a few years now. It requires a few
 > simple changes to gdb (win32-nat.c and the config files). The problem
 > is readline -- the older version had a "minimal" mode, but not available
 > anymore, and I have to essentially use a older version of readline to
 > get the job done. For mingw diffs to gdb-4.1[78], see:
 > 
 >   ftp://ftp.xraylith.wisc.edu/pub/khan/gnu-win32/mingw32/ports/
 > 
 > I haven't had the time to write up ChangeLog entries, but it's pretty
 > self-explanatory. When I get some time, I'll clean up the patches and
 > submit (and there's the assignment issue I keep on forgetting about). 
 > 
 > Unless the internals of gdb in insight is quite different and somehow
 > uses other Cygwin functionality, it should work.  Of course, lots of
 > mingw-specific tcl/tk changes are probably not in the Cygnus tree,
 > so that may involve some work.
 > 

This sounds cool.  How about all the socket and serial code handling
stuff needed for remote targets?  Do you get all that working as well,
or is this just for the native debugger?  I was under the impression
that Windows did not have a Posix'y interface for these functions.

Jim

-- 
++==++==++==++==++==++==++==++==++==++==++==++==++==++==++==++==++==++==++
Jim Ingham                                              jingham@cygnus.com
Cygnus Solutions Inc.