This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[RFA/win32] Use the win32_ops a little less...


When I initially started the port to x86_64-windows, I thought that
I would need to move some of the gdbarch methods to CPU-specific files.
So the first thing I did was to try to implement target-ops inheritance
the same way we do it with ptrace, for instance.  This is to avoid
the use of a target_ops global across several units.

As it turned out, the code was generic enough that having CPU-specific
methods was not necessary.  The only piece that really depended on
the CPU is the "mappings" array.

However, I think that this is a valuable cleanup in itself, as it
gets rid of some unnecessary uses of a global variable with only
a very small number of easy changes.

2009-01-07  Joel Brobecker  <brobecker@adacore.com>

        * win32-nat.c (do_initial_win32_stuff): Add new ops parameter,
        and use it when pushing the target.
        (win32_attach, win32_create_inferior): Update call to
        do_initial_win32_stuff.
        (win32_detach, win32_mourn_inferior): Use our ops parameter
        instead of the global win32_ops to unpush the target.

As all patches that I'm about to submit for windows, I'm well setup
to run the AdaCore testsuite, which is somewhat less extensive than
the official testsuite. But I'm not very well setup to run the testsuite
as it keeps timing out for reasons I don't understand. So, only tested
using AdaCore's testsuite.

OK to apply?

Thanks,
-- 
Joel

Attachment: win32_ops.diff
Description: Text document


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]