This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [RFA/windows] cast of address to DWORD warning (handle_unload_dll)
On Wed, Jan 14, 2009 at 08:46:15AM +0400, Joel Brobecker wrote:
>> >2009-11-14 Joel Brobecker <brobecker@adacore.com>
>> >
>> > * windows-nat.c (handle_unload_dll): Use host_address to string
>> > in order to print the base address of the DLL that was unloaded.
>> >
>> >This one is hard to test, because it's supposed to never happen.
>> >So I tested it by, ahem, visual inspection (I learnt that expression
>> >at my first job, where we were producing safety critical software).
>>
>> I guess this is ok. I just changed other similar occurrences to %p.
>> Maybe all of those should also be changed for hobgoblinish consistency
>> or this one should be %p too.
>
>I don't mind changing the patch to using %p at all. The reason I avoided
>%p is because I was still in this frame of mind that %p depends on
>the implementation - to have consistent output, we have to be careful
>to not use %p in the core code. However, I just realized that it's not
>necessarily a problem to do so in nat files.
>
>I did double-check what %p does on Windows, though, and apparently it
>does not output the leading 0x (but that's only a detail), but also it
>prints the hexa digits in capital letters. This may be from not
>being used to seeing addresses printed with capital letters, but
>I find it harder to read. On the other hand, I have seen this style
>being used in stack dumps printed by Windows, so this may be the
>usual style on this OS.
Ugh. I didn't know Windows behaved that way. I use Cygwin's
implementation internally but I actually don't know what Cygwin does
with %p for the rest of the world and my PC is turned off right now.
Maybe it does the same thing.
I guess that's an argument to change this at some point but for now,
it is, of course, ok to check this in.
cgf