|
From: Nicholas N. <nj...@cs...> - 2005-05-10 13:20:16
|
[cc'ing list] On Tue, 10 May 2005, Julian Seward wrote: >> Vex pollutes the global symbol namespace quite a bit, as this shows: >> >> nm -g coregrind/stage2 | grep -v vg | more > > Wow, it does, doesn't it. A lot. > > I guess .. this is not such a problem as it used to be, since > we're no longer sharing the ld.so with the client -- now we have > our own ld.so instance. And so there's no need to avoid collisions > between stage2 and the guest. > > At least, according to my hazy understanding that's so. > > If that wasn't so, wouldn't all hell break loose when 2.4.0 > runs itself -- since the different V instances would conflict? I guess so. Can someone else confirm... Jeremy? N |
|
From: Jeremy F. <je...@go...> - 2005-05-10 18:46:10
|
Nicholas Nethercote wrote:
>> If that wasn't so, wouldn't all hell break loose when 2.4.0
>> runs itself -- since the different V instances would conflict?
>
>
> I guess so. Can someone else confirm... Jeremy?
Yep, there's no possibility of collision. The VG_(*) prefix stuff is
obsolete from that perspective.
J
|