This is the mail archive of the cygwin-developers mailing list for the Cygwin 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]

Re: [RFC] Cygwin libstdc++ plan (operator new/delete replacement)

On Fri, Oct 02, 2009 at 06:31:21AM +0100, Dave Korn wrote:
>Christopher Faylor wrote:
>>Why isn't this handled by change #1.362 to
>It is! I was using -61 instead of the snapshot.
>>That should always set cxx_malloc unconditionally (which, now that I
>>think of it, might not always be what you want either).
>Yeh, I'd like to replace these hacks with something that copies back
>the updated entries if it finds the malloc stuct pointer has been
>redirected, and then resets the pointer if it has been either nulled or
>redirected.  Also remove the hack in cygwin_external;

Actually, I take it back.  Setting this to cxx_malloc is fine since it
is the contents of that structure that are supposed to be changed, not
the pointer itself.  The cxx_malloc pointer should essentially be
read-only and should never change.

I don't see how you can remove the cygwin_external call as long as there
are dlls out there which reset this field in  You just
need one problem dll and this causes all sorts of bad problems when
loading other dlls.


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