Nicole Willson wrote:
> Actually the message is the header guard corruption message.  For some
> reason hdr->self_ is 0xffffffff on the second time through the for loop.

That's definitely not right. I would suggest setting a breakpoint
on the variable to see when it changes from the valid value first
set in operator_new.

Martin

> 
> 
> -----Original Message-----
> From: Martin Sebor [mailto:sebor@roguewave.com] 
> Sent: Tuesday, July 25, 2006 12:45 PM
> To: stdcxx-dev@incubator.apache.org
> Subject: Re: problem in temp_buffer
> 
> Nicole Willson wrote:
> 
>>The stack trace is:
>>#0  0x0feb0ca8 in raise () from /lib/tls/libc.so.6
>>#1  0x0feb26c4 in abort () from /lib/tls/libc.so.6
>>#2  0x1000eaa4 in find_block (ptr=0x77e00d48, check_heap=true,
>>caller=0x0)
>>    at ../new.cpp:142
> 
> 
> This looks like it's coming from the heap debugging library and could be
> caused by a corrupted header guard (see below). Is there any error
> message? (If not, you might need to flush stderr in the function to see
> it).
> 
> $ svn cat
> http://svn.apache.org/repos/asf/incubator/stdcxx/trunk/tests/src/new.cpp
> 
> | cat -n | head -150 | tail -15
>     136
>     137          if (check_heap) {
>     138              // check the lead guard
>     139              if (hdr->self_ != hdr) {
>     140                  rw_error (0, 0, __LINE__,
>     141                            "%s:%d: header guard corruption at
> %#p: "
>     142                            "expected %#p, got %#p",
>     143                            hdr->ptr_, (const void*)hdr,
> hdr->self_);
>     144                  abort ();
>     145              }
>     146
>     147              // check that the stored pointer has the expected
> value
>     148              if (hdr->ptr_ != hdr + 1) {
>     149                  rw_error (0, 0, __LINE__,
>     150                            "%s:%d: block address mismatch: "
> 
> Martin