Farid Zaripov wrote:
[...]
This is a bug, but it's not present in 4.1.3 release. The bug was
introduced
here: http://svn.apache.org/viewvc?view=rev&revision=380995
I see. My bad.
[...]
The simple test:
#include <sstream>
#include <cassert>
int main (int, char**)
{
std::stringbuf buf;
buf.str ("123");
assert (buf.str() == "123");
buf.str ("456");
assert (buf.str() == "456");
return 0;
}
The second assert fill fail in debug configurations because the buffer
filled by debug content
after __alloc.deallocate ().
Okay. Since it's not a bug in a released version of the library we
don't really need an issue for it. There does need to be a test to
exercise it, though (the yet-to-be-written 27.stringbuf.cpp).
So, please go ahead and commit your fix (thank you! :) separately
from all the other changes and I'll get started on the test.
I'll comment on the other issue(s) in a separate mail.
Martin
