[reSIProcate] String corruption fix

[david Butcher]

Thanks for the fix -- I modified the code a bit more. Instead of reallocating
inline, I called resize for the Share case. resize correctly adds the 1 for
null terminating.

Note that new char[] is called in only Data constructor and resize.
Similarly, delete[] is called in only ~Data and resize.

david


Quoting Christian_Gavin at logitech.com:

> Hi,
>
> Here is the fix I made locally to data.cxx
>
> Data&
> Data::append(const char* str, size_type len)
> {
>   assert(str);
>   if (mCapacity < mSize + len)
>   {
>      // .dlb. pad for future growth?
>      resize(((mSize + len +16)*3)/2, true);
>   }
>   else
>   {
>      if (mMine == Share)
>      {
>         char *oldBuf = mBuf;
>         mCapacity = mSize + len;
>         mBuf = new char[mSize + len + 1];      // << FIX here: added +1 to
> account for extra '\0' !
>         memcpy(mBuf, oldBuf, mSize);
>         mMine = Take;
>      }
>   }
>
>   // could conceivably overlap
>   memmove(mBuf + mSize, str, len);
>   mSize += len;
>   mBuf[mSize] = 0;
>
>   return *this;
> }
>
> It solved the memory corruption problem when appending strings.
>
> Please consider applying this fix to the subversion repository.
>
> Thanks,
> CG
>
> _______________________________________________
> resiprocate-devel mailing list
> resiprocate-devel at list.sipfoundry.org
> https://list.sipfoundry.org/mailman/listinfo/resiprocate-devel
>