Re: compile problems with 2.9.6

[Rafael Laboissiere]

* John W. Eaton <address@hidden> [2006-06-12 11:57]:

> On 12-Jun-2006, Rafael Laboissiere wrote:
> 
> | > And upstream is also informed, although they consider the bug as invalid:
> | > 
> | >     http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27935
> 
> OK, now I'm confused.  The bug report is invalid, but the bug was
> fixed in a later version?  If the bug report was invalid, does that
> mean that the example code in the bug report was invalid?  If so, then
> what is the proper fix so that it is valid code?  Also, for Octave,
> precisely what is the bug and the fix, if it really is a bug in
> Octave?

No, the bug is in the Debian package.  Apparently, a patch was applied that
confused the g++ compiler when size_t is present in the operator delete
argument list.  Here is the relevant part of the Debian changelog:


gcc-4.1 (4.1.1-3) UNRELEASED; urgency=low

  * Update to SVN 20060608, do not revert the fix for PR c++/26068.
    Closes: #372152, #372559.
  [...]
  
 -- Matthias Klose <address@hidden>  Tue,  6 Jun 2006 08:06:33 +0000
       
 
Matthias Klose told me that version 4.1.1-3 will hopefully be uploaded
today. 

-- 
Rafael