Hello fellow coders,

Version 0.2 of the packarray class is available at the same spot:

http://www.netdirect.ca/~cdfrey/software/packarray.cc

This version handles endian issues all at once, and converts double
values to network byte order as well. The basic_field<> class is gone,
as it's not needed any longer.

I ran into an interesting issue with regards to C++ aliasing rules, which
you can read about here:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25235

And here:

http://mail-index.netbsd.org/tech-ke...8/11/0001.html

This makes me rather concerned about the reinterpret_cast<> I have to do
in 4 places in packarray. Any tips on how to clean that up would be very
much appreciated.

Anyway, it works great for me, even with g++ 4.0.2. Testing on non-intel
machines would be great.

Enjoy,
- Chris


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

On Sat, Dec 03, 2005 at 05:32:41AM -0500, Chris Frey wrote:
In addition to the above, I started a thread in comp.lang.c++.moderated
which is turning out to be rather enlightening.

http://groups.google.ca/group/comp.l...14701b02a2a3cc

I hope that's a stable URL from Google. :-)

- Chris


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

Chris Frey wrote:
I just took a look at that, and saw that you seem to have worked out the
technical problems. Are we going to see a v0.3 soon, or were you
waiting on OS X tests with v0.2 from me first?

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

On Thu, Dec 15, 2005 at 07:25:30AM -0700, Warren Young wrote:
I've been sidetracked by flu. The main thing I still don't like about v0.2
is that it uses a reinterpret_cast<> in an inline template function to
provide the application access to double. I fear that could be optimized
away someday, so am pondering putting that string-to-double casting in
a .cpp file, which can be compiled with -fno-strict-aliasing in the library
for safety.

Also, James Kanze in the newsgroup recommended using frexp() and ldexp()
to do the double conversions portably, which would be nice in v0.3, but
I haven't done that yet. That's the implementation that should really be
tested on OS X when it's done.

Thanks,
- Chris


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

Chris Frey wrote:
Sorry to hear that.

Okay, sounds good. No rush from this quarter, just making sure you
weren't waiting on me.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw