[galib] galib development status

gio sarto teofilo68 at yahoo.co.uk
Mon Dec 27 15:42:34 EST 2004


Hi Matthew

I am glad to hear there will be a new release of
GAlib, I have been using it for about  6 months and I
am very much satisfied. 
Concerning bug fixes/improvements here are my
 suggestions:
1) a (I think bug) I documented in 
http://mailman.mit.edu/pipermail/galib/2004-August/001037.html

2) Port everything from float to double, the memory
increase is negligible with current hardware, and it
is straightforward, I have done it myself and it is
almost nothing more than replacing all occurrences of
float with double. But I have not exetensively tested
yet

3) The definitions of GAMin/GAMax in terms of macros
are dangerous because   
  a) if one or both entries are experssions they imply
  recalculation
  b) for the same reason they can have unpleasant side
effects 

 I would suggest to use the STL min(...) and
max(...)templatised functions (probably best choice
for standardisation and portability), or write GAlib's
own templetised
version ( as I did because I was in a hurry and did
not
want to go into the troubles of having STL link
correctly when I link GAlib into my own libs and
application) 

Cheers

Giovanni Sartoni

 --- matthew wall <mwall at oculustech.com> wrote: 
> On 25 Dec 2004, at 10:53, Jonas Karlsson wrote:
> 
> > GALib doesn't compile with the latest versions of
> gcc or MS visual
> > studio.    Are there any plans for a new release,
> or is galib end of
> > life?
> 
> 
> hello everyone,
> 
> recently there have been some questions about the
> state of galib.
> 
> 
> gcc3 and ms vs7 compatibility:
> 
> i have recently installed the latest ms compiler and
> i will soon have 
> the latest ms visual studio.  so i will finally have
> a proper windows 
> environment in which to build and test galib.
> 
> 
> next minor release:
> 
> i plan to release galib 2.4.6 with the following
> fixes:
>   - clean compile with gcc3
>   - clean compile with visual studio 7
>   - incorporation of various bug-fixes
> so this will be a minor, maintenance release.
> 
> i was hoping to have it out in the fall of 2004, but
> various personal 
> incidents and the lack of the windows build/test
> environment have 
> hampered that.
> 
> 
> next major release:
> 
> i have been working on a major galib release.  it
> will probably be 
> called galib 3.0 since it will require not only a
> recompile but also 
> (in some cases) modifications to the code that uses
> galib.
> 
> this will include the following:
>   - saving the state of an evolution to xml (for all
> built-in genome 
> types)
>   - pause/restart of an evolution for any built-in
> types.
>   - removal of the pseudo rtti stuff, and removal of
> other hacks to
>       accommodate ancient compilers.
>   - incorporation of baseline implementations of
> various 'standard'
>       evolutionary algorithm 'best practices' that
> use the components in
>       galib.  in some cases these will
>   - refactoring/rewrite of the examples.
>   - incorporation of a unit testing framework
> similar to junit.
> please let me know if there are specific features
> you would like to see 
> in the release.
> 
> i am sorry, but there is not yet an eta for this
> major release.
> 
> 
> public access to the galib cvs repository:
> 
> the hardware for lancet.mit.edu will soon be
> upgraded.  when this 
> happens, the galib cvs repository will be made
> available so that you 
> can get the latest snapshots and so that we can have
> easier direct 
> contribution to galib.
> 
> 
> if anyone has suggestions or other feedback, please
> let me know.
> 
> matthew
> 
> _______________________________________________
> galib mailing list
> galib at mit.edu
> http://mailman.mit.edu/mailman/listinfo/galib
>  



	
	
		
___________________________________________________________ 
ALL-NEW Yahoo! Messenger - all new features - even more fun! http://uk.messenger.yahoo.com


More information about the galib mailing list