[GiNaC-devel] GiNaC 1.3.3 release

[Richard B. Kreckel]

Dear Chris,

Chris Dams wrote:

>Dear developers,
>
>On Mon, 24 Oct 2005, Richard B. Kreckel wrote:
>
>  
>
>>GiNaC 1.3.3 is out and available. The changes are:
>>    
>>
>
>Is the GiNaC development team making a habit out of ignoring patches?
>
Not really.  Rather, we're lucky that we got the thing out the door at 
last.  It's turned out more difficult than it used to.  Sorry if we 
dropped a patch, but given your recent flash of activity it was not 
crystal clear which one was ready and supposed to enter 1.3.

>I
>realize that because of adding a new overloaded version of
>rename_dummy_indices_uniquely this probably can't go into 1.3.3 but the
>separate symmetrization of different index types could have gone into it,
>I think. Also, I would appreciate it if changes were added to the
>1.4.0-branch. This spares me the trouble of managing a locally changed
>version of GiNaC.
>
>Here is the latest diff of my local copy of GiNaC with the CVS tree. New 
>changes are:
>(1) Made an automatic symplification that turns an indexed object without 
>indices into its op(0). This is because it is sometimes useful to make an 
>indexed object out of an exvector and an indexed object without indices is 
>quite silly, really.
>
>(2) Made the return_type and return_type_tinfo of an indexed object return
>these values as they are of op(0). Reason: if the user wants to define his
>own noncommutative functions/symbols and wants them to take indices, that
>user probably wants these indexed objects to be noncummutative.
>  
>
I don't know how Jens managed patches, but I find cumulative patches are 
confusing.  Would it be possible for you to separate those patches and 
make it clear which ones don't break binary compatibility and are for 
1.3 and which ones are for HEAD?

Jens?

  -richy.

-- 
Richard B. Kreckel
<http://www.ginac.de/~kreckel/>