Re: locale support patch for [fs]sscanf function

[CdeMills]

John W. Eaton wrote
> 
> 
> There is a prevailing style for the ChangeLog entry part of a commit
> message.  Please use it, as it makes finding what was changed somewhat
> easier.  There's also usually no need to go into great detail in a
> ChangeLog entry.  For example, instead of the above, write a brief
> one-line summary for Mercurial followed by the ChangeLog info like
> this (but not indented in the actual commit message):
> 
>   handle locale [fs]scanf functions
> 
>   * src/oct-stream.h (octave_base_stream::imbue): New virtual function.
>   (octave_stream::imbue): New function.
>   * src/file_io.cc (Ffscanf): Handle locale, warning if requested locale
>   is not available.  Update docstring. 
>   (Fscanf) Likewise.  Update docstring.
>   New tests.
> 
>   The "F" prefix is used for DEFUN functions because "F" is prepended to
>   the function names by the DEFUN macro.
> 
> 
> The "undefined-return-values" warning ID is currently used for
> functions that do things like this:
> 
>   function [a, b, c] = f ()
>     a = 1;
>     c = 2;
>   end
> 
>   [a,b,c] = f
>   warning: f: some elements in list of return values are undefined
>   a =  1
>   b = [](0x0)
>   c =  2
> 
> so I don't see how it is appropriate for this case.
> 
> jwe
> 


The idea is to be able to transform the warning into an error to catch it
during the 'make check' phase. To this end, an Id is required. I agree that
"undefined return value" is not very appropriate, but other existing Id have
lesser relevance with the case. Could you suggest or create something
appropriate ?

Regards

Pascal

--
View this message in context: 
http://octave.1599824.n4.nabble.com/locale-support-patch-for-fs-sscanf-function-tp4303471p4310669.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.