From: Paul K. <pki...@ja...> - 2002-09-20 15:51:05
|
On Fri, Sep 20, 2002 at 04:43:53PM +0100, Etienne Grossmann wrote: > > Hi all > > On Thu, Sep 19, 2002 at 11:37:56PM -0400, Paul Kienzle wrote: > # Hi all, > # > [snip] > # > # Off the top of my head, struct/*.cc and FIXES/fsolve.cc are broken. A more > ^^ > In 2.1.37, you mean? Perhaps. I've built from CVS so it's not clear what version I'm working with. > > # > # Unfortunately m-files will also have to change: > # * replace ... with varargin/varargout > # * replace all_va_args with varargin{:} > # * for [v,k]=struct ... end now sets v to a list rather than a value > # * anything else? > # > # We could use version tests in our code. > > ^^ > Can that do 'function foo (x,...)' and 'function foo (x,varargin)' Good point. Preprocessor it is. [snip] > # Anything else? > > What about supporting 2.1.36, and the day 2.1.37 comes out, we fix > things (duration : a few weeks) and release a new octave-forge when > it's done. No worries until 2.1.37, none one month after it comes > either. Or am I over-optimistic? Okay, so 2.1.37 is not a show stopper for this release, but 2.1.36 issues are. No reason not to try supporting 2.1.37 if like me you are building from CVS since it will have to be done eventually. - Paul |