vm team mailing list archive

[Uday Reddy <u.s.reddy@xxxxxxxxxxxxx>]

On 6/19/2010 6:26 AM, Tim X wrote:

> I looked at some of the warnings. Although I may be wrong, I think a few
> of them, especially the faces related ones and some of the ones with
> wrong number of arguments are due to differences between GNU Emacs and
> Xemacs. Not sure what we should do. Probably the most reliable method
> would be to either test for type of emacs and call whichever is
> appropriate. Another possiblity may be to break things up into xemacs
> and gnu emacs code files - this might be particularly useful for the
> faces stuff. While it has been sometime since I checked, there was
> considerable difference between emacs and xemcs in that area (in fact,
> emacspeak no longer works with xemacs largely because of this. Emacspeak
> uses face information to alter speech properties. A sort of voice
> locking, where different voices are used to indicate differences in text
> in a similar way to colors ini font-locking. Differences in how xemacs
> and emacs dealt with this area made it very difficult to maintain a
> version that would run on both).

What we generally do is to define functions with vm- prefix, and put in 
vm-misc.el the case-based implementations for different versions of Emacs.  If 
you look in lp:~reddyuday/vm/warnings, you will see a few definitions I have 
added to vm-misc.el.  There are others there already.

> The other thing I was wondering about is if there is any position on
> backwards compatibility? Many of the warnings are probably due to
> changes occuring in versions. While not a big issue yet, it is likely to
> become more so as time passes. This could be especially so with the
> improved support for multibyte characters and encodings etc. We probably
> should have some standard on this so that anyone who wants to contribute
> can make a call on whether they need to add code to support an older
> version or just ignore possible negative impact on an older version in
> exchange for cleaner and more maintainiable code for later versions. For
> example, I would think only supporting emacs 21 onwards would be
> reasonable for future versions. Anyone running emacs 20 or less (would
> anyone be still running emacs 19?) can always stick with earlier
> versions. Is this something we should canvas with the user community?

There is an older thread by Ulrich in the newsgroup, where we declared which 
version we will support for VM 8.2.0.  I think Gnu Emacs 21 won't be supported.

> BTW, did you get any response to your request for Xemacs user
> assistance? Would be really handy to have an xemacs user to at least
> test some of the work that will need to be done to cleanup some of those
> compiler warnings.

No, I haven't received any responses.  The XEmacs users are a quite bunch.

On the other hand, I have negotiated with Aidan Kehoe on the XEmacs team that 
he will maintain the XEmacs package provided we won't give him more than two 
releases a year to work with.  He is going to put in 8.1.1 right now.  So, we 
will have another chance when we want to add 8.2.x.

There are a few XEmacs users who do download the recent releases.  But it is 
hopeless to get them to do any work ahead of time.  However, after downloading 
a release, they will no doubt report problems they run into.

Cheers,
Uday