Re: visual-line-mode

[David Reitter]

On 5 Jul 2008, at 19:41, Chong Yidong wrote:
> Good, I've checked it in (together with a couple other minor fixes).

Can you post a patch for these minor fixes?
I'm running this on 22 (because the trunk does not run on my system).

At least in the previous version, there's also a problem with posn-x- 
y, which returns a position too far when given coordinates that are to  
the right of a word-wrapped line (i.e. in the blank space).  This can  
also be seen if you just click into that space (mouse-set-point).

I can post a detailed recipe after I've made sure I've got the up-to- 
date patch.  My new code isn't iterative.  It's faster and shouldn't  
lead to excessive garbage collections as before.

> One thing that's not clear to me is whether your minor mode correctly
> handles invisible/intangible text and point motion hooks.  One of the
> reasons line-move-1 and line-move-finish is so complicated is,
> unfortunately, the need to DTRT for these.  I don't see anything
> analogous in your code; am I missing something?


OK, I'll bind `inhibit-point-motion-hooks' except for the last  
movement command.

Regarding invisible text, the original code went a long way to handle  
this efficiently.  With the new method (I'll post some code soon),  
this is handled elegantly and without any hassle thanks to posn-x-y.   
Regarding intangible text, this works fine with my working version.
I'm not sure about "fields", perhaps because I don't know much about  
their implementation.  Can you point me to a test case?  Customize  
buffers perhaps?

Thanks for pointing out these issues.

smime.p7s
Description: S/MIME cryptographic signature