Re: [Bug-zile] Nested buffers

[Reuben Thomas]

On 5 February 2014 22:25, Gary V. Vaughan <address@hidden> wrote:

Me too!  I even tried to use Plan 9 (via http://swtch.com/plan9port/) for a while, because Acme does a much better job of being a terminal and a text editor at the same time.  It didn't really stick because Acme likes a 3 button mouse, and emulation with a track pad on a laptop is a bit awkward.

I read about it and loved it but thought I'd need a while to get used to an editor with so few built-in commands. I guess it must've influenced my thinking about Zee more than I thought, as I'd almost forgotten it.

 

> as does structured editing (which I've listed as an aim for Zee).

...as in XML?  Bleh.  Anything XML can do, Lisp and/or JSON can do better.

No, I just mean structured editing, i.e. an editor that is aware of hetereogeneous documents, so if you have one language embedded in another, or an image in a coment, or a terminal session in a web page, you get the appropriate nested editor at that point.

 

Are you thinking of the terminal program with mouse interaction with scroll back?

I don't think so, nor was I thinking of zooming interfaces (though I like that idea very much too).

 

Permanent undo is interesting too, and I've been reading about some clever data structures (Splay Ropes, and other functional tree structures) that provide a similar feature.  So much cool stuff to do, and so little time!! :)

Indeed. I'm in the middle of doing to FontForge what I did to C Zile, namely, removing all its memory management and replacing it with libgc. FontForge is a program I only use (I know nothing about programming fonts), but it's half a million lines of sheer misery (George Williams, the original and main author, now retired, spent 10 years working on it solidly, bless him, and didn't stop to learn how to program; he wrote an entire toolkit (Qt wasn't free and Gtk didn't do Unicode when he started) and more code that I'll probably write ever, but it is HIDEOUS to read), so I thought I'd try to give the current developers (the project has fortunately come back from the dead) some permanent leverage (I already went through and removed lots of dead code). I'm hoping to end up about 20kLOC down from where I started, which is only really scratching the surface, but never having to write another "free" or track another block of memory or debug another double-free or memory leak will, I hope, save them a lot of time.

--
http://rrt.sc3d.org