RE: Roadmap for 0.99-17

[SANGOI DINO LEONARDO <SANGOID@xxxxxxxxxxxxxxxxx>]

> -----Original Message-----
> From: John E. Davis
> Sent: lunedì 12 settembre 2005 16.51
> To: jed-users@xxxxxxxxxxx
> Subject: Re: Roadmap for 0.99-17
> 
[SNIP]
>    1.  How well is unicode (UTF-8) handled?  Unfortunately, I am
>        unable to answer this question since everything I edit is 7-bit
>        ASCII.  Does anyone on this list use jed in UTF-8 mode?

Well, I use to edit some UTF-8 encoded files with Jed... but those are
mostly Italian text, so it surely does not use a lot of UTF-8...

But I can reverse the question somehow: Are we sure that using Jed B0.99-17
*NOT* in UTF-8 mode works well as before? this means both ASCII only and
national character sets. If the answer is "yes", well, I think that jed is
mostly ready for a new "stable" release, and let fix the UTF-8 problems that
may appear after.
(By using Jed and WJed snapshots daily, I feel very confident that there are
no big regressions...)

>    2.  Some sort of iconv support would be nice to permit one to edit
>        a file encoded with a national character set while the editor
>        is running in unicode mode.   This would be accomplished by
>        having iconv convert the text in the national character-set to
>        and from UTF-8.

Do you remember that I did an iconv-module for slang? my latest version (I
don't remember if I posted it on this ML) works both on windows and Linux
(using the glibc iconv support, but should work well with and extenal iconv
library).
I did also a slang mode for jed to allow editing files with an on-disk
character set with an UTF-8 version of jed... and I use it to edit windows
registry files (UTF-16 encoded).

I think that maybe you should consider adding a native SLang interface to
iconv, instead of a module. With an option on sl-feat.h we can choose to
have a completely stub iconv implementation, some simple, harcoded
conversion (e.g. UTF-8 <-> iso-8859-15 and a few others) of a full,
iconv-based version.

>    3.  The issue of the conditional loading of jed.conf.  A
>        commandline switch has been proposed but I simply am unable to
>        see why this is necessary.  Is there a concrete example that
>        illustrates why the loading of jed.conf should be conditional?

This was never a problem for me, and I used to do a lot of batch editing
with jed... but maybe I simply never hit the problem...

> I imagine that adding unicode font support to Xjed will not be
> difficult, but will take some time to research the appropriate X11
> interface.  Hopefully I will find time to do this in the next few
> weeks and then make another 0.99-17 snapshot available.

I tried a hackish version of this... it's easy if you have a xfree/xorg
Xlib: those have Xutf8* functions... I don't know other X implementations...
We can do three things:
1) support xjed UTF-8 mode only when compiling on XFree/Xorg (and maybe
other X11 implementations that have Xutf8* funtions).
2) Always convert the UTF-8 text to the national character set used by X,
and draw text using it.
   (btw this can also be done for plain jed running on a non-utf8 terminal).
3) use 1) if available, else use 2).

Note that option 2) may be very difficult to implement (this means also
converting input data), and surely some characters could not be converted
from UTF-8 -> national character. 

Anyways I will try to find my UTF-8 Xjed version, and post it here tomorrow.

Thanks,
							Dino