Re: elastic tabstops implemented for GTK

["Nick Gravgaard" <me nickgravgaard com>]

On 29/08/2007, Alex Jones <alex weej com> wrote:
>
>  On Wed, 2007-08-29 at 12:06 +0100, Gustavo J. A. M. Carneiro wrote:
>  And how will this look e.g. to Python, which generally doesn't even like
> tabs?  Variable width tabs will likely cause havoc in Python programs...
>
>  As Nick said, Python is just fine with tabs!
>  Now, if you did the exact same thing, but behind the scenes inserted
> spaces instead of tab characters, then it would be nice.
>
>  His GEdit plugin supports this, but it's a lossy conversion. When reading
> back the 0x20s, you have to try to figure out what was supposed to be a tab,
> and I'm sure you might get some false positives eventually, unless there's
> something I'm missing.

No, you're exactly correct. If we look at the following text (in a
monospaced font):

a       zzzz
aa      zzzz
aaa     zzzz
aaaa    zzzz
aaaaa   zzzz
aaaaaa  zzzz
aaaaaaa zzzz

I have no (easy) way of determining if the one space on the last line
is being used for alignment or not. The plugin I've made for gedit
which converts to and from the different formats understands this
issue - a tab is never converted into just one space (it will shift
the text in that column to the next multiple of N instead), and when
converting spaces->tabs (using elastic tabstops) it assumes that any
run of more than one space is being used for alignment purposes. It
will be an issue for text that's converted using other tools though,
but at the moment there's not a lot I can do about it. In practice it
seems to work pretty well.

>  I really can't wait to get this functionality, it finally puts an end to
> all of the nonsense over how many spaces there should be in a tab!

Thanks - I'm working on it! :)