[Nano-devel] Re: bug in indent/unindent

[Benno Schulenberg]

David Lawrence Ramsey wrote:
> Benno Schulenberg wrote:
>  > At the moment, in CVS, M-[ indents and M-] unindents; to me
>  > the other way around would seem more logical.
>
> I have a good reason for the current order.  See below.

Below I haven't seen an explanation for this, or it should be that 
users are more likely to unindent than indent, and for this reason 
the unindent key is the undelayed one?

The reason for my preferring M-[ for unindent and M-] for indent is 
that M-( means to start and M-) means to end of paragraph, that is 
move far left and move far right; M-< means previous and M-> means 
next buffer, that is also kind of left and right, at least for 
people that read and write left to right.

>  > *** glibc detected *** realloc(): invalid size: 0x080b7018 ***
>  > Aborted
>
> I don't get this message, but, after looking into it, I noticed
> that the mark handling didn't account for an underflow when
> subtracting, so the beginning mark coordinate would end up far
> out of range.  I've fixed that problem in current CVS.  Could you
> try it and let me know whether this still happens?

It happens no more.  Looks like it's fixed.

>  > Also, when hitting M-[ say five times in a row, after a
>  > seconds pause it inserts four brackets ([[[[) at the cursor
>  > position.
>
> There's always a delay after pressing Meta-[, because many escape
> sequences start with Esc [.  This is hardcoded into curses, and
> there's no way to get rid of it without breaking escape sequence
> support.  This is why { and } are first, since they don't have
> the problem, even if you have to use the Shift key to get them. 
> Sorry for the inconvenience, but it's another terminal oddity...

Weird.  In my opinion this delay makes M-[ into an unusable key.  
Having to wait for it to take effect, without any indication that 
one has to wait...  Better use the ticks then: M-` and M-'.

Benno