Yesterdays observation of lower left fringe

Yesterdays observation of lower left fringe

[Lennart Borgman (gmail)]

I am not quite comfortable with an observation I made yesterday. I have
set indicate-buffer-boundaries to top left and bottom left. The bottom
left used a right upper corner symbol yesterday.

Today it uses a lower left. I like that more.

But it is frustrating that I am using the same Emacs today as yesterday.
(But I have rebooted my w32 pc.)

Re: Yesterdays observation of lower left fringe

[David De La Harpe Golden]

Lennart Borgman (gmail) wrote:
> I am not quite comfortable with an observation I made yesterday. I have
> set indicate-buffer-boundaries to top left and bottom left. The bottom
> left used a right upper corner symbol yesterday.
> 
> Today it uses a lower left. I like that more.
> 
> But it is frustrating that I am using the same Emacs today as yesterday.
> (But I have rebooted my w32 pc.)
> 
> 

Um. Doesn't the bottom left only use the upper right corner symbol when
the last char in the file isn't a newline?  At least, that's the
behaviour on my fortnight-old emacs build (in the middle of apartment
move, kinda busy...).

i.e. it is (or at least was) actively indicating something.

Re: Yesterdays observation of lower left fringe

[Lennart Borgman (gmail)]

David De La Harpe Golden wrote:
> Lennart Borgman (gmail) wrote:
>> I am not quite comfortable with an observation I made yesterday. I have
>> set indicate-buffer-boundaries to top left and bottom left. The bottom
>> left used a right upper corner symbol yesterday.
>>
>> Today it uses a lower left. I like that more.
>>
>> But it is frustrating that I am using the same Emacs today as yesterday.
>> (But I have rebooted my w32 pc.)
>>
>>
> 
> Um. Doesn't the bottom left only use the upper right corner symbol when
> the last char in the file isn't a newline?  At least, that's the
> behaviour on my fortnight-old emacs build (in the middle of apartment
> move, kinda busy...).
> 
> i.e. it is (or at least was) actively indicating something.


Eh, you are right. But ... signs that do not talk should be crystal
clear ...

Isn't there someone who have a better suggestion for that
"there-is-no-ending-new-line" symbol?

Personally I would be less surprised by a lower right and a lower left
corner, but still confusing. Here is another idea:

If the last line has an ending newline char then place the fringe symbol
one line down (and let it be the same symbol). What about that?

Re: Yesterdays observation of lower left fringe

[David De La Harpe Golden]

Lennart Borgman (gmail) wrote:

> 
> If the last line has an ending newline char then place the fringe symbol
> one line down (and let it be the same symbol). What about that?
> 
> 
That would make the normal case where there is a terminating newline
present*, kind of ugly IMO.   I'd agree that the upper right symbol is
a bit wierd, but better use a different symbol when there's no final
newline, maybe something like |... , not move a line down IMO. i.e. a
dotted horizontal bar on the "L" because the line hasn't been terminated
yet.

* Technically, a bit like C vs. Pascal ';', most programs treat "\n" as
a line terminator, some as a line separator.   Having the fringe marker
move another line down for a present final newline would make more sense
if newline were considered a separator, as programs observing that
convention might think there was a new, empty line if the last line in
the file had a newline at the end.  But AFAIK considering it a
terminator is far more common than separator.

Re: Yesterdays observation of lower left fringe

[Lennart Borgman (gmail)]

David De La Harpe Golden wrote:
> Lennart Borgman (gmail) wrote:
> 
>> If the last line has an ending newline char then place the fringe symbol
>> one line down (and let it be the same symbol). What about that?
>>
>>
> That would make the normal case where there is a terminating newline
> present*, kind of ugly IMO.   I'd agree that the upper right symbol is
> a bit wierd, but better use a different symbol when there's no final
> newline, maybe something like |... , not move a line down IMO. i.e. a
> dotted horizontal bar on the "L" because the line hasn't been terminated
> yet.

If we think of the ending newline as terminating the line then maybe
using a lower right corner for that and a lower left corner for the
"unterminated" line would be "more mnemonic".

> * Technically, a bit like C vs. Pascal ';', most programs treat "\n" as
> a line terminator, some as a line separator.   Having the fringe marker
> move another line down for a present final newline would make more sense
> if newline were considered a separator, as programs observing that
> convention might think there was a new, empty line if the last line in
> the file had a newline at the end.  But AFAIK considering it a
> terminator is far more common than separator.
> 
> 
> 
> 

Re: Yesterdays observation of lower left fringe

[David De La Harpe Golden]

Lennart Borgman (gmail) wrote:


> If we think of the ending newline as terminating the line then maybe
> using a lower right corner for that and a lower left corner for the
> "unterminated" line would be "more mnemonic".
> 

You mean in different fringes for the two cases?  In one fringe, that
would be plain ugly, spoiling the nice bracketing "[" effect for the
common case.

Re: Yesterdays observation of lower left fringe

[Lennart Borgman (gmail)]

David De La Harpe Golden wrote:
> Lennart Borgman (gmail) wrote:
> 
> 
>> If we think of the ending newline as terminating the line then maybe
>> using a lower right corner for that and a lower left corner for the
>> "unterminated" line would be "more mnemonic".
>>
> 
> You mean in different fringes for the two cases?  In one fringe, that
> would be plain ugly, spoiling the nice bracketing "[" effect for the
> common case.


My sense for this is more logical consistency. I think the symbol should
be the same whether they are shown to the right or the left.

And the symbols that fits together best with the initial upper left
corner here are the lower corners, left and right.

I can agree that "[" is a nice visual effect. But logically if we want
to distinguish between terminated and unterminated lines it more feels
like lower right is "more" terminating than lower left.

But I would not mind much switching that so that left lower corner is
for the normal case and right lower corner is for the "missing" new line
case.

Re: Yesterdays observation of lower left fringe

[David De La Harpe Golden]

Lennart Borgman (gmail) wrote:

> I can agree that "[" is a nice visual effect. But logically if we want
> to distinguish between terminated and unterminated lines it more feels
> like lower right is "more" terminating than lower left.
> 

Relative, though - how about just a (thick/matching) "|" for
unterminated?. Then it's "completed" with the horizontal bar of the "L"
when there's a newline?  i.e. an "L" is in turn "more" terminating than
a "|" :-)

 _
|  hello
   this is
|  an example
==
==


 _
|  hello
   this is
|_ an example
==
==

Re: Yesterdays observation of lower left fringe

[Lennart Borgman (gmail)]

David De La Harpe Golden wrote:
> Lennart Borgman (gmail) wrote:
> 
>> I can agree that "[" is a nice visual effect. But logically if we want
>> to distinguish between terminated and unterminated lines it more feels
>> like lower right is "more" terminating than lower left.
>>
> 
> Relative, though - how about just a (thick/matching) "|" for
> unterminated?. Then it's "completed" with the horizontal bar of the "L"
> when there's a newline?  i.e. an "L" is in turn "more" terminating than
> a "|" :-)
> 
>  _
> |  hello
>    this is
> |  an example
> ==
> ==

I can see what you mean, but lower right seems more "balanced", more
related to me - and therefor easier to guess. You already have to know
what the symbol is for to guess that a "|" means not closed.

During this discussion I have come to think that guessing is more
important than knowing the exact meaning. I think I have almost never
worried about ending a file with a newline.

 _
|   hello
    this is
 _| an example
==
==


>  _
> |  hello
>    this is
> |_ an example
> ==
> ==