Re: How to test if the current line contains only white-spache?

[Marcin Borkowski <mbork@mbork.pl>]

On 2015-11-20, at 00:58, Emanuel Berg <embe8573@student.uu.se> wrote:

> Marcin Borkowski <mbork@mbork.pl> writes:
>
>> On 2015-11-19, at 14:54, Rolf Ade
>> <rolf@pointsman.de>
>> wrote:
>>
>>> Yes, this is, what my instinct also telling me.
>>> But this instinct was trained by other programming
>>> languages and environments. I'm aware, that it may
>>> be not applicable to emacs lisp / emacs.
>>
>> And you're right: it's not.
>>
>>> Discussing programming concepts at such 'meta level'
>>> is difficult. But maybe some may express their
>>> thinking about this. Is this 'moving point around'
>>> in emacs lisp programming code the "right thing" to
>>> do [...]
>>
>> Yes it is.
>
> I don't think you can be categorical about it.

Of course I can.  Though maybe I shouldn't.  ;-)

But I don't say that this is a general rule, always applicable.  Just
that it is a fine pattern in Elisp coding.

> `save-excursion' and `save-window-excursion' are there
> because they are useful. But just because they are
> useful doesn't mean they should always be used.
>
> As for "side-effects" that is the phantom menace the
> Haskell buffs came up with, when actually side effects
> are the reason we use computers.
>
> Rather, in this case it is clear which is preferable
> of
>
>     (save-excursion (beginning-of-line) (point))
>
> and
>
>     (line-beginning-position)
>
> The code is easier and faster to write, read,
> and understand.

Is it really clear?  I'm not sure.

Besides, I've just checked the /implementation/ of
`line-beginning-position', and I have some bad news for you.  I'm not
fluent in C, nut AFAICT, it's even uglier than the Elisp variant: there
is no `save-excursion', just a temporary variable for the original
position of point (which is how `save-excursion' probably has to work
anyway).

So, how is the C `line-beginning-position' better than

(defun elisp-line-beginning-position ()
  (save-excursion (beginning-of-line) (point)))?

(apart from being less general)?  It is just /faster/, but not more
/elegant/ or /pure/.

> This is even more true when the example isn't as basic
> as that, but when everything appears in nested forms
> and loops and what have you. If you rely on moving
> point around, this will demand of you to be a better
> programmer than you would have to be, less it will get
> out of hands pretty quickly.

Aren't /abstractions/ (like functions and/or macros) there to help me
with that?

> That is one aspect. The other aspect is that often
> moving point to the correct place to extract some data
> seems straightforward in the lab. In the field all
> sorts of unexpected situations appear and suddenly
> point doesn't land where it should (or remotely so!).
> Now the code once again can quickly get out of hands
> if you try to compensate if `looking-at' something
> `forward-char' else - and so on: not good!

Could you give a concrete example where your technique is superior
because of that?

> Note that save-excursion isn't only used to get data
> but perhaps most often to seemingly reset the state
> after something is done so tho the user is aware it
> happened (because he struck the command) he won't be
> bothered with the buffer flicking by in the process.

Of course, many commands (like `mark-word', for instance), are
/supposed/ to do something without moving the point.

Best,

-- 
Marcin Borkowski
http://octd.wmi.amu.edu.pl/en/Marcin_Borkowski
Faculty of Mathematics and Computer Science
Adam Mickiewicz University