[cperl-mode] cperl-indent-for-comment

[cperl-mode] cperl-indent-for-comment

[LanX]

Hi

Two questions;

1) I wanna make # more DWIM and automatically doing an indent-for-
comment when at the end of the line. I thought about
  a) checking if the preceding character was a ";" or a paren.
and
  b) checking with describe-text-property that I'm not within a string
or regex.

is b) the recommended way to do this?

2) cperl-indent-for-comment produces weird results when started within
a Perl-string (M-;), it indents but repeated calls  produce a multiple
concatenated "# " at the end.

Is this the intended functionality???

Cheers
  Rolf

Re: cperl-indent-for-comment

[LanX]

Hi

this is my first approach to question 1...

Any suggestions?

-----------------------------------------------------------
(defun dwim-indent-for-comment (arg)
 "Indent comment after \; { ( [ otherwise insert #"
 (interactive "p")
 (if (and (memq (preceding-char)  '( ?\; ?{ ?\( ?\[ ) )
	   (not   (get-text-property (point) 'face))
	  )
		(indent-for-comment)
		(insert "#"))
  )

(local-set-key (kbd "#") 'dwim-indent-for-comment)
-----------------------------------------------------------

Re: [cperl-mode] cperl-indent-for-comment

[Ilya Zakharevich]

On 2010-05-05, LanX <lanx.perl@googlemail.com> wrote:
> 1) I wanna make # more DWIM and automatically doing an indent-for-
> comment when at the end of the line. I thought about
>   a) checking if the preceding character was a ";" or a paren.

Why not do it more agressively, and just avoid it after $ ?

> and
>   b) checking with describe-text-property that I'm not within a string
> or regex.

Keep in mind that one needs to force update of syntaxification before
checking text-properties.

        (cperl-update-syntaxification (point) (point))

> 2) cperl-indent-for-comment produces weird results when started within
> a Perl-string (M-;), it indents but repeated calls  produce a multiple
> concatenated "# " at the end.

Not in my version.

Yours,
Ilya

Re: cperl-indent-for-comment

[LanX]

Hi

> Why not do it more agressively, and just avoid it after $ ?

hmm ... technically I was worrying about #-delimiters like in
regexes ... but testing the face should already be sufficient.

I th this DWIM behavior should be restricted to the same end-of-line
scenarios like auto-newline.

If I need it explicitly from within the code I can use C-#.

But indeed thats a matter of taste and could be made customizable ...

> Keep in mind that one needs to force update of syntaxification before
> checking text-properties.
>
>         (cperl-update-syntaxification (point) (point))

Good point 8)

-------------------------------------- Version 0.02
(defun my-indent-for-comment (arg)
  "dwim indent-for-comment"
  (interactive "p")
  (cperl-update-syntaxification (point) (point))
  (if (and (memq (preceding-char)  '( ?\; ?{ ?\( ?\[ ) )
	   (not   (get-text-property (point) 'face)))
      (indent-for-comment)
    (insert "#"))
  )

(local-set-key (kbd "#") 'my-indent-for-comment)
----------------------------------------

TODO:
- checking for after code (whitespace till end of line)
- adding this case to cperl-electric-backspace
- handling the case of auto-newline.


> Not in my version.

your still running emacs 19 right?

For completeness:

putting the cursor right after the doublequote and pressing M-;
results in this
------------------------------------
$a="				# #
------------------------------------

every new try appends a new " #"

I started "emacs22 -q" and emacs23 and tested it with cperl "5.23 and
"6.2". (perl-mode is OK)

But this is not a really important bug...

Re: [cperl-mode] cperl-indent-for-comment

[Stefan Monnier]

>   b) checking with describe-text-property that I'm not within a string
> or regex.

You might want to try (nth 8 (syntax-ppss)) for that.


        Stefan

Re: cperl-indent-for-comment

[Ilya Zakharevich]

On 2010-05-06, LanX <lanx.perl@googlemail.com> wrote:
>> Not in my version.

> your still running emacs 19 right?

I have access to computers with v19, v20, and v21 around.  I had to
one with v22, but it lost libpng (or somesuch) and Emacs won't load
anymore.  Now let me recheck...  AHA!  That guy was updated, and now
has v23 - and it may actually be loaded!

Now I can start real testing...

> For completeness:
>
> putting the cursor right after the doublequote and pressing M-;
> results in this
> ------------------------------------
> $a="				# #
> ------------------------------------
>
> every new try appends a new " #"
>
> I started "emacs22 -q" and emacs23 and tested it with cperl "5.23 and
> "6.2". (perl-mode is OK)
>
> But this is not a really important bug...

For me it is.  I get quite used to ALL my stuff working 100% of
time...  (Or at least being quickly enough upgraded after problems are
discovered.)  ( -  Well, of course it is a pipe dream with scores of my
stuff on CPAN...)

Yours,
Ilya

Re: cperl-indent-for-comment

[LanX]

Hi

I posted the the code on perlmonks:

http://perlmonks.org/index.pl?node_id=839203

I had to extend it to assure that there is no syntax-type at point

       (not (get-text-property (point) 'syntax-type))

I'm wondering now if checking

       (memq  (get-text-property (point) 'face)
	      '(cperl-my-trailing-spaces-face nil))

is redundant now, for what I see all other cperl-faces only occur when
syntax-type is not set! (?)

Cheers
  Rolf

Re: cperl-indent-for-comment

[Ilya Zakharevich]

On 2010-05-10, LanX <lanx.perl@googlemail.com> wrote:
> I had to extend it to assure that there is no syntax-type at point
>
>        (not (get-text-property (point) 'syntax-type))

BTW, there is a sketch of documentation before the code of
`cperl-unwind-to-safe'.

Hope this helps,
Ilya

[cperl] electric-backspace and variable-refactoring

[LanX]

> BTW, there is a sketch of documentation before the code of
> `cperl-unwind-to-safe'.

Thanks a lot Ilya, I expanded now electric-backspace with a wrapper to
handle indented-comments.

see  http://perlmonks.org/?node_id=839593

BTW: I'm trying to realize variable refactoring by searching for the
last my declaration of a given variable.

I noticed that %a and @a are handled differently from $a

(print (sexp-at-point)) will return "$a" but only "a" in the other
cases. Same problem with (backward-sexp)

Cheers
  Rolf
I need this to check if a variable is preceded by a "my" or "our"
statement.

Any suggestions?