`character-fold-search' should be a user option, like `case-fold-search'

`character-fold-search' should be a user option, like `case-fold-search'

[Drew Adams]

Subject line says it all.  Users should be able to customize
this variable, to decide for themselves, without fiddling with
`setq', whether search starts by default with or without
character folding.  They should be able to find this easily,
as an option in custom group `matching', right alongside
`case-fold-search'.

No?

I customized `case-fold-search' to nil long ago, and I will
likely do the same for `char-fold-search': set it to nil.
By default I want search to distinguish different characters.

Different users have different uses and different preferences.

Fortunately, the behavior can be toggled.  But the default
behavior should also be customizable.  Users should not need
to use `setq' for `char-fold-search' but be able to customize
`case-fold-search'.  This is a (minor) step backward, IMO.

Re: `character-fold-search' should be a user option, like `case-fold-search'

[Juri Linkov]

> Subject line says it all.  Users should be able to customize
> this variable, to decide for themselves, without fiddling with
> `setq', whether search starts by default with or without
> character folding.  They should be able to find this easily,
> as an option in custom group `matching', right alongside
> `case-fold-search'.
>
> No?
>
> I customized `case-fold-search' to nil long ago, and I will
> likely do the same for `char-fold-search': set it to nil.
> By default I want search to distinguish different characters.
>
> Different users have different uses and different preferences.
>
> Fortunately, the behavior can be toggled.  But the default
> behavior should also be customizable.  Users should not need
> to use `setq' for `char-fold-search' but be able to customize
> `case-fold-search'.  This is a (minor) step backward, IMO.

Like with ‘case-fold-search’ and ‘isearch-case-fold-search’ we could
make ‘char-fold-search’ customizable and ‘isearch-char-fold-search’
togglable, i.e. when ‘char-fold-search’ is t:

‘C-s a’     matches [aä]
‘C-s a M-'’ matches [a]
‘C-s ä’     matches [ä]
‘C-s ä M-'’ matches [aä]

when ‘char-fold-search’ is nil:

‘C-s a’     matches [a]
‘C-s ä’     matches [ä]

RE: `character-fold-search' should be a user option, like `case-fold-search'

[Drew Adams]

> > Subject line says it all.  Users should be able to customize
> > this variable, to decide for themselves, without fiddling with
> > `setq', whether search starts by default with or without
> > character folding.  They should be able to find this easily,
> > as an option in custom group `matching', right alongside
> > `case-fold-search'.
> >
> > No?
> >
> > I customized `case-fold-search' to nil long ago, and I will
> > likely do the same for `char-fold-search': set it to nil.
> > By default I want search to distinguish different characters.
> >
> > Different users have different uses and different preferences.
> >
> > Fortunately, the behavior can be toggled.  But the default
> > behavior should also be customizable.  Users should not need
> > to use `setq' for `char-fold-search' but be able to customize
> > `case-fold-search'.  This is a (minor) step backward, IMO.
> 
> Like with ‘case-fold-search’ and ‘isearch-case-fold-search’ we could
> make ‘char-fold-search’ customizable and ‘isearch-char-fold-search’
> togglable, i.e. when ‘char-fold-search’ is t:
> 
> ‘C-s a’     matches [aä]
> ‘C-s a M-'’ matches [a]
> ‘C-s ä’     matches [ä]
> ‘C-s ä M-'’ matches [aä]
> 
> when ‘char-fold-search’ is nil:
> 
> ‘C-s a’     matches [a]
> ‘C-s ä’     matches [ä]

Read this quick, but I think you are saying the same thing
you said in your earlier post - with which I agree.

Re: `character-fold-search' should be a user option, like `case-fold-search'

[Artur Malabarba]

2015-09-01 16:49 GMT+01:00 Drew Adams <drew.adams@oracle.com>:
> Subject line says it all.  Users should be able to customize
> this variable, to decide for themselves, without fiddling with
> `setq', whether search starts by default with or without
> character folding.  They should be able to find this easily,
> as an option in custom group `matching', right alongside
> `case-fold-search'.
>
> No?

Yes, it should be.
I didn't make it a defcustom initially because I'm not sure how
defcustoms interact with autoloading (I know it's bad in external
packages, but I have no idea if it's ok in the emacs core).

Re: `character-fold-search' should be a user option, like `case-fold-search'

[Artur Malabarba]

> Like with ‘case-fold-search’ and ‘isearch-case-fold-search’ we could
> make ‘char-fold-search’ customizable

Yes.

> and ‘isearch-char-fold-search’ togglable

I'm not sure I understand. Character folding is already toggleable
during isearch.
Do you mean we specifically need a variable called
isearch-char-fold-search? isearch-case-fold-search is an internal
variable, not a user option, so I don't think there needs to be a
char-fold analogue.

RE: `character-fold-search' should be a user option, like `case-fold-search'

[Drew Adams]

> > Subject line says it all.  Users should be able to customize
> > this variable, to decide for themselves, without fiddling with
> > `setq', whether search starts by default with or without
> > character folding.  They should be able to find this easily,
> > as an option in custom group `matching', right alongside
> > `case-fold-search'.  No?
> 
> Yes, it should be.
> I didn't make it a defcustom initially because I'm not sure how
> defcustoms interact with autoloading (I know it's bad in external
> packages, but I have no idea if it's ok in the emacs core).

Great.  Hoping this becomes the case.

Re: `character-fold-search' should be a user option, like `case-fold-search'

[Juri Linkov]

>> Like with ‘case-fold-search’ and ‘isearch-case-fold-search’ we could
>> make ‘char-fold-search’ customizable
>
> Yes.
>
>> and ‘isearch-char-fold-search’ togglable
>
> I'm not sure I understand. Character folding is already toggleable
> during isearch.
> Do you mean we specifically need a variable called
> isearch-char-fold-search? isearch-case-fold-search is an internal
> variable, not a user option, so I don't think there needs to be a
> char-fold analogue.

I thought ‘isearch-char-fold-search’ might be useful as an internal
variable to propagate its value to ‘isearch-occur’ (to check it and
not display a gibberish regexp in the *Occur* header).  But maybe
after renaming ‘isearch-word’ to a more suitable name like
‘isearch-regexp-based’ it could be used for the same purpose.