CamelHump word movement in Emacs

CamelHump word movement in Emacs

[Tassilo Horn]

Hi all,

in many editors for "modern" programming languages like Java or C#, the
normal word movement commands also stop on CamelHumps.  This is very
convenient, because nowadays the convention for identifiers names more
and more turns to using camelCaseNaming.

Here's an example:

  int foo_bar_baz() {}

With the usual forward/backward-word commands, point always stops at the
_ (when moving forward) or the first char of the component word (when
moving backward).  I really like that behavior.

Unfortunately, it doesn't work if the function uses camelCase naming.

  int fooBarBaz() {}

Here, forward/backward-word jump over the complete identifier.  What I
would really like to have, was that those commands move point to the
capital letters (the 2 Bs), too.

IMO, that would be the right thing to do, because camelCase is nearly
never used for something else than separating words.

What do you think?

Bye,
Tassilo

Re: CamelHump word movement in Emacs

[Giuseppe Scrivano]

have you considered "c-subword-mode"?

Cheers,
Giuseppe



Tassilo Horn <tassilo@member.fsf.org> writes:

> Hi all,
>
> in many editors for "modern" programming languages like Java or C#, the
> normal word movement commands also stop on CamelHumps.  This is very
> convenient, because nowadays the convention for identifiers names more
> and more turns to using camelCaseNaming.
>
> Here's an example:
>
>   int foo_bar_baz() {}
>
> With the usual forward/backward-word commands, point always stops at the
> _ (when moving forward) or the first char of the component word (when
> moving backward).  I really like that behavior.
>
> Unfortunately, it doesn't work if the function uses camelCase naming.
>
>   int fooBarBaz() {}
>
> Here, forward/backward-word jump over the complete identifier.  What I
> would really like to have, was that those commands move point to the
> capital letters (the 2 Bs), too.
>
> IMO, that would be the right thing to do, because camelCase is nearly
> never used for something else than separating words.
>
> What do you think?
>
> Bye,
> Tassilo

Re: CamelHump word movement in Emacs

[Tassilo Horn]

Giuseppe Scrivano <gscrivano@gnu.org> writes:

Hi Giuseppe,

> have you considered "c-subword-mode"?

Nope, I didn't know it till now, but it does exactly what I want.

I think this feature is generally useful and has nothing to do with C,
so I propose to rename it (and its commands and the file) to
`subword-mode' or `camelcase-mode'.

Also, it would be nice to have it as a global minor mode, too.

Any objections?  If not, I'd volunteer to do that.

Bye,
Tassilo

> Tassilo Horn <tassilo@member.fsf.org> writes:
>
>> Hi all,
>>
>> in many editors for "modern" programming languages like Java or C#, the
>> normal word movement commands also stop on CamelHumps.  This is very
>> convenient, because nowadays the convention for identifiers names more
>> and more turns to using camelCaseNaming.
>>
>> Here's an example:
>>
>>   int foo_bar_baz() {}
>>
>> With the usual forward/backward-word commands, point always stops at the
>> _ (when moving forward) or the first char of the component word (when
>> moving backward).  I really like that behavior.
>>
>> Unfortunately, it doesn't work if the function uses camelCase naming.
>>
>>   int fooBarBaz() {}
>>
>> Here, forward/backward-word jump over the complete identifier.  What I
>> would really like to have, was that those commands move point to the
>> capital letters (the 2 Bs), too.
>>
>> IMO, that would be the right thing to do, because camelCase is nearly
>> never used for something else than separating words.
>>
>> What do you think?
>>
>> Bye,
>> Tassilo

Re: CamelHump word movement in Emacs

[Stefan Monnier]

> I think this feature is generally useful and has nothing to do with C,
> so I propose to rename it (and its commands and the file) to
> `subword-mode' or `camelcase-mode'.

IIRC the same suggestion to make it generic came up right around the
Emacs-23 freeze, which is why it didn't happen.  If someone wants to
make this generic *right now* he'd better hurry.  Otherwise it'll have
to wait for Emacs-24 (or Emacs-23.3 maybe).

> Also, it would be nice to have it as a global minor mode, too.
> Any objections?  If not, I'd volunteer to do that.

IIRC there was a general agreement that it would be a desirable feature.


        Stefan

Re: CamelHump word movement in Emacs

[Tassilo Horn]

[-- Attachment #1: Type: text/plain, Size: 806 bytes --]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

Hi Stefan,

>> I think this feature is generally useful and has nothing to do with
>> C, so I propose to rename it (and its commands and the file) to
>> `subword-mode' or `camelcase-mode'.
>
> IIRC the same suggestion to make it generic came up right around the
> Emacs-23 freeze, which is why it didn't happen.  If someone wants to
> make this generic *right now* he'd better hurry.

Ok, here it is (attached to this mail).  I removed all cc-mode
dependencies (which were all about XEmacs and old emacs version
compatibility anyway).

Is it ok to commit?

If yes, how do I do that with CVS?  cvs delete progmodes/cc-subword.el
and cvs add subword.el?  There seems to be no renaming command.

And what do I have to write into the ChangeLog?

Bye,
Tassilo

[-- Attachment #2: subword.el --]
[-- Type: application/emacs-lisp, Size: 9804 bytes --]

Re: CamelHump word movement in Emacs

[Stefan Monnier]

>>> I think this feature is generally useful and has nothing to do with
>>> C, so I propose to rename it (and its commands and the file) to
>>> `subword-mode' or `camelcase-mode'.
>> IIRC the same suggestion to make it generic came up right around the
>> Emacs-23 freeze, which is why it didn't happen.  If someone wants to
>> make this generic *right now* he'd better hurry.
> Ok, here it is (attached to this mail).  I removed all cc-mode
> dependencies (which were all about XEmacs and old emacs version
> compatibility anyway).

> Is it ok to commit?

It needs to add some compatibility aliases or adjust the CC-mode (and
maybe other?) code that refers to those functions with their old name.
Also, I'd prefer to rename the functions so that they all share
a `subword-' prefix.

> If yes, how do I do that with CVS?  cvs delete progmodes/cc-subword.el
> and cvs add subword.el?

Yes.  Please make sure you commit the add and the delete as a single
commit rather than as two separate commits.

> There seems to be no renaming command.

Indeed, there's no support for renaming.

> And what do I have to write into the ChangeLog?

Usually I write one entry for the new name that says something like
"Rename from <oldname>".  You can also add second entry for the old name
that says "Rename to <newname>".  You'll also want to add to the entry
for the new name the changes performed (renaming of the functions, ...),
of course.


        Stefan

Re: CamelHump word movement in Emacs

[Tassilo Horn]

Stefan Monnier <monnier@IRO.UMontreal.CA> writes:

>> Is it ok to commit?
>
> It needs to add some compatibility aliases or adjust the CC-mode (and
> maybe other?) code that refers to those functions with their old name.

I've grepped all code, and it seems to be only cc-mode.  There, I
changed the code to use the new names.

> Also, I'd prefer to rename the functions so that they all share
> a `subword-' prefix.

In general I agree, but in this case I think fitting to naming of the
word movement functions is more desirable.

I'll try to compile emacs with those changes now, and then I'll commit.

Bye,
Tassilo

Re: CamelHump word movement in Emacs

[Eli Zaretskii]

> From: Tassilo Horn <tassilo@member.fsf.org>
> Date: Fri, 20 Nov 2009 08:25:19 +0100
> Cc: Giuseppe Scrivano <gscrivano@gnu.org>, emacs-devel@gnu.org
> 
> Stefan Monnier <monnier@IRO.UMontreal.CA> writes:
> 
> >> Is it ok to commit?
> >
> > It needs to add some compatibility aliases or adjust the CC-mode (and
> > maybe other?) code that refers to those functions with their old name.
> 
> I've grepped all code, and it seems to be only cc-mode.  There, I
> changed the code to use the new names.
> 
> > Also, I'd prefer to rename the functions so that they all share
> > a `subword-' prefix.
> 
> In general I agree, but in this case I think fitting to naming of the
> word movement functions is more desirable.
> 
> I'll try to compile emacs with those changes now, and then I'll commit.

It looks like you didn't make the necessary changes in
lisp/Makefile.in and lisp/makefile.w32-in.  They still refer to
cc-subword.el and cc-subword.elc.

Re: CamelHump word movement in Emacs

[Tassilo Horn]

Eli Zaretskii <eliz@gnu.org> writes:

Hi Eli,

>> I'll try to compile emacs with those changes now, and then I'll
>> commit.
>
> It looks like you didn't make the necessary changes in
> lisp/Makefile.in and lisp/makefile.w32-in.  They still refer to
> cc-subword.el and cc-subword.elc.

Oh, indeed.  I've changed that right now.

Thanks for the pointer,
Tassilo

Re: CamelHump word movement in Emacs

[Eli Zaretskii]

> From: Tassilo Horn <tassilo@member.fsf.org>
> Date: Fri, 20 Nov 2009 15:05:49 +0100
> Cc: emacs-devel@gnu.org, monnier@IRO.UMontreal.CA, gscrivano@gnu.org
> 
> > It looks like you didn't make the necessary changes in
> > lisp/Makefile.in and lisp/makefile.w32-in.  They still refer to
> > cc-subword.el and cc-subword.elc.
> 
> Oh, indeed.  I've changed that right now.

Thanks.

There's one other problem: compiling cc-cmds.el displays this warning:

  In end of data:
  cc-cmds.el:4575:1:Warning: the following functions are not known to be
      defined: forward-subword, backward-subword

Re: CamelHump word movement in Emacs

[Stefan Monnier]

> I've grepped all code, and it seems to be only cc-mode.

That was my impression as well.

> There, I changed the code to use the new names.

Great, thanks.

>> Also, I'd prefer to rename the functions so that they all share
>> a `subword-' prefix.
> In general I agree, but in this case I think fitting to naming of the
> word movement functions is more desirable.

No, this is a separate package not part of the core, so it has to follow
the coding convention of using a distinguished prefix.


        Stefan

Re: CamelHump word movement in Emacs

[Tassilo Horn]

Eli Zaretskii <eliz@gnu.org> writes:

> There's one other problem: compiling cc-cmds.el displays this warning:
>
>   In end of data:
>   cc-cmds.el:4575:1:Warning: the following functions are not known to be
>       defined: forward-subword, backward-subword

I declare-functioned them, and the byte-compiler is quiet again.

Bye,
Tassilo

Re: CamelHump word movement in Emacs

[Tassilo Horn]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

Hi Stefan!

>>> Also, I'd prefer to rename the functions so that they all share
>>> a `subword-' prefix.
>> In general I agree, but in this case I think fitting to naming of the
>> word movement functions is more desirable.
>
> No, this is a separate package not part of the core, so it has to
> follow the coding convention of using a distinguished prefix.

Ok, chief, I'll rename them all! ;-)

What naming do you have in mind?  Simply turning around?

  subword-backward-kill
  subword-backward
  subword-capitalize
  subword-downcase
  subword-forward
  subword-kill
  subword-mark
  subwords-transpose
  subword-upcase

Or keep the names and only add a prefix?

  subword-backward-kill-subword
  subword-backward-subword
  subword-capitalize-subword
  subword-downcase-subword
  subword-forward-subword
  subword-kill-subword
  subword-mark-subword
  subword-transpose-subwords
  subword-upcase-subword

Bye,
Tassilo

Re: CamelHump word movement in Emacs

[Chong Yidong]

Tassilo Horn <tassilo@member.fsf.org> writes:

> Ok, here it is (attached to this mail).  I removed all cc-mode
> dependencies (which were all about XEmacs and old emacs version
> compatibility anyway).

The file should go into progmodes/, not the main lisp/ directory.

Re: CamelHump word movement in Emacs

[Tassilo Horn]

Chong Yidong <cyd@stupidchicken.com> writes:

Hi!

>> Ok, here it is (attached to this mail).  I removed all cc-mode
>> dependencies (which were all about XEmacs and old emacs version
>> compatibility anyway).
>
> The file should go into progmodes/, not the main lisp/ directory.

I don't think so.  CamelCase in words is common in programming
languages, but not restricted to them.  I enabled global-subword-mode,
because I want this behavior also when writing mail or plain text files.

Of course, YMMV, and if you still think it should go into progmodes/,
I'll move it there.

Bye,
Tassilo

Re: CamelHump word movement in Emacs

[Stefan Monnier]

> What naming do you have in mind?  Simply turning around?

Ys, except for ..

>   subword-backward-kill
>   subword-backward
>   subword-capitalize
>   subword-downcase
>   subword-forward
>   subword-kill
>   subword-mark
>   subwords-transpose
    ^^^^^^^^
No s at the end.


        Stefan

Re: CamelHump word movement in Emacs

[Dan Nicolaescu]

Tassilo Horn <tassilo@member.fsf.org> writes:

  > Stefan Monnier <monnier@iro.umontreal.ca> writes:
  > 
  > Hi Stefan,
  > 
  > >> I think this feature is generally useful and has nothing to do with
  > >> C, so I propose to rename it (and its commands and the file) to
  > >> `subword-mode' or `camelcase-mode'.
  > >
  > > IIRC the same suggestion to make it generic came up right around the
  > > Emacs-23 freeze, which is why it didn't happen.  If someone wants to
  > > make this generic *right now* he'd better hurry.
  > 
  > Ok, here it is (attached to this mail).  I removed all cc-mode
  > dependencies (which were all about XEmacs and old emacs version
  > compatibility anyway).

BTW, I think this mode should also be listed in `mode-line-mode-menu',
glasses-mode which helps in similar ways is listed there.

Re: CamelHump word movement in Emacs

[Tassilo Horn]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> What naming do you have in mind?  Simply turning around?
>
> Ys, except for ..
>
>>   subword-backward-kill
>>   subword-backward
>>   subword-capitalize
>>   subword-downcase
>>   subword-forward
>>   subword-kill
>>   subword-mark
>>   subwords-transpose
>     ^^^^^^^^
> No s at the end.

Sure, I'll do it after lunch.

Bye,
Tassilo

Re: CamelHump word movement in Emacs

[Chong Yidong]

Tassilo Horn <tassilo@member.fsf.org> writes:

> I don't think so.  CamelCase in words is common in programming
> languages, but not restricted to them.  I enabled global-subword-mode,
> because I want this behavior also when writing mail or plain text files.

The cap-words, glasses, and hideshow packages are also in progmodes/.
Being in that directory does not mean the package is *only* useful for
programming, just that it's mostly useful for it.

Subword mode should go into progmodes/.

Re: CamelHump word movement in Emacs

[Miles Bader]

Hmm, how about

(defalias 'studly-caps-mode 'subword-mode)
(defalias 'sTuDLyCapS-mode 'subword-mode)

?

-- 
Bacchus, n. A convenient deity invented by the ancients as an excuse for
getting drunk.