[per.starback@gmail.com: completion of coding systems]

[per.starback@gmail.com: completion of coding systems]

[Richard Stallman]

Would someone please do this in Emacs 22, then ack?

------- Start of forwarded message -------
X-Spam-Status: No, score=0.6 required=5.0 tests=NO_REAL_NAME,
	UNPARSEABLE_RELAY autolearn=no version=3.1.0
Date: Thu, 11 Oct 2007 13:00:14 +0200
From: per.starback@gmail.com
To: bug-gnu-emacs@gnu.org
Subject: completion of coding systems

In GNU Emacs 22.1:

This happened to a user here recently:

  C-x RET f UTF-8 RET	=>  [No match]

(since the coding system is really called "utf-8" in lowercase).

Small suggestion: Set completion-ignore-case during
read-coding-system.

Reason: Some coding systems are often given with uppercase letters
outside of Emacs so one might thing that is the proper form to use.
It wouldn't hurt anyway, since all coding systems in Emacs have
all-lower-case names anyway.
------- End of forwarded message -------

Re: [per.starback@gmail.com: completion of coding systems]

[Glenn Morris]

Richard Stallman wrote:

> Would someone please do this in Emacs 22, then ack?

Is this right?


*** coding.c	25 Jul 2007 07:49:38 -0000	1.351.2.2
--- coding.c	13 Oct 2007 03:11:31 -0000
***************
*** 6563,6573 ****
--- 6563,6577 ----
       Lisp_Object prompt, default_coding_system;
  {
    Lisp_Object val;
+   int count = SPECPDL_INDEX ();
+ 
    if (SYMBOLP (default_coding_system))
      default_coding_system = SYMBOL_NAME (default_coding_system);
+   specbind (intern ("completion-ignore-case"), Qt);
    val = Fcompleting_read (prompt, Vcoding_system_alist, Qnil,
  			  Qt, Qnil, Qcoding_system_history,
  			  default_coding_system, Qnil);
+   unbind_to (count, Qnil);
    return (SCHARS (val) == 0 ? Qnil : Fintern (val, Qnil));
  }
  

> From: per.starback@gmail.com
> Subject: completion of coding systems
> To: bug-gnu-emacs@gnu.org
> Date: Thu, 11 Oct 2007 13:00:14 +0200
[...]
> Small suggestion: Set completion-ignore-case during read-coding-system.

Re: [per.starback@gmail.com: completion of coding systems]

[Richard Stallman]

    +   int count = SPECPDL_INDEX ();
    + 
	if (SYMBOLP (default_coding_system))
	  default_coding_system = SYMBOL_NAME (default_coding_system);
    +   specbind (intern ("completion-ignore-case"), Qt);
	val = Fcompleting_read (prompt, Vcoding_system_alist, Qnil,
			      Qt, Qnil, Qcoding_system_history,
			      default_coding_system, Qnil);
    +   unbind_to (count, Qnil);

That would work, but it would be cleaner to use 

    specbind (Qcompletion_ignore_case, Qt);

instead of interning it again.

Re: [per.starback@gmail.com: completion of coding systems]

[Glenn Morris]

Richard Stallman wrote:

> That would work, but it would be cleaner to use 
>
>     specbind (Qcompletion_ignore_case, Qt);
>
> instead of interning it again.

I was copying the behaviour of read-file-name in fileio.c.

You mean you would rather have this:


--- coding.c	25 Jul 2007 07:49:38 -0000	1.351.2.2
+++ coding.c	15 Oct 2007 07:13:18 -0000
@@ -386,6 +386,8 @@
    decided.  */
 Lisp_Object eol_mnemonic_undecided;
 
+Lisp_Object Qcompletion_ignore_case;
+
 /* Format of end-of-line decided by system.  This is CODING_EOL_LF on
    Unix, CODING_EOL_CRLF on DOS/Windows, and CODING_EOL_CR on Mac.
    This has an effect only for external encoding (i.e. for output to
@@ -6558,16 +6560,22 @@
 
 DEFUN ("read-coding-system", Fread_coding_system, Sread_coding_system, 1, 2, 0,
        doc: /* Read a coding system from the minibuffer, prompting with string PROMPT.
-If the user enters null input, return second argument DEFAULT-CODING-SYSTEM.  */)
+If the user enters null input, return second argument DEFAULT-CODING-SYSTEM.
+Ignores case when completing coding systems (all Emacs coding systems
+are lower-case).  */)
      (prompt, default_coding_system)
      Lisp_Object prompt, default_coding_system;
 {
   Lisp_Object val;
+  int count = SPECPDL_INDEX ();
+
   if (SYMBOLP (default_coding_system))
     default_coding_system = SYMBOL_NAME (default_coding_system);
+  specbind (Qcompletion_ignore_case, Qt);
   val = Fcompleting_read (prompt, Vcoding_system_alist, Qnil,
 			  Qt, Qnil, Qcoding_system_history,
 			  default_coding_system, Qnil);
+  unbind_to (count, Qnil);
   return (SCHARS (val) == 0 ? Qnil : Fintern (val, Qnil));
 }
 
@@ -7733,6 +7741,9 @@
   /* Target FILENAME is the third argument.  */
   Fput (Qwrite_region, Qtarget_idx, make_number (2));
 
+  Qcompletion_ignore_case = intern ("completion-ignore-case");
+  staticpro (&Qcompletion_ignore_case);
+
   Qcall_process = intern ("call-process");
   staticpro (&Qcall_process);
   /* Target PROGRAM is the first argument.  */

Re: [per.starback@gmail.com: completion of coding systems]

[Richard Stallman]

There is already a variable Qcompletion_ignore_case.
You just have to use it.

Re: [per.starback@gmail.com: completion of coding systems]

[Glenn Morris]

Richard Stallman wrote:

> There is already a variable Qcompletion_ignore_case.

Only in some #ifdef VMS code in dired.c AFAICS.

> You just have to use it.

I'm probably missing your point by several miles...

Re: [per.starback@gmail.com: completion of coding systems]

[Richard Stallman]

    > There is already a variable Qcompletion_ignore_case.

    Only in some #ifdef VMS code in dired.c AFAICS.

I did not see it was #ifdef VMS.

We only need one (unconditional) definition of the variable in Emacs.
You may as well unconditionalize the one in dired.c.

Re: [per.starback@gmail.com: completion of coding systems]

[Glenn Morris]

Richard Stallman wrote:

> We only need one (unconditional) definition of the variable in Emacs.
> You may as well unconditionalize the one in dired.c.

It made more sense to me to define it in minibuf.c, so that's what I did.