Re: null terminated strings (was: argz SMOB)

Re: null terminated strings (was: argz SMOB)

[Roland Orre]

On Fri, 2004-01-16 at 10:10, Andreas Voegele wrote:

> "Remove calls to SCM_STRING_COERCE_0TERMINATION_X.  Since the
> substring type is gone, all strings are 0-terminated anyway."

Such a statement is very worrying. What happened to the promises
that all substrings will be shared strings? 

	Roland Orre





_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-user

argz SMOB

[Brian S McQueen]

I have the beginnings of a SMOB library for GNU libc's argz.  I will also
be adding the envz functionality.

Brian McQueen


_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-user

Re: argz SMOB

[Daniel Skarda]

Hello,

Brian S McQueen <bqueen@nas.nasa.gov> writes:
> I have the beginnings of a SMOB library for GNU libc's argz.  I will also
> be adding the envz functionality.

 I am sorry I do not see what is it good for. Could you please explain what is
your itch you are trying to scratch with this SMOB library?

0.


_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-user

Re: argz SMOB

[Brian S McQueen]

Since scheme strings can contain the null character, I noticed there was
no need for an argz smob.  I removed it by using scheme strings instead
and it works excellently.  The C functions that produce argzs, are all
callable from scheme since the argz is defined by a pointer and a length,
just as are scheme strings.  The C functions are wrapped in simple
functions which use the guile API, and the trailing null char in each argz
is dropped.

I am sure that nobody is particularly interested in argzs, but some
readers may be interested in a real life example of the guile API.  Below
are some examples of how I gained access to some C function from guile.
If any of you veterans have any more constructive advice, I would be glad
to hear it, but don't ask me to get rid of the argz.  They are here to
stay!  Particularly, I wonder about the best way to produce a null
terminated C string from a scheme string.  I used scm_must_malloc, memcpy,
memset.  I expected a ready made guile call for this purpose, but
I did not find any.

A simple call to a function which is expecting an argz and returns
nothing:

static SCM printer_hostile_printer(SCM scm_out_buff) {

  struct argz_holder out_buff;

  SCM_ASSERT (SCM_STRINGP (scm_out_buff), scm_out_buff, SCM_ARG1,
"printer_hostile_printer");

  out_buff.argz = SCM_STRING_CHARS(scm_out_buff);
  out_buff.argz_len = SCM_STRING_LENGTH(scm_out_buff);

  output_printer(&out_buff);

  return SCM_UNDEFINED;

}

A call to a database query function which returns an argz full of query
results:

static SCM get_from_db(SCM scm_login) {

  char *login_chrs;
  int login_len;

  char * login;

  struct db_parm_holder  login_parm = { 0 };
  struct db_parm_holder argz_parm = { 0 };

  SCM ret_val;

  SCM_ASSERT (SCM_STRINGP (scm_login), scm_login, SCM_ARG1,
"get_from_db");

  login_chrs = SCM_STRING_CHARS(scm_login);
  login_len = SCM_STRING_LENGTH(scm_login);

  login  = (char *)scm_must_malloc(login_len + 1, "get_from_db");
  memcpy(login, login_chrs, login_len);
  memset(login + login_len, '\0', 1);

  set_db_in_parm_str(&login_parm, "@login", USER_ID_LEN, login);
  set_db_ret_parm_envz(&argz_parm, NULL);
  db_query_va("get_from_db", &login_parm, &argz_parm, NULL);

  //don't take the last null term on the argz
  ret_val =  scm_mem2string(argz_parm.data, argz_parm.data_len - 1);

  free(argz_parm.data);

  scm_must_free(login);

  return ret_val;

}





_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-user

Re: argz SMOB

[Paul Jarc]

Brian S McQueen <bqueen@nas.nasa.gov> wrote:
> Particularly, I wonder about the best way to produce a null
> terminated C string from a scheme string.

SCM_STRING_COERCE_0TERMINATION_X(scheme_string);
char* s=SCM_STRING_CHARS(scheme_string);

Or do you need a separate copy?

> static SCM printer_hostile_printer(SCM scm_out_buff) {
>
>   struct argz_holder out_buff;
>
>   SCM_ASSERT (SCM_STRINGP (scm_out_buff), scm_out_buff, SCM_ARG1,
> "printer_hostile_printer");

You could also write that as:
#define FUNC_NAME s_printer_hostile_printer
SCM_DEFINE(printer_hostile_printer, "printer_hostile_printer",
           1, 0, 0, (SCM scm_out_buff))
{
  struct argz_holder out_buff;
  SCM_VALIDATE_STRING(SCM_ARG1, scm_out_buff);


paul


_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-user

null terminated strings (was: argz SMOB)

[Andreas Voegele]

Paul Jarc writes:

> Brian S McQueen <bqueen@nas.nasa.gov> wrote:
>> Particularly, I wonder about the best way to produce a null
>> terminated C string from a scheme string.
>
> SCM_STRING_COERCE_0TERMINATION_X(scheme_string);
> char* s=SCM_STRING_CHARS(scheme_string);

Is it necessary or desirable to use the macro
SCM_STRING_COERCE_0TERMINATION_X if one uses Guile 1.6?  I've found
the following statement in the file libguile/strings.c that comes with
Guile 1.6.4:

"[...] we promise that strings are null-terminated."

And in guile-readline/Changelog:

"Remove calls to SCM_STRING_COERCE_0TERMINATION_X.  Since the
substring type is gone, all strings are 0-terminated anyway."


_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-user