Re: [NEWBIE] Updating code

From: Tony Robbins (
Date: 09/08/01

From: "NeoStar" <>
> Yes, I do have ASCII pfiles, but I thought the patch for it took care of
> all that kind of stuff

It did, for what it applied to.  When you add stuff, you have to take those
considerations yourself.

> > written example of it, anyway.  A bit of reordering is in order:
> >
> >   if (load_char(argument, &tmp_store) > -1) {
> >     CREATE(victim, struct char_data, 1);
> >     clear_char(victim);
> >     store_to_char(&tmp_store, victim);
> Why reorder it? The char has to be created before it can be used for
> comparisons!

In stock code, load_char() takes a name (argument) and a pointer to a
char_file_u structure.  What it does with this data is get the character
information from the player file and copy it to the char_file_u struct.

So, to use load_char(), no `victim' (struct char_data *) is needed, because
it's neither an argument nor a returned-value (as can be gleaned from the
"> -1" above, load_char returns a numeric value).

CREATE() is a front-end macro for memory allocation, (without looking at the
source, I believe it actually ends up calling calloc() rather than
malloc()).  Now, we have a pointer to a char_data structure which has
actually had memory allocated for it.  This means, at some point, it'll have
to be free()'d as well.

clear_char(victim) simply makes sure everything is set to a default (zero)
value on the char_data struct.  Technically, if I'm right about calloc(), I
think everything is zeroed already, but I could be wrong.

store_to_char(), in stock (and only in stock, the ascii player file patch
you're using doesn't have store_to_char()) copies information over from the
char_file_u (tmp_store) to the char_data struct which `victim' is a pointer

The reason why one should, in this case, arrange as above, is that you don't
allocate and free memory you're not going to use, unnecessarily.
load_char() returns -1 (that should be NOBODY, but that's just a #define
of -1 anyway) on failure, so we know, once it is done, that no player was

In that event, we won't allocate memory and try to get information that
isn't there.

> > BTW, this is why some wise person the other day suggested that one only
> > apply patches that one understands and can write oneself.  If you don't
> > understand how they interact, you rapidly cut-and-paste yourself into a
> Might I remind you that, in accordance with list rules, I added a flag to
> the subject of my email, indicating it was a NEWBIE topic. You shouldn't
> expect a lot.

Except effort is expected (and encouraged).

(and aliteration is awesome!)

   | FAQ: |
   | Archives: |

This archive was generated by hypermail 2b30 : 12/06/01 PST