Re: [ng-spice] History and command completion


To ng-spice@ieee.ing.uniroma1.it
From "GLAO S. Dezai" <dezai@hotbot.com>
Date Mon, 13 Nov 2000 06:22:40 -0800
Delivered-To mailing list ng-spice@ieee.ing.uniroma1.it
Mailing-List contact ng-spice-help@ieee.ing.uniroma1.it; run by ezmlm
Organization HotBot Mail (http://mail.hotbot.mailcity.lycos.com:80)
Reply-To ng-spice@ieee.ing.uniroma1.it

 
--

On Sat, 11 Nov 2000 11:00:46   widlok wrote:
>Hi,

>    Yes, You are right about the pointers. Other solution could be to change 
>cp_unquote in such way that it will not making a copy of the string but
>rather overwriting given string. I am not sure, but I will try to check it.
>
>You wrote (I don't remember where) that the calloc() might cause problems 
>with tfree(). I am not sure, but I think that it is not possible. As far as 
>I know calloc() uses malloc() inside, but don't take it for sure.
>
>    Regards
>            Michael
>

+ Al Davis calloc code. (Thanks to you)

Good, so i feel better now. The patches i sent should run. You may check only 
in variable.c why we have the segfault. But it seems that paolo has fixed the 
problem. 
I think that the cp_unquote should be left as it is. My patch had resolved 
the memory leak anywhere this function is called. So try to fix only the 
variable.c problem instead of reversing all the patch. Reversing the patch 
seems to be a big job as Arno said.
The original code has tried to free the cp_unquote result in some files. see 
for example com_chdir.c and xgraph.c. Unfortunately this effort has not been 
done in many files.
I think that fixing variable.c problem is better than changing cp_unquote 
code.
Best regards.
Dezai Glao




HotBot - Search smarter.
http://www.hotbot.com

Partial thread listing: