httpd: memory hole in function addEnv() and more

[Dirk Clemens]

Rob Landley wrote:

>>
>>We know the maximum needed size of the temp buffer within one
>>call of sendCgi():
> 
> 
> Considering that the URL is of arbitrar arbitrary length...
> 
I have done that.

> 
>> needed_size = max_env_var_name + maximum( stdlen(uri), MAX_PATH) + 2;

I mean strlen(uri) and not stdlen(uri), sorry.

The problem is, that the space for a possible large URI should not
be reserved by alloca(). See the other posting a few minutes
before this.

>>  snprintf(envbuf,envbuf_size,"NAME=%s",value);
>>  putenv(envbuf);
>
>
> It's still possible that numerous calls to one wrapper with two
arguments is
> going to be smaller than numerous calls to two functions with a total
of four
> arguments.  The wrapper isn't all that big, and a simplification that
makes
> the binary bigger is a hard sell here...

It is only an idea to replace then current addEnv()
(perhaps again as wrapper) without many alloc+free cycles.


Dirk