In message 20040513143724.GP27774@pengutronix.de you wrote:

• Patch by Kai-Uwe Bloem, Robert Schwebel, 13 May 2004: Copy data only when we _have_ a default environment, otherwhise set environment to 0.

Why is this patch needed?

memset (env_ptr, 0, sizeof(env_t));

Here the environment gets zeroed.

• memcpy (env_ptr->data,
  

• 	default_environment,
  

• 	sizeof(default_environment));
  

• if (sizeof(default_environment) > 0)
  

• 	memcpy (env_ptr->data,
  

• 		default_environment,
  

• 		sizeof(default_environment));
  

• else
  

• 	env_ptr->data[0] = '\0';
  

For sizeof(default_environment)>0 your patch does not change anything; for sizeof(default_environment)==0 it adds the "env_ptr->data[0] = '\0'" statement which seems to be redundand to me since the whole area was zeroed out earlier. [And "memcpy(d,s,0)" is a noop, right?]

Am I missing something?

Best regards,

Wolfgang Denk