On 02/22/2013 01:18 PM, Bernhard Voelker wrote:
On 02/21/2013 08:40 PM, Ondrej Oprala wrote:
so let's do your approach ;)
Hi Ondrej,
thank you.
In your patch, you missed the case when fork failed:
the target would yet be left behind.
Therefore, and to have slightly smaller code, I inverted the
logic in strip(), and used a bool return value instead of int.
The inclusion of unistd.h was unnecessary BTW:
maint.mk: the above are already included via system.h
Since this was reported in Redhat's Bugzilla, and it changes
the behavior in the case of such an error for the user,
I've also added a NEWS entry.
Finally, I've removed the error suppression in the test,
as it may help to diagnose problems in future.
I also changed the target file name to avoid interference
with the first test (paranoia, I know... ;-)
Is this okay for you?
looks good.
thanks,
Pádraig.
