Hi Martijn,

Martijn van Duren wrote on Thu, Jul 16, 2020 at 11:13:31PM +0200:

> I gave a quick look at replacing prompt with readpassphrase(3), but that
> would be more trouble than it's worth. (adjusting pledge, semantics
> change in where the "?..." would be printed).
> 
> Minor nits inline, but either way OK martijn@

Thanks to all three of you for checking, i just committed the patch.

> On Thu, 2020-07-16 at 21:49 +0200, Ingo Schwarze wrote:
[...]
>>  * There is no need to put a marker "return value ignored on
>>    purpose" where ignoring the return value almost never indicates
>>    a bug, like for fprintf, fflush, fclose.

> If you're going to change this, please also remove them from run()

Done.

>> -    (void)fflush(stderr);
>> +    fflush(stderr);

> According to setvbuf(3):
> The standard error stream stderr is initially unbuffered.
> And since setvbuf is not called in xargs this seems superfluous to me.
> If this is correct, there's a second fflush that can also be removed.

I didn't mix that into this diff because i didn't want to have two
unrelated, potentially non-trivial changes in the same commit.

But if you think this is worth cleaning up, by all means, send a
diff for it.  I'm not very experienced with stdio buffering, so i'd
hope you might get an OK from another developer.

Yours,
  Ingo

Reply via email to