[wt4hq] Broken Pipe
Mario Lorenz
ml at vdazone.org
Mon Jun 18 20:09:00 CEST 2007
Am 18. Jun 2007, um 09:46:56 schrieb Olivier F5MZN:
> Hi !
>
> About EINTR signal: this should be handled by the "default" case
> statement, which will issue in a disconnect. Do you suggest to handle it
> differently?
Yes, I do. Albeit it is unlikely in your code that a signal would ever occur,
the common Unix idiom is to handle EINTR like EAGAIN, since EINTR is not
a failure indication, its just that the operation was interrupted by some
signal occuring and couldn't be completed this time.
Therefore, imho, trying again is much more sensible,
otherwise you kill your client connection for no reason, and run
the risk of losing a QSO (or, worse, a GAB)...
(When reading through my mails again and again, you might notice I'm
sounding like an advanced case of paranoia, but hey, thats OK :)
On a related note, Olivier, will you be in Friedrichshafen at Ham Radio
next weekend?
Mario
--
Mario Lorenz Internet: <ml at vdazone.org>
Ham Radio: DL5MLO at DB0ERF.#THR.DEU.EU
The deorbiting of MIR was brought to you by radio FFH and MIRcrosoft, your
specialist for controlled crashes! (local radio station, on the very day)
More information about the Wt4hq
mailing list