On Wed, 05 Mar 2014 11:31:29 +0100 Peter Marschall <pe...@adpm.de> wrote:
> Oops, now you got me. > I did not check doings tests, but relied on the manual pages of > IO::Socket::IP only. > > Maybe I got confused by those two facts > * In older versions, of IO::Socket::IP, it mentioned the Timeout > option in the list of incompatibilities with IO::Socket::INET*. > (This changed in versions > 0.25 ( > * In recent version of the man page this warning has gone, but > unfortunately Timeout is not mentioned in the list of supported > options either. > > Paul, > from the mail above, I get it that Timeout is working in > IO::Socket::IP. Can you confirm? > Additionally, may I ask you to add Timeout to the documentation? Ah yes; this is actually down to a misunderstanding on my part. Originally I had presumed that IO::Socket::IP would have to implement the "Timeout" option, because IO::Socket::INET lists it in its documentation. However, on reading the source code I discovered that ::INET doesn't implement it; the code to actually implement that option appears entirely within IO::Socket itself, which ::INET inherits from. Because ::IP is also a subclass of plain IO::Socket, it automatically inherits the Timeout option. Ideally core's documentation should be changed to document Timeout in IO::Socket rather than IO::Socket::INET, which would also solve this issue. But long story short: IO::Socket::IP supports the Timeout option in the same way that IO::Socket::INET documents for itself. -- Paul "LeoNerd" Evans leon...@leonerd.org.uk ICQ# 4135350 | Registered Linux# 179460 http://www.leonerd.org.uk/
