Possible Bug in Ruby?


(Yukihiro Matsumoto) #1

Hi,

···

In message “Re: Possible Bug in Ruby?” on 02/05/31, nobu.nokada@softhome.net nobu.nokada@softhome.net writes:

Thanks for your patch! It worked like a charm to solve my problem. Will
it ever be included in the current or release version of ruby?

Persuade Matz of the requirement. :slight_smile:

IMO, Sockets can be half-baked at exceptions, unless the patch.

Sorry, I didn’t follow the discussion. Persuade me.
Or, just commit the fix, if you think it’s the right thing.

						matz.

(Nobuyoshi Nakada) #2

Hi,

Thanks for your patch! It worked like a charm to solve my problem. Will
it ever be included in the current or release version of ruby?

Persuade Matz of the requirement. :slight_smile:

IMO, Sockets can be half-baked at exceptions, unless the patch.

half-baked wasn’t be proper word.

Sorry, I didn’t follow the discussion. Persuade me.

Since ruby_connect() is interruptible an exception can raise
there, a socket descriptor and an addrinfo in init_inetsock()
aren’t bound to any Object and won’t be freed in such cases.

It may be possible to bind them to any Object and leave it to
GC, but I’d like to ensure they are freed locally.

···

At Sun, 2 Jun 2002 21:14:11 +0900, Yukihiro Matsumoto wrote:


Nobu Nakada