Re: [CODE] Warmboot Patch

From: Mark A. Heilpern (heilpern@MINDSPRING.COM)
Date: 07/14/98


At 08:29 PM 7/13/98 -0400, you wrote:
>As the socket descriptor itself is an integer, writing down the actual
>descriptor to a file, and not closing the socket, and writing down the
>inet_addr, and maybe enough to know which character to reload, would
>enable win95 to a hot reboot, but I really have never tried this, because
>I have no reason to.  However, it's something new and cool, and ify, so if
>anyone has any time and a win95 box(just say it's your friend's :P) then I
>would be interested in the results, as I am sure are other people in the
>list.

Simply not closing the sockets prior to termination is not sufficient. You
have
to be able to get the new boot of the mud to have them as open descriptors;
otherwise the act of terminating the "old" boot of the mud will sever
everyone's
link. (Regardless, if the "new" boot doesn't inherit those open sockets, it
has
no way to get access to them.) The unix exec*() calls [thanks for the
correction, Erwin, to my {failing} memory :)] calls start up a child
process in
a manner that allows I/O path inheritance.


     +------------------------------------------------------------+
     | Ensure that you have read the CircleMUD Mailing List FAQ:  |
     | http://democracy.queensu.ca/~fletcher/Circle/list-faq.html |
     +------------------------------------------------------------+



This archive was generated by hypermail 2b30 : 12/15/00 PST