Re: lynx-dev Cookie problems?

[Bela Lubkin]

Mike Castle wrote:

> If a multiple directories isn't a solution, how about a client/server
> solution.
> 
> Lynx fires up, finds no "lynx acting as server", opens up a receiving port
> and is willing to accept connections on a socket which it publishes in a
> file somewhere.  That server will always maintain accesses to the cookie
> file.  On systems that support it, lynx could fork off a system just to
> handle that, and perhaps DNS look ups as well (preforking).
> 
> When a lynx starts up, it checks to see if something is registered,
> verifies it's still there, then uses it to access cookies.
> 
> Just a random thought (and something not easily implemented, but probably
> a better overall solution than locking files).

That might be better when it worked: but the level of difficulty in
implementing it across all the platforms Lynx supports would be much
worse than implementing locks.  Then, even on a system where it worked,
it would be flaky under stressful conditions.

>Bela<