Re: Global Proxy Settings

[Michael Meeks <michael ximian com>]

Hi James,

On Fri, 2001-12-28 at 15:48, James M. Cape wrote:
> Looking at the README, HTML docs, and headers, I *think* it's too
> high-level for what I'm doing.

	Quite possibly - I was merely curious as to whether you had looked at
it.

> Mainly, it doesn't appear as though I
> could emit signals based on the gethostbyname & connect return, or do
> forward & reverse DNS w/ caching (all of which libgtcpsocket do now)
> without rewriting some-to-a-lot of the code in linc.

	Yes true - and linc has the major disadvantage of being frozen.

> Also, I'm not
> exactly sure how linc_main_loop_run() works with the Glib main loop, if
> the gethostbyname call in linc is threaded (as it is in libgtcpsocket),
> or how/if threads factor into linc (it doesn't appear as though they do,
> really, it just seems as though it supports them).

	Yes - we just block like a pig when doing host name lookups - which
sucks really; one thing it'd be nice to fix now we can cache data going
out on the connection.

> IOW, I'd like to use it (even though it would mean I "wasted" months of
> time figuring out TCP/IP from a programming standpoint :-)), but it
> doesn't appear to allow me enough flexibility for the user experience
> I'm aiming for with libgtcpsocket. It does look really useful for
> behind-the-scenes stuff, like ORBit, though.

	Fair enough; sounds reasonable. I'd like to re-factor it / expand the
hostname lookup code to make it possible to share code at some stage if
you're interested - but it seems to make no sense for now.

	Thanks for taking a look,

	Regards,

		Michael.

-- 
 mmeeks gnu org  <><, Pseudo Engineer, itinerant idiot