Christopher Browne <········@acm.org> wrote in message news:<··············@chvatal.cbbrowne.com>...
> Your "hack" of dropping the data into a file, and having another
> daemon periodically look at that file to notify you sounds like a
> reasonable, if not spectacularly elegant, solution.
Thanks. By the way, with biff/comsat not working I found the need to
write a functiion that shows me the key header lines (*) of any new e-mail
that arrives in my inbox, using code I already wrote for parsing BSD
format mail-files, and bleeping my terminal whenever doing so. Using
that primitive, I then wrote a loop that checks for new e-mail and
bleeps if so, and then sleeps for a while (typically one minute) before
checking again. I can then background my LISP running that while I go
off to do something else. With that already implemented now, it seems
like having Web users execute a mailto command will now be the easiest
way for somebody to let me know they need my immediate attention such as
to get on IRC with them. I might write an HTML chat, since Yahoo's HTML
chat doesn't work and there seems to be no other working HTML chat, and
there are some people I'd like to talk with live who don't have IRC.
If and when I do that, it'll be trivial to include a mailto to get my
attention when they're in my HTML-chat wanting me to join them. (In fact
I could have it mailto me whenever anyone whatsoever gets in my HTML chat
when I'm online.)
Nevertheless, for other purposes, I still might also implement the hack
to write info into a file which my daemon would check for at the same time
it checks for new e-mail, and bleep me if either. (Partly becaues mailto
requires manual confirmation by the Web user whereas I can implement
other methods that just happen without the Web user having to do anything
special, like just connecting to my WebServer application might be something
I'd want to know about promptly.)