Re: fooled by shifting date
- From: "Alexandre Ferrieux" <alexandre.ferrieux@xxxxxxxxx>
- Date: 6 Dec 2006 02:58:55 -0800
On Dec 6, 7:16 am, "Roy Terry" <royte...@xxxxxxxxxxxxx> wrote:
Broken "after" commands can be a big thorn.
Roy
Thanks Roy. Maybe I wasn't being paranoid after all :-)
[Kevin]
A possible solution would be to have a separate [at] command to wake
up according to the wall clock, and make [after] strictly dependent
on the interval timer.
Nice idea.
Perhaps available before 9.0 as loadable extension?
Seconded. I find it a bit frustrating that such a sanitization of a
core primitive be delayed with tons of bells and whistles like 9.0...
I would love to volunteer an earlier patch. But I need some technical
advice, on how to do it:
- a jiffies()-like call (which one on which OS ?)
- using setitimer() and the interruption of select() by the
SIGALRM, or the handler firing outside the select(). But what about
Windows ? Yet another worker thread ?
-Alex
.
- References:
- [after] fooled by shifting date
- From: Alexandre Ferrieux
- Re: [after] fooled by shifting date
- From: Kevin Kenny
- Re: [after] fooled by shifting date
- From: Roy Terry
- [after] fooled by shifting date
- Prev by Date: Re: ANN: iaxclient 0.2
- Next by Date: Re: A Problem with tkhtml 2.0 in ActiveTcl 8.4.13 Distribution
- Previous by thread: Re: [after] fooled by shifting date
- Next by thread: Re: [after] fooled by shifting date
- Index(es):
Relevant Pages
|
|