On Fri, 21 Jun 2002, Stas Bekman wrote:
> what kind of problem?
mod_ssl for example would segfault if added to httpd.conf of a server that
was started without it, couldn't change server certs/keys, etc.
for mod_perl, much of dso nightmare(s) was caused by these assumptions.
> Nothing much, just want to do something on server 'start'. The Apache's
> immediate restart is not a logical thing, but an internal thing. I still
> want to distinguish between logical start (the server starts when it
> wasn't running) and restart (the server restarts when it was already
> running).
that's part of the problem, you can't distinguish. consider when httpd
has been started without modperl, then modperl is added to httpd.conf and
a restart signal given to httpd.
> This is an important distinction for administration needs.
> I'll give you an example:
>
> I want to be notified by email when the server restarts, and I pay
> special attention to those notifications because that usually means that
> something went wrong and the watchdog restarted the server. However I
> don't want to get this kind of notifications when the server didn't run
> and just starting.
>
> Well, I guess this is not the best example, but you get the idea.
so if you're lucky, you get two less emails if supported the {Re}Starting
variables. i don't see the point.
> The punch line is that user can code this logic by himself, by checking
> the flag in the pconf pool, which can be set on start. But I see no
> reason why not giving users this as an internal flag.
i don't see any reason to attempt setting such an internal flag.
or to suggest one to use such logic.
i'm also drawing blanks on why the variables were introduced in the first
place, maybe that would shed some light?
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
