httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@lnd.com>
Subject RE: Reviewing the exit() calls in 2.0
Date Tue, 30 May 2000 17:57:43 GMT
> From: rbb@covalent.net [mailto:rbb@covalent.net]
> Sent: Tuesday, May 30, 2000 12:41 PM
> 
> OK, after OtherBill suggested that somebody review where we are just
> calling exit in 2.0, I did it.  This is a long message that basically
> details every call to exit() in the server and gives my opinion on 
> whether or not it needs to change to something that cleans up after 
> Apache.  None of these are gospel, just my opinion.  If somebody is 
> looking for something to do, this is some low-hanging fruit.

Thank you for your *VERY* thorough review of these scenarios.  The
Win32 truly cares, since it is hanging on to a whole lot more than
some memory pools that will go away in any case (the winsock interface,
the service control manager, monitor win threads etc...

The group that scared me most was certainly mod_proxy... it sure
doesn't make sense that a module should be able to knock off the
server, but you clarified (most of) the points, and it will give
me a chance to eyeball the worst offenders you've identified.
Don't let that stop anyone else from picking them off the list,
though :-)

> ./src/os/win32/service.c:272:        exit(0);
> 
> Haven't got a clue what is going on here, somebody else want 
> to comment?

Everything in service.c I've recoded, and winnt.c (mpm) I am 
reviewing, so I'll be catching anything that is nonsensical 
(I hope!)

Kudos!  Spared me from updating this issue in STATUS (I'll be
reviewing all the potential lifetime pool allocation issues at
the same time)... as I suggested, there "appeared" to be a
showstopper issue here.

Bill


Mime
View raw message