I should have also mentioned that not only does the "returned" come back out of order there are also more than there should be. There's 3 calls and 4 results.
Worse it now appears that if I hit the server long enough to come back to the same process the value has stayed put. I assume this is why the suggested use of $r->pnotes? If so how do I get access to $r (which is shifted off early in the .pl) when this is deep within a module (without a global that is)? This was exactly the kind of problem I was striving (not very well) to solve in the first place. Matt. > -----Original Message----- > From: Perrin Harkins [mailto:[EMAIL PROTECTED] > Sent: Thursday, 9 October 2003 2:22 PM > To: Morton-Allen, Matt > Cc: Matisse Enzer; [EMAIL PROTECTED] > Subject: RE: Using class wide variables under mod_perl is safe? > > On Wed, 2003-10-08 at 23:33, Morton-Allen, Matt wrote: > > I am using Apache::DBI and would be at a loss without it (we use Oracle > > so the connection overhead is serious). With this in mind is it such a > > bad thing to hang on to the one handle for an execution? > > No that's fine. I suggest you do it with $r->pnotes(), so that you can > be sure it will be cleaned up at the end of the request. > > > Secondly, I use the inherited object to achieve some other general but > > more complicated logic like logging and error handling which might be > > why it looked like overkill. > > I still wouldn't use inheritance here. Inheritance is not a good way to > add functions like this. That's better done by making separate packages > and calling them, and saving inheritance for things where "is a" > literally sounds right. But I don't know the whole story with your data > model, so I won't lecture you about it. > > > I assume that if I did continue to use inheritance then the following > > (after the inheritance was set up) would be the equivalent and that the > > fact it was inherited at the time will make no difference? > > > > my $parser = $self->get_parser(); > > Yes. > > > There's newlines outputted at the very end of the execution so it's > > confusing as to how the top line which is part of the process for the > > previous hit can be a) output after the newlines and b) output only > > after the next request is made. I tried turning off buffering to no > > avail. > > This buffering is probably done by apache's logging functions. I don't > think you can affect it. > > - Perrin
