Hi all, On Mon, 24 Nov 2003, Geoffrey Young wrote: > [snip, snip] > Elizabeth Mattijsen wrote: > > > > What _I_ am interested in, is a flag with which I could see whether it > > would make sense to load as much as possible, rather than on demand. > > #ifdef USE_ITHREADS > static void modperl_init_clones(server_rec *s, apr_pool_t *p) > > I'm more in favor of the > Apache->mpm_is_threaded;
I'm a little concerned by two things here. 1. The, er, blind reliance on the mantra that because of a, b and c then it makes sense to do d. It must be safer to instrument the system and see how it behaves under real conditions than do something just because of the folklore. Having read some of Elizabeth's posts in the past I'm not suggesting that she would fail to investigate if necessary, but there are a couple of thousand people reading this List who, er, seem to just read it. 2. The fact that we might have to worry at all about whether we're using a pool of interpreters or not. There are (as I've previously opined on this List) already too many things to trip up the unwary Perl scripter in a mod_perl system. I feel that rather than providing more and more API to debug, document and confuse there should be a concerted effort applied to reducing the differences perceptible to a Perl script between its environment and behaviour when run from the command line and those when it runs as the result of some query made on a Web server. Granted there must be differences; but if one has to just 'use ModPerl;' or something similar and at worst the interpreter says something like "Variable will not stay shared..." then I believe that there will be a much greater acceptance of the package as a whole and, if that's a Good Thing, that other Good Things will follow. This is about design, not about implementation (so maybe it's on the wrong List, and maybe I'm talking about mod_perl version 3:), and here performance should take a back seat - next year your hardware is going to knock most of the software performance improvements into a cocked hat anyway, and you'll *still* be debugging it. Having said that I'll go back to my cave before anyone asks me what I'm going to do about it... :) 73, Ged. PS: I like the one about the hammer and the duct tape. -- Reporting bugs: http://perl.apache.org/bugs/ Mail list info: http://perl.apache.org/maillist/modperl.html
