I've decided that I would understand this stuff a lot better if I had a clearer sense of what happens along the way, from your computer to the page your browser shows you. Apache is fine on the hair-splitting details, but not so great when it comes to putting things into words of two syllables.
At the beginning, you ask your browser to give you such-and-such URL.
The domain name goes the rounds of DNSs until it meets one that says "Sure, I know him, he lives at 12.34.56.78". Or not, in which case the browser puts up a rude message suggesting that you can't type, or the cat stepped on your router switch.
The request now dashes off to 12.34.56.78 and knocks on the door, where it waits for an answer. In the two browsers that I tried (using an URL that I knew would time out), it waits 75 seconds; I couldn't find a place to change it. If no answer after 75 seconds, the browser puts up another message, this time with rude comments along the lines of "I haven't got all day, you know". I don't know what happens if the browser's time limit is longer than the time limit at the far end. (Apache's default is 300 seconds.) Rude message from Apache instead?
OK... now what happens?
If you're looking for a particular file, does the config file tell you upfront if it doesn't exist, or do you have to go in and look? Along the way, do you have to obey any htaccess instructions you happen to meet, whether or not your requested file is where it's supposed to be?
If the Directory Slash Redirect is enabled, do you pick up the slash immediately, or do you have to look for yourself and confirm that your destination is a directory? (Apache docs seem to imply that this is already known, though they don't say by whom. All I know for sure is that it doesn't come through as a 301 the way a server-level www redirect does.)
If the request is for a directory, does the config file say "Go on in, you're looking for a file called index.htm"? Or does it give you a list of possibilities? (In the case of my server, the request must have picked up this information before it reaches the top-level htaccess, because the rules only work for \.html)
What if there isn't a named Index file? Does the config file say "Here's your auto-generated Index, you'll need it"? Or do you go in, look around, and then go all the way back and report that there's nobody there, but along the way you passed an htaccess file that said Options +Indexes, so I'll take an auto-index now if it isn't too much trouble? Or do you have to keep running back and forth until you've found out whether there's a FallbackResource (mod_dir), and only then get to ask if mod_autoindex even exists?
Now, about those htaccess files.
Does the request have to go all the way up to its destination on the off chance that there might be an htaccess file containing SetEnvIf instructions, and only then go back to the beginning to pick up Rewrites, and then look for Options and so on? Or does the server tell you what's where?
If you meet a RewriteRule that ends in [F], do you not pass go, not collect $200, and proceed directly to the 403 page, or are there further detours before you are allowed to drop dead?
If you get redirected within the same named domain, how far back do you go? Do you knock on the 12.34.56.78 door all over again? Or do you get kicked back to the server? to the lobby of the userspace (assuming shared hosting)? to the top level of the domain?
What if you get rewritten? Where do you go?
And, finally, what about all those core-level Admits and Denys? Do those happen up front, or only after you've jumped through the whole htaccess alphabet? Or both before and after, to allow for mod_setenvif activity?
