From: "Ian Holsman" <[EMAIL PROTECTED]> Sent: Wednesday, January 09, 2002 9:51 PM
> we did a LOT of testing before we decided SetEnvIf really really sucked > in 1.3 and wrote the replacement. > > the config change actually came from our custom module, I'm trying to > convince the developers over here to use the standard one, but wrowe's > veto nixed that. Ian, a much bigger point is; we have a method of Logging and including or excluding entries based on variables. Perhaps too flexibly, and perhaps not considering enough variables. And rather than clutter up the list of variables, perhaps we focus on more expliit directives to do something similar to what you proposed by family, such as input header fields, output headers (like mime) or whatnot. Although I expressed my opinion as -1, if not a veto, you already have several others who think this method is duplicitous and yet too narrow to be generally useful. Rather than fix the directives (SetEnvIf and it's testing logic) we are adding a one-shot-single-use directive (much like BrowserMatch) that isn't flexible enough to do much, but crosses over the directives that exist. Put this patch in status for a vote with the message ids, you know three of us are leaning -1, although I have not explicitly vetoed the patch, but add Lars' comments, his short paragraph explained my opinion perfectly. See what other opinions come up in a few days, and consider how to address Lars' short and to the point comment; it's the same issue that Josh and I share. Then consider how to be more diplomatic than pointing at 'wrowe's or 'lars's veto. It's a debate of how to approach your legit gripe with setenvif, not personalities; I have the same issue with the DocumentRoot in a <Location> directive debate. I'm waiting until I can stop, clearly explain the benefits of the DocumentRoot change (probably choosing a different name for the Directive) and then let folks comment again when they have more details to chew on. Study the objection and see if you can refocus your patch. Bill
