On 3/7/06, Jim Jagielski <[EMAIL PROTECTED]> wrote: > Feh. No way. The argument would hold true if mod_jk > did not provide just the sort of control that you say > we shouldn't provide to users/admins, and control that > *varies* from setup to setup.
Why must we offer a shiny knob that the average admin will have no real hope of understanding why it is there or when to enable it? As witnessed the JIRA issue, certain people (who should have known better) were extremely quick to blame all errors on the flushing code - when the flushing code was working exactly as expected. > Yes, the protocol itself is "incomplete" but a the current, > non-changable setup is not the correct way to address it. Until the protocol is fixed, we should do the right thing - and that means we shouldn't ever allow the entire response to be spooled in memory. -- justin
