Proposal: mod_setenvif and mod_browser

[Rodent of Unusual Size <co...@decus.org>]

    I've been looking at Paul's setenvif module, and I'd like to
    re-propose something that was mentioned 'way back at the beginning
    of the [calendar] year.

    BrowserMatch[NoCase] is a special case of what mod_setenvif does
    (set envariables based upon header-field values).  I would like to
    see mod_browser replaced with mod_setenvif, and the "BrowserMatch*"
    directives added to setenvif as special cases for compatibility's
    sake.

    I've done some fiddling with mod_setenvif to make it use RAW_ARGS
    rather than a new ITERATE3 argument type, and added a UnSetEnvIfZero
    flag directive to control whether "banned=0" means banned has a
    value or is un-defined.  (I agree with Paul that the current
    treatment of zero as true is a little ambiguous.)

    Comments?  I think being able to set envariables based upon *any* of
    the header-fields would be a big win.

    #ken    :-)}

Re: Proposal: mod_setenvif and mod_browser

[Dean Gaudet <dg...@arctic.org>]

+1 on mod_setenvif replacing mod_browser, I was going to suggest it too. 

+0 on UnSetEnvIfZero.

Dean

On Tue, 29 Jul 1997, Rodent of Unusual Size wrote:

>     I've been looking at Paul's setenvif module, and I'd like to
>     re-propose something that was mentioned 'way back at the beginning
>     of the [calendar] year.
> 
>     BrowserMatch[NoCase] is a special case of what mod_setenvif does
>     (set envariables based upon header-field values).  I would like to
>     see mod_browser replaced with mod_setenvif, and the "BrowserMatch*"
>     directives added to setenvif as special cases for compatibility's
>     sake.
> 
>     I've done some fiddling with mod_setenvif to make it use RAW_ARGS
>     rather than a new ITERATE3 argument type, and added a UnSetEnvIfZero
>     flag directive to control whether "banned=0" means banned has a
>     value or is un-defined.  (I agree with Paul that the current
>     treatment of zero as true is a little ambiguous.)
> 
>     Comments?  I think being able to set envariables based upon *any* of
>     the header-fields would be a big win.
> 
>     #ken    :-)}
>