Re: [mod-xslt dev] patch: add environment variable override

["Denis F. Latypoff" <latypoff@xxxxxxxxx>]

Hello colin,

Tuesday, January 30, 2007, 1:36:33 AM, you wrote:

> The patch below adds SetEnvIf environment variable support that can
> be used to disable mod_xslt. Effectively, this can be used to mimics
> the functionality of XSLTAddRule but can also be used for
> dynamically assigned xslt. With this patch you can AddOutputFilter
> and then disable it based on any criteria necessary. This allows for
> more customized control over the execution of the mod_xslt module.

> For example:
> <Location ~ "\.php|/$|^[^.]*$">
>         SetOutputFilter mod-xslt
> </Location>
> SetEnvIf Request_URI /Rss.php no-xslt

I am using 'AddOutputFilterByType mod-xslt text/xml'
and when my apps sends 'ContentType: text/xml' mod_xslt is processing
output document, so I can easily change content type to any other and
see non-processed document (it's for debug mainly). I don't use any
exotic mod-xslt's commands such as XSLTAddRule etc.

> This will enable the OutputFilter for all .php files (incl welcome
> pages) but disable it for the rss.php. While this is a primitive
> example, my scenario required adjustment based on Headers and
> parameters which is much more complicated for <Location> matches.

> Also this patch also moves the removal of the Content-Length to
> after the quick exit block. This is important so that pipelining is
> preserved even if the mod_xslt is dynamically removed from the chain.

> --- modxslt-2005072700/sapi/apache2/modxslt.c.orig   2005-07-27 
> 04:27:57.000000000 -0600
> +++ modxslt-2005072700/sapi/apache2/modxslt.c        2007-01-24 
> 15:03:35.000000000 -0600
> @@ -138,7 +138,8 @@
>    const char * forcestyle, * defaultstyle;
>    void * get;

> -  apr_table_unset(f->r->headers_out, "Content-Length");
> +  if (apr_table_get(f->r->subprocess_env, "no-xslt"))
> +    return ap_pass_brigade(f->next, brigade);

>    if(f->r->header_only)
>      return ap_pass_brigade(f->next, brigade);
> @@ -148,6 +149,8 @@
>    if(f->r->status != HTTP_OK)
>      return ap_pass_brigade(f->next, brigade);

> +  apr_table_unset(f->r->headers_out, "Content-Length");
> +
>    data=(apr_bucket_brigade *)f->ctx;
>    ap_save_brigade(f, &data, &brigade, f->r->pool);
>    f->ctx=data;

thanks for the patch, it's really useful ;)

> __________ NOD32 2010 (20070126) Information __________

> This message was checked by NOD32 antivirus system.
>   part000.txt - is OK

> http://www.eset.com




-- 
Best regards,
 Denis Latypoff                          mailto:latypoff@xxxxxxxxx