Maybe my mistake is to use getResponseHeader function rather than
getHeader, the latter returns the headers of shidig cache?.
Emilio
On Wed, Jul 2, 2008 at 2:41 PM, Chris Chabot <[EMAIL PROTECTED]> wrote:
> Thats not the only thing though .. there's also the case that the remote
> server didn't set a Last-Modified header, so the check for ifModified and
> ifNoneMatch still fails, since there's no last-modified to check against.
>
> Maybe that's ok .. but i'm trying to find out if we can't handle that better
> our selves, would be a shame to loose being able to cache efficiently
> because someone didn't configure their server correctly
>
> -- Chris
>
> On Jul 2, 2008, at 7:22 PM, Emilio Daniel González wrote:
>
>> jaja! sorry, i forgot modify the function in the last code from svn.
>>
>> Can you fix that?, here you got the "real" function:
>>
>> private function setCachingHeaders($etag = null)
>> {
>> // TODO: Re-implement caching behavior if appropriate.
>> header("Cache-Control: private; max-age=0", true);
>> header("Expires: " . gmdate("D, d M Y H:i:s", time() -
>> 3000) . " GMT", true);
>> if ($etag !== null)
>> header("ETag: $etag", true);
>> }
>>
>> cheers!
>>
>> On Wed, Jul 2, 2008 at 2:15 PM, Chris Chabot (JIRA) <[EMAIL PROTECTED]>
>> wrote:
>>>
>>> [
>>> https://issues.apache.org/jira/browse/SHINDIG-374?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12609978#action_12609978
>>> ]
>>>
>>> Chris Chabot commented on SHINDIG-374:
>>> --------------------------------------
>>>
>>> Hi Emilio,
>>>
>>> I noticed something funny about this patch:
>>>
>>> $this->setCachingHeaders($etag);
>>> ..
>>> private function setCachingHeaders()
>>> {
>>> // TODO: Re-implement caching behavior if appropriate.
>>> header("Cache-Control: private; max-age=0", true);
>>> header("Expires: " . gmdate("D, d M Y H:i:s", time() - 3000) . " GMT",
>>> true);
>>> }
>>>
>>> In other words the e-tag is never set, so the code never could have been
>>> tested .... And really i much prefer to see patches that you at least tried
>>> out in a few different browsers, and actually ran your self :)
>>>
>>> I'm taking a look at the code now but it seems i'll have to modify it a
>>> good bit to cover all the border cases.
>>>
>>>> Proxy ignores e-tags / if-modified-since
>>>> ----------------------------------------
>>>>
>>>> Key: SHINDIG-374
>>>> URL: https://issues.apache.org/jira/browse/SHINDIG-374
>>>> Project: Shindig
>>>> Issue Type: Bug
>>>> Components: Gadget Rendering Server (PHP)
>>>> Reporter: Chris Chabot
>>>> Attachments: shindig-http-304-support.patch
>>>>
>>>>
>>>> The ProxyHandler class does it's own header control instead of using the
>>>> generic Servlet one because it requires full control over what headers are
>>>> send.
>>>> However this currently means it doesn't honor if-modified-since and
>>>> e-tag headers, which causes the proxied files to always be re-downloaded by
>>>> the web browser.
>>>> On straight get's it should really:
>>>> - If if-modified-since is set, and the cached file time is within the
>>>> refreshInterval params value, return not-modified
>>>> - Create it's own e-tag's (md5(responseBody)), and if the e-tag's match
>>>> return a not-modified
>>>
>>> --
>>> This message is automatically generated by JIRA.
>>> -
>>> You can reply to this email to add a comment to the issue online.
>>>
>>>
>
>
