Nils Goroll
Nils Goroll
FTR: My concern is that backends might send no body in error with the valid Etag/LM. In that case, the current code would avoid the error to be exposed for...
@rezan would it be possible to get the response times for a statistically relevant number, like 1000 requests? Also, could you clarify, is AWS pricing different based on the response...
Thank you, @rezan . That would not appear to be a relevant difference, or would it?
Bugwash discussed this ticket again: WONTFIX because no valid case was shown and we seem to all agree that a 304 for an empty body is pointless.
first reaction (I have a strong feeling that I might be overlooking something and might need to add more later): I have also seen the `ttl=0` case in production and...
sorry, I had hit the wrong button
> Any idea how many places in the code it could have been used already ? I have pushed a commit changing five places in the tree, but the macro...
> Did you consider a coccinelle patch? Yes, but I deemed the patterns too different - for the given example https://github.com/varnishcache/varnish-cache/pull/3815/commits/afe8b2ca8fce84043b9467753097d519bbb5817e there are three different ways of error handling.
see also #3017 : It appears to me that this PR should be dropped in favor of the new error handling agreed at VDD19Q3
I think this is another instance of #3101