Hi Stuart, 

Thanks for your response. It worked.

Best regards,
Antoine

On Mon, Oct 10, 2016 at 9:13 AM, Stuart Douglas <sdouglas@redhat.com> wrote:
The max-content-size predicate will always return false for chunked
encoding. You need something like 'exists(%{o,Transfer-encoding}) or
max-content-size(5)'.

Stuart

On Sun, Oct 9, 2016 at 10:55 PM, Antoine Girard
<antoine.girard@ymail.com> wrote:
> Dear Undertow devs,
>
> I originally posted the below question on the Wildfly forum and was advised
> to do it here instead.
>
> I am using the Undertow EncodingHandler with a Gzip Encoding provider, as
> follows (sample code taken from an Undertow unit test):
>
>
>
> final EncodingHandler handler =
>     new EncodingHandler(new ContentEncodingRepository()
>       .addEncodingHandler("gzip",
>           new GzipEncodingProvider(), 50,
>           Predicates.parse("max-content-size[5]")))
>       .setNext(path);
>
>
>
> It works fine for very small response bodies, however, when Undertow starts
> to use the Chunked Transfer-Encoding, the Gzip compression is simply
> ignored.
>
> My question is basically: how to apply a Gzip encoding (or other
> compression) before Undertow starts chunking the response?
>
>
> Could it be a bug?
>
>
> Also, and I could not really track it down from the source code, how does
> Undertow decide to start chunking the response?
>
>
> Thanks in advance!
>
>
> Cheers,
>
> Antoine
>
>
>
> _______________________________________________
> undertow-dev mailing list
> undertow-dev@lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/undertow-dev