On Tue, Apr 7, 2020 at 5:22 PM Filip Hanik <fha...@pivotal.io> wrote:
> On Tue, Apr 7, 2020 at 8:15 AM Rémy Maucherat <r...@apache.org> wrote:
>
>> On Tue, Apr 7, 2020 at 10:16 AM Mark Thomas <ma...@apache.org> wrote:
>>
>>> > Noted, I think a compromise may be in order. Where we simply add a
>>> > constructor that avoids the Class.forName
>>> > and that allows the developer to explicitly invoke a constructor that
>>> > avoids it.
>>>
>>> My thinking was more along the following lines...
>>>
>>> Rewrite the Connector (and possibly some/all other components) so that
>>> if a known class name was specified, rather than using the specified
>>> name directly via reflection we could do something like:
>>>
>>> if ("org.apache.coyote.http11.Http11NioProtocol".equals(className)) {
>>> protocol = new Http11NioProtocol();
>>> } else if ("org...
>>> ...
>>> } else {
>>> // OK. Unrecognised class name. Use reflection
>>> ...
>>> }
>>>
>>> Would that style of approach help at all?
>>>
>>
>> Proposed patch:
>> https://pastebin.com/ydCYnBD9
>>
>
> That patch works, I would include the check for the full classname too
>
> Was: if ("HTTP/1.1".equals(protocol) || protocol == null)
> Could be: if (protocol == null || "HTTP/1.1".equals(protocol) ||
> "HTTP/1.1".equals(Http11NioProtocol.class.getName))
>
> However, what still is awkward is that Connector.java deals with the
> interface ProtocolHandler.
> The loading of both the Http11NioConnector and the Ajp13NioConnector seems
> to defy the purpose of having that interface being present
>
> Is: protected final ProtocolHandler protocolHandler;
>
> Does the connector need to know about the actual implementations?
>
Ideally no, but it removes the reflection you say is bad for Graal.
Rémy
>
> Filip
>
>
>
>
>>
>> Rémy
>>
>>
>>>
>>> Mark
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
>>> For additional commands, e-mail: dev-h...@tomcat.apache.org
>>>
>>>
