Jetty ALPN (was: Roadmap 2016/2017)

[sebb <se...@gmail.com>]

This is straying from the original topic

On 23 August 2016 at 21:58, Philippe Mouawad <ph...@gmail.com> wrote:
> Regarding jetty-alpn:
>
> The ALPN implementation relies on modification of a few OpenJDK classes and
> on a few new classes that need to live in the sun.security.ssl package.
> These classes are released under the same GPLv2+exception license of
> OpenJDK. The ALPN class and its nested classes are released under same
> license as the classes of the Jetty project.

Which means it is not suitable for JMeter, because the sun.* classes
don't exist in all JVMs.

We cannot rely on a particular implementation of Java.

> Regards
>
> On Tue, Aug 23, 2016 at 9:05 PM, Philippe Mouawad <
> philippe.mouawad@gmail.com> wrote:
>
>> Hello,
>> AFAICR all mentioned dependencies are Apache Licensed:
>> - Open SSL
>> - netty-tcnative
>>
>> I opened an issue for the jetty-alpn (which AFAIU is not the most
>> performing one):
>> https://github.com/jetty-project/jetty-alpn/issues/10
>>
>> Regards
>> Philippe
>>
>> On Fri, Aug 19, 2016 at 12:32 PM, sebb <se...@gmail.com> wrote:
>>
>>> It looks like Netty has additional requirements for HTTP/2 with TLS
>>>
>>> http://netty.io/wiki/requirements-for-4.x.html
>>>
>>> We need to ensure that these additional dependencies are available
>>> under a suitable license before considering whether to use it for
>>> JMeter.
>>>
>>> On 19 August 2016 at 10:24, Philippe Mouawad <ph...@gmail.com>
>>> wrote:
>>> > For info:
>>> > Http2Client :
>>> > - http://netty.io/5.0/xref/io/netty/example/http2/client/Http2
>>> Client.html
>>> >
>>> > Regards
>>> >
>>> > On Fri, Aug 19, 2016 at 9:45 AM, Philippe Mouawad <
>>> > philippe.mouawad@gmail.com> wrote:
>>> >
>>> >>
>>> >>
>>> >> On Fri, Aug 19, 2016 at 12:19 AM, sebb <se...@gmail.com> wrote:
>>> >>
>>> >>> On 18 August 2016 at 22:47, Philippe Mouawad <
>>> philippe.mouawad@gmail.com>
>>> >>> wrote:
>>> >>> > Hello,
>>> >>> > @sebb, as chief of PMC, could you expose your vision of JMeter's
>>> roadmap
>>> >>> > and future ?
>>> >>>
>>> >>> Yes, I am the chair of the PMC, but that is not the same as a chief.
>>> >>> The chair is not a technical role, it is purely administrative.
>>> >>> As such the chair has the same standing as any other PMC member.
>>> >>>
>>> >>> ==
>>> >>>
>>> >> It is good to know that
>>> >>
>>> >>
>>> >>>
>>> >>> In the short term, I would like to see the HC4 code fixed to remove
>>> >>> all the deprecations.
>>> >>>
>>> >> +1
>>> >>
>>> >>>
>>> >>> HTTP2 can wait until there is a suitable library.
>>> >>>
>>> >>
>>> >> There is already some, see:
>>> >>
>>> >> https://github.com/http2/http2-spec/wiki/Implementations
>>> >>
>>> >> the most famous being netty no ?
>>> >>
>>> >>
>>> >>
>>> >>> It would be nice to fix undo/redo, but as I suspected this is going to
>>> >>> be really difficult.
>>> >>>
>>> >> +1
>>> >>
>>> >>>
>>> >>> As a general rule, we should be cautious about adding code to support
>>> >>> infrequent use cases, especially where there is a reasonable
>>> >>> alternative.
>>> >>>
>>> >>
>>> >> Ok if it's not a matter of living with broken features just to avoid
>>> >> breaking changes.
>>> >> It's not because there is a workaround or a complex way to do something
>>> >> that we should not try to make it easier.
>>> >>
>>> >>
>>> >>>
>>> >>> > Thank you
>>> >>> >
>>> >>> > On Thursday, July 28, 2016, Philippe Mouawad <
>>> >>> philippe.mouawad@gmail.com>
>>> >>> > wrote:
>>> >>> >
>>> >>> >>
>>> >>> >>
>>> >>> >> On Thursday, July 28, 2016, Felix Schumacher <felix.schumacher@
>>> >>> >> internetallee.de
>>> >>> >> <javascript:_e(%7B%7D,'cvml','felix.schumacher@internetallee.de
>>> ');>>
>>> >>> >> wrote:
>>> >>> >>
>>> >>> >>> Am 27.07.2016 um 14:41 schrieb Philippe Mouawad:
>>> >>> >>>
>>> >>> >>>> Hello,
>>> >>> >>>> I think we should work on a roadmap for JMeter to ensure:
>>> >>> >>>> - we prioritize some urgent work on it.
>>> >>> >>>> - we give visibility on future of JMeter to users
>>> >>> >>>>
>>> >>> >>>>
>>> >>> >>>> I see the following things from most to less urgent:
>>> >>> >>>>
>>> >>> >>>> - http2 support. We depend on httpclient for this
>>> >>> >>>>
>>> >>> >>> +1 but I think it is httpclient or jdk where the work is.
>>> >>> >>
>>> >>> >>
>>> >>> >> Afaik, it's a priority of Httpclient no?
>>> >>> >> If not, maybe we should look at other options although my
>>> preference
>>> >>> >> clearly goes to hc for simplicity and uniformity.
>>> >>> >>
>>> >>> >> - possible rework of core architecture to at least introduce a
>>> pool of
>>> >>> >>>> threads or switch to async model allowing us to take advantage of
>>> >>> async
>>> >>> >>>> io
>>> >>> >>>>
>>> >>> >>> lot of work (I think really a lot of work), but it is probably a
>>> >>> worthy
>>> >>> >>> goal for the long run
>>> >>> >>
>>> >>> >> I think so
>>> >>> >>
>>> >>> >>>
>>> >>> >>> - start a migration to JavaFX , a good opportunity would be to
>>> >>> replace the
>>> >>> >>>> old browser used for html rendering
>>> >>> >>>>
>>> >>> >>> replacement of the browser component with javafx is a good thing.
>>> >>> >>
>>> >>> >> Yes that was my main intention. I don't think it's a big deal.
>>> >>> >>
>>> >>> >>>
>>> >>> >>> Replacement of every thing? I don't know. Will javafx really be
>>> the
>>> >>> next
>>> >>> >>> big java gui? Would it be worth trying to get to a html/http gui
>>> and
>>> >>> get
>>> >>> >>> rid of swing completely?
>>> >>> >>
>>> >>> >>
>>> >>> >> It would be a webapp ?  How do you see it ?
>>> >>> >>
>>> >>> >>>
>>> >>> >>> For the near future:
>>> >>> >>>
>>> >>> >>> * get a bugfix release of 3.0
>>> >>> >>
>>> >>> >> I would like to commit an enhancement to generate reports from gui
>>> >>> before.
>>> >>> >>
>>> >>> >>
>>> >>> >>> * complete migration to httpclients new api
>>> >>> >>
>>> >>> >> +1
>>> >>> >>
>>> >>> >>> * make recording of jsf sites easier
>>> >>> >>
>>> >>> >> maybe introduce more generally a Framework correlator where jsf
>>> would
>>> >>> be 1
>>> >>> >> implementation.
>>> >>> >>
>>> >>> >>
>>> >>> >>> * discuss replacement of logging framework
>>> >>> >>
>>> >>> >> +1
>>> >>> >>
>>> >>> >>> * look at memory consumption of the tree view listener
>>> >>> >>>   ideas where:
>>> >>> >>>     - store only the last/first X entries
>>> >>> >>>     - store only marked entries (might be a header field)
>>> >>> >>>
>>> >>> >>>
>>> >>> >>>
>>> >>> >>>> As side features:
>>> >>> >>>> - DSL ?
>>> >>> >>>>
>>> >>> >>> nice idea, but this is a lot of work
>>> >>> >>
>>> >>> >>
>>> >>> >>
>>> >>> >>> - JSON format instead of XMLfor jmx plans if dsl is too heavy
>>> change ?
>>> >>> >>>>
>>> >>> >>> I think xml has served us well. Every other format has to prove,
>>> that
>>> >>> it
>>> >>> >>> really can compete.
>>> >>> >>>
>>> >>> >>>> - Fix undo /redo feature
>>> >>> >>>>
>>> >>> >>> +1 for looking at undo/redo
>>> >>> >>
>>> >>> >> I would really love to see this one fixed or dropped.
>>> >>> >> But I don't want to spend too much energy on it.
>>> >>> >>
>>> >>> >>
>>> >>> >>
>>> >>> >>
>>> >>> >>>
>>> >>> >>> Regards,
>>> >>> >>>  Felix
>>> >>> >>>
>>> >>> >>>> - ...
>>> >>> >>>>
>>> >>> >>>> Ideas welcome
>>> >>> >>>> Regards
>>> >>> >>>> Philippe M.
>>> >>> >>>>
>>> >>> >>>>
>>> >>> >>>>
>>> >>> >>>>
>>> >>> >>>
>>> >>> >>
>>> >>> >> --
>>> >>> >> Cordialement.
>>> >>> >> Philippe Mouawad.
>>> >>> >>
>>> >>> >>
>>> >>> >>
>>> >>> >>
>>> >>> >
>>> >>> > --
>>> >>> > Cordialement.
>>> >>> > Philippe Mouawad.
>>> >>>
>>> >>
>>> >>
>>> >>
>>> >> --
>>> >> Cordialement.
>>> >> Philippe Mouawad.
>>> >>
>>> >>
>>> >>
>>> >
>>> >
>>> > --
>>> > Cordialement.
>>> > Philippe Mouawad.
>>>
>>
>>
>>
>> --
>> Cordialement.
>> Philippe Mouawad.
>>
>>
>>
>
>
> --
> Cordialement.
> Philippe Mouawad.

Re: Jetty ALPN (was: Roadmap 2016/2017)

[Philippe Mouawad <ph...@gmail.com>]

On Tuesday, August 23, 2016, sebb <se...@gmail.com> wrote:

> This is straying from the original topic
>
> On 23 August 2016 at 21:58, Philippe Mouawad <philippe.mouawad@gmail.com
> <javascript:;>> wrote:
> > Regarding jetty-alpn:
> >
> > The ALPN implementation relies on modification of a few OpenJDK classes
> and
> > on a few new classes that need to live in the sun.security.ssl package.
> > These classes are released under the same GPLv2+exception license of
> > OpenJDK. The ALPN class and its nested classes are released under same
> > license as the classes of the Jetty project.
>
> Which means it is not suitable for JMeter, because the sun.* classes
> don't exist in all JVMs.
>
> We cannot rely on a particular implementation of Java.

Anyway it's not the recommended approach by netty , it's the alternative.
The first option is compatible in terms of licenses.



>
> > Regards
> >
> > On Tue, Aug 23, 2016 at 9:05 PM, Philippe Mouawad <
> > philippe.mouawad@gmail.com <javascript:;>> wrote:
> >
> >> Hello,
> >> AFAICR all mentioned dependencies are Apache Licensed:
> >> - Open SSL
> >> - netty-tcnative
> >>
> >> I opened an issue for the jetty-alpn (which AFAIU is not the most
> >> performing one):
> >> https://github.com/jetty-project/jetty-alpn/issues/10
> >>
> >> Regards
> >> Philippe
> >>
> >> On Fri, Aug 19, 2016 at 12:32 PM, sebb <sebbaz@gmail.com <javascript:;>>
> wrote:
> >>
> >>> It looks like Netty has additional requirements for HTTP/2 with TLS
> >>>
> >>> http://netty.io/wiki/requirements-for-4.x.html
> >>>
> >>> We need to ensure that these additional dependencies are available
> >>> under a suitable license before considering whether to use it for
> >>> JMeter.
> >>>
> >>> On 19 August 2016 at 10:24, Philippe Mouawad <
> philippe.mouawad@gmail.com <javascript:;>>
> >>> wrote:
> >>> > For info:
> >>> > Http2Client :
> >>> > - http://netty.io/5.0/xref/io/netty/example/http2/client/Http2
> >>> Client.html
> >>> >
> >>> > Regards
> >>> >
> >>> > On Fri, Aug 19, 2016 at 9:45 AM, Philippe Mouawad <
> >>> > philippe.mouawad@gmail.com <javascript:;>> wrote:
> >>> >
> >>> >>
> >>> >>
> >>> >> On Fri, Aug 19, 2016 at 12:19 AM, sebb <sebbaz@gmail.com
> <javascript:;>> wrote:
> >>> >>
> >>> >>> On 18 August 2016 at 22:47, Philippe Mouawad <
> >>> philippe.mouawad@gmail.com <javascript:;>>
> >>> >>> wrote:
> >>> >>> > Hello,
> >>> >>> > @sebb, as chief of PMC, could you expose your vision of JMeter's
> >>> roadmap
> >>> >>> > and future ?
> >>> >>>
> >>> >>> Yes, I am the chair of the PMC, but that is not the same as a
> chief.
> >>> >>> The chair is not a technical role, it is purely administrative.
> >>> >>> As such the chair has the same standing as any other PMC member.
> >>> >>>
> >>> >>> ==
> >>> >>>
> >>> >> It is good to know that
> >>> >>
> >>> >>
> >>> >>>
> >>> >>> In the short term, I would like to see the HC4 code fixed to remove
> >>> >>> all the deprecations.
> >>> >>>
> >>> >> +1
> >>> >>
> >>> >>>
> >>> >>> HTTP2 can wait until there is a suitable library.
> >>> >>>
> >>> >>
> >>> >> There is already some, see:
> >>> >>
> >>> >> https://github.com/http2/http2-spec/wiki/Implementations
> >>> >>
> >>> >> the most famous being netty no ?
> >>> >>
> >>> >>
> >>> >>
> >>> >>> It would be nice to fix undo/redo, but as I suspected this is
> going to
> >>> >>> be really difficult.
> >>> >>>
> >>> >> +1
> >>> >>
> >>> >>>
> >>> >>> As a general rule, we should be cautious about adding code to
> support
> >>> >>> infrequent use cases, especially where there is a reasonable
> >>> >>> alternative.
> >>> >>>
> >>> >>
> >>> >> Ok if it's not a matter of living with broken features just to avoid
> >>> >> breaking changes.
> >>> >> It's not because there is a workaround or a complex way to do
> something
> >>> >> that we should not try to make it easier.
> >>> >>
> >>> >>
> >>> >>>
> >>> >>> > Thank you
> >>> >>> >
> >>> >>> > On Thursday, July 28, 2016, Philippe Mouawad <
> >>> >>> philippe.mouawad@gmail.com <javascript:;>>
> >>> >>> > wrote:
> >>> >>> >
> >>> >>> >>
> >>> >>> >>
> >>> >>> >> On Thursday, July 28, 2016, Felix Schumacher <felix.schumacher@
> >>> >>> >> internetallee.de
> >>> >>> >> <javascript:_e(%7B%7D,'cvml','felix.schumacher@internetallee.de
> <javascript:;>
> >>> ');>>
> >>> >>> >> wrote:
> >>> >>> >>
> >>> >>> >>> Am 27.07.2016 um 14:41 schrieb Philippe Mouawad:
> >>> >>> >>>
> >>> >>> >>>> Hello,
> >>> >>> >>>> I think we should work on a roadmap for JMeter to ensure:
> >>> >>> >>>> - we prioritize some urgent work on it.
> >>> >>> >>>> - we give visibility on future of JMeter to users
> >>> >>> >>>>
> >>> >>> >>>>
> >>> >>> >>>> I see the following things from most to less urgent:
> >>> >>> >>>>
> >>> >>> >>>> - http2 support. We depend on httpclient for this
> >>> >>> >>>>
> >>> >>> >>> +1 but I think it is httpclient or jdk where the work is.
> >>> >>> >>
> >>> >>> >>
> >>> >>> >> Afaik, it's a priority of Httpclient no?
> >>> >>> >> If not, maybe we should look at other options although my
> >>> preference
> >>> >>> >> clearly goes to hc for simplicity and uniformity.
> >>> >>> >>
> >>> >>> >> - possible rework of core architecture to at least introduce a
> >>> pool of
> >>> >>> >>>> threads or switch to async model allowing us to take
> advantage of
> >>> >>> async
> >>> >>> >>>> io
> >>> >>> >>>>
> >>> >>> >>> lot of work (I think really a lot of work), but it is probably
> a
> >>> >>> worthy
> >>> >>> >>> goal for the long run
> >>> >>> >>
> >>> >>> >> I think so
> >>> >>> >>
> >>> >>> >>>
> >>> >>> >>> - start a migration to JavaFX , a good opportunity would be to
> >>> >>> replace the
> >>> >>> >>>> old browser used for html rendering
> >>> >>> >>>>
> >>> >>> >>> replacement of the browser component with javafx is a good
> thing.
> >>> >>> >>
> >>> >>> >> Yes that was my main intention. I don't think it's a big deal.
> >>> >>> >>
> >>> >>> >>>
> >>> >>> >>> Replacement of every thing? I don't know. Will javafx really be
> >>> the
> >>> >>> next
> >>> >>> >>> big java gui? Would it be worth trying to get to a html/http
> gui
> >>> and
> >>> >>> get
> >>> >>> >>> rid of swing completely?
> >>> >>> >>
> >>> >>> >>
> >>> >>> >> It would be a webapp ?  How do you see it ?
> >>> >>> >>
> >>> >>> >>>
> >>> >>> >>> For the near future:
> >>> >>> >>>
> >>> >>> >>> * get a bugfix release of 3.0
> >>> >>> >>
> >>> >>> >> I would like to commit an enhancement to generate reports from
> gui
> >>> >>> before.
> >>> >>> >>
> >>> >>> >>
> >>> >>> >>> * complete migration to httpclients new api
> >>> >>> >>
> >>> >>> >> +1
> >>> >>> >>
> >>> >>> >>> * make recording of jsf sites easier
> >>> >>> >>
> >>> >>> >> maybe introduce more generally a Framework correlator where jsf
> >>> would
> >>> >>> be 1
> >>> >>> >> implementation.
> >>> >>> >>
> >>> >>> >>
> >>> >>> >>> * discuss replacement of logging framework
> >>> >>> >>
> >>> >>> >> +1
> >>> >>> >>
> >>> >>> >>> * look at memory consumption of the tree view listener
> >>> >>> >>>   ideas where:
> >>> >>> >>>     - store only the last/first X entries
> >>> >>> >>>     - store only marked entries (might be a header field)
> >>> >>> >>>
> >>> >>> >>>
> >>> >>> >>>
> >>> >>> >>>> As side features:
> >>> >>> >>>> - DSL ?
> >>> >>> >>>>
> >>> >>> >>> nice idea, but this is a lot of work
> >>> >>> >>
> >>> >>> >>
> >>> >>> >>
> >>> >>> >>> - JSON format instead of XMLfor jmx plans if dsl is too heavy
> >>> change ?
> >>> >>> >>>>
> >>> >>> >>> I think xml has served us well. Every other format has to
> prove,
> >>> that
> >>> >>> it
> >>> >>> >>> really can compete.
> >>> >>> >>>
> >>> >>> >>>> - Fix undo /redo feature
> >>> >>> >>>>
> >>> >>> >>> +1 for looking at undo/redo
> >>> >>> >>
> >>> >>> >> I would really love to see this one fixed or dropped.
> >>> >>> >> But I don't want to spend too much energy on it.
> >>> >>> >>
> >>> >>> >>
> >>> >>> >>
> >>> >>> >>
> >>> >>> >>>
> >>> >>> >>> Regards,
> >>> >>> >>>  Felix
> >>> >>> >>>
> >>> >>> >>>> - ...
> >>> >>> >>>>
> >>> >>> >>>> Ideas welcome
> >>> >>> >>>> Regards
> >>> >>> >>>> Philippe M.
> >>> >>> >>>>
> >>> >>> >>>>
> >>> >>> >>>>
> >>> >>> >>>>
> >>> >>> >>>
> >>> >>> >>
> >>> >>> >> --
> >>> >>> >> Cordialement.
> >>> >>> >> Philippe Mouawad.
> >>> >>> >>
> >>> >>> >>
> >>> >>> >>
> >>> >>> >>
> >>> >>> >
> >>> >>> > --
> >>> >>> > Cordialement.
> >>> >>> > Philippe Mouawad.
> >>> >>>
> >>> >>
> >>> >>
> >>> >>
> >>> >> --
> >>> >> Cordialement.
> >>> >> Philippe Mouawad.
> >>> >>
> >>> >>
> >>> >>
> >>> >
> >>> >
> >>> > --
> >>> > Cordialement.
> >>> > Philippe Mouawad.
> >>>
> >>
> >>
> >>
> >> --
> >> Cordialement.
> >> Philippe Mouawad.
> >>
> >>
> >>
> >
> >
> > --
> > Cordialement.
> > Philippe Mouawad.
>


-- 
Cordialement.
Philippe Mouawad.